upload model weights after 2400 training steps with frozen pretrained weights, this performs worse than checkpoint 1000 on test set T_T
Browse files- codebert-base-cross-entropy/checkpoint-3800/config.json +27 -0
- codebert-base-cross-entropy/checkpoint-3800/merges.txt +0 -0
- codebert-base-cross-entropy/checkpoint-3800/model.safetensors +3 -0
- codebert-base-cross-entropy/checkpoint-3800/optimizer.pt +3 -0
- codebert-base-cross-entropy/checkpoint-3800/rng_state.pth +3 -0
- codebert-base-cross-entropy/checkpoint-3800/scaler.pt +3 -0
- codebert-base-cross-entropy/checkpoint-3800/scheduler.pt +3 -0
- codebert-base-cross-entropy/checkpoint-3800/special_tokens_map.json +51 -0
- codebert-base-cross-entropy/checkpoint-3800/tokenizer.json +0 -0
- codebert-base-cross-entropy/checkpoint-3800/tokenizer_config.json +58 -0
- codebert-base-cross-entropy/checkpoint-3800/trainer_state.json +2799 -0
- codebert-base-cross-entropy/checkpoint-3800/training_args.bin +3 -0
- codebert-base-cross-entropy/checkpoint-3800/vocab.json +0 -0
- codebert-base-cross-entropy/checkpoint-4000/config.json +27 -0
- codebert-base-cross-entropy/checkpoint-4000/merges.txt +0 -0
- codebert-base-cross-entropy/checkpoint-4000/model.safetensors +3 -0
- codebert-base-cross-entropy/checkpoint-4000/optimizer.pt +3 -0
- codebert-base-cross-entropy/checkpoint-4000/rng_state.pth +3 -0
- codebert-base-cross-entropy/checkpoint-4000/scaler.pt +3 -0
- codebert-base-cross-entropy/checkpoint-4000/scheduler.pt +3 -0
- codebert-base-cross-entropy/checkpoint-4000/special_tokens_map.json +51 -0
- codebert-base-cross-entropy/checkpoint-4000/tokenizer.json +0 -0
- codebert-base-cross-entropy/checkpoint-4000/tokenizer_config.json +58 -0
- codebert-base-cross-entropy/checkpoint-4000/trainer_state.json +2951 -0
- codebert-base-cross-entropy/checkpoint-4000/training_args.bin +3 -0
- codebert-base-cross-entropy/checkpoint-4000/vocab.json +0 -0
- codebert-base-cross-entropy/training.log +114 -60
codebert-base-cross-entropy/checkpoint-3800/config.json
ADDED
|
@@ -0,0 +1,27 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"RobertaForSequenceClassification"
|
| 4 |
+
],
|
| 5 |
+
"attention_probs_dropout_prob": 0.1,
|
| 6 |
+
"bos_token_id": 0,
|
| 7 |
+
"classifier_dropout": null,
|
| 8 |
+
"dtype": "float32",
|
| 9 |
+
"eos_token_id": 2,
|
| 10 |
+
"hidden_act": "gelu",
|
| 11 |
+
"hidden_dropout_prob": 0.1,
|
| 12 |
+
"hidden_size": 768,
|
| 13 |
+
"initializer_range": 0.02,
|
| 14 |
+
"intermediate_size": 3072,
|
| 15 |
+
"layer_norm_eps": 1e-05,
|
| 16 |
+
"max_position_embeddings": 514,
|
| 17 |
+
"model_type": "roberta",
|
| 18 |
+
"num_attention_heads": 12,
|
| 19 |
+
"num_hidden_layers": 12,
|
| 20 |
+
"output_past": true,
|
| 21 |
+
"pad_token_id": 1,
|
| 22 |
+
"position_embedding_type": "absolute",
|
| 23 |
+
"transformers_version": "4.56.0",
|
| 24 |
+
"type_vocab_size": 1,
|
| 25 |
+
"use_cache": true,
|
| 26 |
+
"vocab_size": 50265
|
| 27 |
+
}
|
codebert-base-cross-entropy/checkpoint-3800/merges.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
codebert-base-cross-entropy/checkpoint-3800/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6031ed9355f4bb93cc424d55e3a52229dd6d62203e13678081e0176f8498ca62
|
| 3 |
+
size 498612824
|
codebert-base-cross-entropy/checkpoint-3800/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e82f4498e7198e81b562310f4ff40c7971d53a8a84bff0b71e58281976613451
|
| 3 |
+
size 4741859
|
codebert-base-cross-entropy/checkpoint-3800/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2b427a0205e14d6d4032ab551a6a7140c9e60c809edd788f977f847d1512d0fb
|
| 3 |
+
size 14645
|
codebert-base-cross-entropy/checkpoint-3800/scaler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:dbb840e1609cf0fcdfb28db1eba7346a1598ec64f055428e0b0daef59592d19e
|
| 3 |
+
size 1383
|
codebert-base-cross-entropy/checkpoint-3800/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bcbdb018f18551e0320a4c96f210f7c7ba00d1c99f1a27c69af2ce94b0a1ff75
|
| 3 |
+
size 1465
|
codebert-base-cross-entropy/checkpoint-3800/special_tokens_map.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": true,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"cls_token": {
|
| 10 |
+
"content": "<s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": true,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"eos_token": {
|
| 17 |
+
"content": "</s>",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": true,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"mask_token": {
|
| 24 |
+
"content": "<mask>",
|
| 25 |
+
"lstrip": true,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"pad_token": {
|
| 31 |
+
"content": "<pad>",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": true,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
},
|
| 37 |
+
"sep_token": {
|
| 38 |
+
"content": "</s>",
|
| 39 |
+
"lstrip": false,
|
| 40 |
+
"normalized": true,
|
| 41 |
+
"rstrip": false,
|
| 42 |
+
"single_word": false
|
| 43 |
+
},
|
| 44 |
+
"unk_token": {
|
| 45 |
+
"content": "<unk>",
|
| 46 |
+
"lstrip": false,
|
| 47 |
+
"normalized": true,
|
| 48 |
+
"rstrip": false,
|
| 49 |
+
"single_word": false
|
| 50 |
+
}
|
| 51 |
+
}
|
codebert-base-cross-entropy/checkpoint-3800/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
codebert-base-cross-entropy/checkpoint-3800/tokenizer_config.json
ADDED
|
@@ -0,0 +1,58 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"added_tokens_decoder": {
|
| 4 |
+
"0": {
|
| 5 |
+
"content": "<s>",
|
| 6 |
+
"lstrip": false,
|
| 7 |
+
"normalized": true,
|
| 8 |
+
"rstrip": false,
|
| 9 |
+
"single_word": false,
|
| 10 |
+
"special": true
|
| 11 |
+
},
|
| 12 |
+
"1": {
|
| 13 |
+
"content": "<pad>",
|
| 14 |
+
"lstrip": false,
|
| 15 |
+
"normalized": true,
|
| 16 |
+
"rstrip": false,
|
| 17 |
+
"single_word": false,
|
| 18 |
+
"special": true
|
| 19 |
+
},
|
| 20 |
+
"2": {
|
| 21 |
+
"content": "</s>",
|
| 22 |
+
"lstrip": false,
|
| 23 |
+
"normalized": true,
|
| 24 |
+
"rstrip": false,
|
| 25 |
+
"single_word": false,
|
| 26 |
+
"special": true
|
| 27 |
+
},
|
| 28 |
+
"3": {
|
| 29 |
+
"content": "<unk>",
|
| 30 |
+
"lstrip": false,
|
| 31 |
+
"normalized": true,
|
| 32 |
+
"rstrip": false,
|
| 33 |
+
"single_word": false,
|
| 34 |
+
"special": true
|
| 35 |
+
},
|
| 36 |
+
"50264": {
|
| 37 |
+
"content": "<mask>",
|
| 38 |
+
"lstrip": true,
|
| 39 |
+
"normalized": false,
|
| 40 |
+
"rstrip": false,
|
| 41 |
+
"single_word": false,
|
| 42 |
+
"special": true
|
| 43 |
+
}
|
| 44 |
+
},
|
| 45 |
+
"bos_token": "<s>",
|
| 46 |
+
"clean_up_tokenization_spaces": false,
|
| 47 |
+
"cls_token": "<s>",
|
| 48 |
+
"eos_token": "</s>",
|
| 49 |
+
"errors": "replace",
|
| 50 |
+
"extra_special_tokens": {},
|
| 51 |
+
"mask_token": "<mask>",
|
| 52 |
+
"model_max_length": 512,
|
| 53 |
+
"pad_token": "<pad>",
|
| 54 |
+
"sep_token": "</s>",
|
| 55 |
+
"tokenizer_class": "RobertaTokenizer",
|
| 56 |
+
"trim_offsets": true,
|
| 57 |
+
"unk_token": "<unk>"
|
| 58 |
+
}
|
codebert-base-cross-entropy/checkpoint-3800/trainer_state.json
ADDED
|
@@ -0,0 +1,2799 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 2000,
|
| 3 |
+
"best_metric": 0.743516265066894,
|
| 4 |
+
"best_model_checkpoint": "./output_checkpoints/codebert-base-cross-entropy/checkpoint-2000",
|
| 5 |
+
"epoch": 0.4863688723921669,
|
| 6 |
+
"eval_steps": 2000,
|
| 7 |
+
"global_step": 3800,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.0012799180852425445,
|
| 14 |
+
"grad_norm": 3.6954128742218018,
|
| 15 |
+
"learning_rate": 3.840409643695328e-08,
|
| 16 |
+
"loss": 0.7763,
|
| 17 |
+
"step": 10
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.002559836170485089,
|
| 21 |
+
"grad_norm": 3.3660645484924316,
|
| 22 |
+
"learning_rate": 8.10753147002347e-08,
|
| 23 |
+
"loss": 0.7763,
|
| 24 |
+
"step": 20
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.0038397542557276334,
|
| 28 |
+
"grad_norm": 4.491067886352539,
|
| 29 |
+
"learning_rate": 1.2374653296351612e-07,
|
| 30 |
+
"loss": 0.7889,
|
| 31 |
+
"step": 30
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.005119672340970178,
|
| 35 |
+
"grad_norm": 1.5920521020889282,
|
| 36 |
+
"learning_rate": 1.6641775122679754e-07,
|
| 37 |
+
"loss": 0.7888,
|
| 38 |
+
"step": 40
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.006399590426212722,
|
| 42 |
+
"grad_norm": 4.410898208618164,
|
| 43 |
+
"learning_rate": 2.0908896949007894e-07,
|
| 44 |
+
"loss": 0.7692,
|
| 45 |
+
"step": 50
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.007679508511455267,
|
| 49 |
+
"grad_norm": 2.5669264793395996,
|
| 50 |
+
"learning_rate": 2.517601877533604e-07,
|
| 51 |
+
"loss": 0.7735,
|
| 52 |
+
"step": 60
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.008959426596697812,
|
| 56 |
+
"grad_norm": 6.190986156463623,
|
| 57 |
+
"learning_rate": 2.944314060166418e-07,
|
| 58 |
+
"loss": 0.7679,
|
| 59 |
+
"step": 70
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 0.010239344681940356,
|
| 63 |
+
"grad_norm": 5.736413478851318,
|
| 64 |
+
"learning_rate": 3.371026242799232e-07,
|
| 65 |
+
"loss": 0.744,
|
| 66 |
+
"step": 80
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"epoch": 0.011519262767182901,
|
| 70 |
+
"grad_norm": 4.278866767883301,
|
| 71 |
+
"learning_rate": 3.7977384254320464e-07,
|
| 72 |
+
"loss": 0.7542,
|
| 73 |
+
"step": 90
|
| 74 |
+
},
|
| 75 |
+
{
|
| 76 |
+
"epoch": 0.012799180852425445,
|
| 77 |
+
"grad_norm": 5.565455436706543,
|
| 78 |
+
"learning_rate": 4.22445060806486e-07,
|
| 79 |
+
"loss": 0.752,
|
| 80 |
+
"step": 100
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.01407909893766799,
|
| 84 |
+
"grad_norm": 4.031781196594238,
|
| 85 |
+
"learning_rate": 4.651162790697675e-07,
|
| 86 |
+
"loss": 0.7614,
|
| 87 |
+
"step": 110
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.015359017022910534,
|
| 91 |
+
"grad_norm": 2.484126329421997,
|
| 92 |
+
"learning_rate": 5.077874973330489e-07,
|
| 93 |
+
"loss": 0.7664,
|
| 94 |
+
"step": 120
|
| 95 |
+
},
|
| 96 |
+
{
|
| 97 |
+
"epoch": 0.016638935108153077,
|
| 98 |
+
"grad_norm": 2.7525155544281006,
|
| 99 |
+
"learning_rate": 5.504587155963304e-07,
|
| 100 |
+
"loss": 0.7521,
|
| 101 |
+
"step": 130
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"epoch": 0.017918853193395624,
|
| 105 |
+
"grad_norm": 1.6374385356903076,
|
| 106 |
+
"learning_rate": 5.931299338596117e-07,
|
| 107 |
+
"loss": 0.7436,
|
| 108 |
+
"step": 140
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"epoch": 0.019198771278638168,
|
| 112 |
+
"grad_norm": 2.370774269104004,
|
| 113 |
+
"learning_rate": 6.358011521228932e-07,
|
| 114 |
+
"loss": 0.7429,
|
| 115 |
+
"step": 150
|
| 116 |
+
},
|
| 117 |
+
{
|
| 118 |
+
"epoch": 0.02047868936388071,
|
| 119 |
+
"grad_norm": 2.8948566913604736,
|
| 120 |
+
"learning_rate": 6.784723703861745e-07,
|
| 121 |
+
"loss": 0.7333,
|
| 122 |
+
"step": 160
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.021758607449123255,
|
| 126 |
+
"grad_norm": 2.7481186389923096,
|
| 127 |
+
"learning_rate": 7.21143588649456e-07,
|
| 128 |
+
"loss": 0.7294,
|
| 129 |
+
"step": 170
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.023038525534365802,
|
| 133 |
+
"grad_norm": 2.0159404277801514,
|
| 134 |
+
"learning_rate": 7.638148069127374e-07,
|
| 135 |
+
"loss": 0.731,
|
| 136 |
+
"step": 180
|
| 137 |
+
},
|
| 138 |
+
{
|
| 139 |
+
"epoch": 0.024318443619608346,
|
| 140 |
+
"grad_norm": 3.6898131370544434,
|
| 141 |
+
"learning_rate": 8.064860251760189e-07,
|
| 142 |
+
"loss": 0.7153,
|
| 143 |
+
"step": 190
|
| 144 |
+
},
|
| 145 |
+
{
|
| 146 |
+
"epoch": 0.02559836170485089,
|
| 147 |
+
"grad_norm": 3.2842342853546143,
|
| 148 |
+
"learning_rate": 8.491572434393003e-07,
|
| 149 |
+
"loss": 0.7245,
|
| 150 |
+
"step": 200
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"epoch": 0.02559836170485089,
|
| 154 |
+
"eval_accuracy": 0.47695,
|
| 155 |
+
"eval_loss": 0.7148504853248596,
|
| 156 |
+
"eval_macro_f1": 0.3229290091066048,
|
| 157 |
+
"eval_precision": 0.238475,
|
| 158 |
+
"eval_recall": 0.5,
|
| 159 |
+
"eval_runtime": 1641.324,
|
| 160 |
+
"eval_samples_per_second": 60.926,
|
| 161 |
+
"eval_steps_per_second": 0.476,
|
| 162 |
+
"step": 200
|
| 163 |
+
},
|
| 164 |
+
{
|
| 165 |
+
"epoch": 0.026878279790093433,
|
| 166 |
+
"grad_norm": 2.294558048248291,
|
| 167 |
+
"learning_rate": 8.918284617025817e-07,
|
| 168 |
+
"loss": 0.718,
|
| 169 |
+
"step": 210
|
| 170 |
+
},
|
| 171 |
+
{
|
| 172 |
+
"epoch": 0.02815819787533598,
|
| 173 |
+
"grad_norm": 2.9467175006866455,
|
| 174 |
+
"learning_rate": 9.344996799658632e-07,
|
| 175 |
+
"loss": 0.7076,
|
| 176 |
+
"step": 220
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.029438115960578524,
|
| 180 |
+
"grad_norm": 2.6889193058013916,
|
| 181 |
+
"learning_rate": 9.771708982291445e-07,
|
| 182 |
+
"loss": 0.7019,
|
| 183 |
+
"step": 230
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 0.030718034045821067,
|
| 187 |
+
"grad_norm": 0.7491384744644165,
|
| 188 |
+
"learning_rate": 1.0198421164924258e-06,
|
| 189 |
+
"loss": 0.7046,
|
| 190 |
+
"step": 240
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 0.031997952131063614,
|
| 194 |
+
"grad_norm": 2.0363433361053467,
|
| 195 |
+
"learning_rate": 1.0625133347557074e-06,
|
| 196 |
+
"loss": 0.7085,
|
| 197 |
+
"step": 250
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 0.033277870216306155,
|
| 201 |
+
"grad_norm": 1.1431337594985962,
|
| 202 |
+
"learning_rate": 1.1051845530189888e-06,
|
| 203 |
+
"loss": 0.7008,
|
| 204 |
+
"step": 260
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 0.0345577883015487,
|
| 208 |
+
"grad_norm": 3.7735233306884766,
|
| 209 |
+
"learning_rate": 1.1478557712822702e-06,
|
| 210 |
+
"loss": 0.7015,
|
| 211 |
+
"step": 270
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 0.03583770638679125,
|
| 215 |
+
"grad_norm": 0.7889328002929688,
|
| 216 |
+
"learning_rate": 1.1905269895455517e-06,
|
| 217 |
+
"loss": 0.699,
|
| 218 |
+
"step": 280
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 0.03711762447203379,
|
| 222 |
+
"grad_norm": 0.8796226382255554,
|
| 223 |
+
"learning_rate": 1.233198207808833e-06,
|
| 224 |
+
"loss": 0.6922,
|
| 225 |
+
"step": 290
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 0.038397542557276336,
|
| 229 |
+
"grad_norm": 0.7966652512550354,
|
| 230 |
+
"learning_rate": 1.2758694260721145e-06,
|
| 231 |
+
"loss": 0.6949,
|
| 232 |
+
"step": 300
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 0.039677460642518876,
|
| 236 |
+
"grad_norm": 0.7029069066047668,
|
| 237 |
+
"learning_rate": 1.318540644335396e-06,
|
| 238 |
+
"loss": 0.695,
|
| 239 |
+
"step": 310
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 0.04095737872776142,
|
| 243 |
+
"grad_norm": 2.04327392578125,
|
| 244 |
+
"learning_rate": 1.3612118625986772e-06,
|
| 245 |
+
"loss": 0.6944,
|
| 246 |
+
"step": 320
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 0.04223729681300397,
|
| 250 |
+
"grad_norm": 0.918025553226471,
|
| 251 |
+
"learning_rate": 1.4038830808619588e-06,
|
| 252 |
+
"loss": 0.6863,
|
| 253 |
+
"step": 330
|
| 254 |
+
},
|
| 255 |
+
{
|
| 256 |
+
"epoch": 0.04351721489824651,
|
| 257 |
+
"grad_norm": 0.9528706073760986,
|
| 258 |
+
"learning_rate": 1.4465542991252401e-06,
|
| 259 |
+
"loss": 0.6907,
|
| 260 |
+
"step": 340
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 0.04479713298348906,
|
| 264 |
+
"grad_norm": 1.071359634399414,
|
| 265 |
+
"learning_rate": 1.4892255173885215e-06,
|
| 266 |
+
"loss": 0.6873,
|
| 267 |
+
"step": 350
|
| 268 |
+
},
|
| 269 |
+
{
|
| 270 |
+
"epoch": 0.046077051068731605,
|
| 271 |
+
"grad_norm": 1.9875625371932983,
|
| 272 |
+
"learning_rate": 1.531896735651803e-06,
|
| 273 |
+
"loss": 0.6942,
|
| 274 |
+
"step": 360
|
| 275 |
+
},
|
| 276 |
+
{
|
| 277 |
+
"epoch": 0.047356969153974145,
|
| 278 |
+
"grad_norm": 2.0222439765930176,
|
| 279 |
+
"learning_rate": 1.5745679539150842e-06,
|
| 280 |
+
"loss": 0.6898,
|
| 281 |
+
"step": 370
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 0.04863688723921669,
|
| 285 |
+
"grad_norm": 1.7827857732772827,
|
| 286 |
+
"learning_rate": 1.6172391721783658e-06,
|
| 287 |
+
"loss": 0.6909,
|
| 288 |
+
"step": 380
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"epoch": 0.04991680532445923,
|
| 292 |
+
"grad_norm": 1.065674901008606,
|
| 293 |
+
"learning_rate": 1.6599103904416472e-06,
|
| 294 |
+
"loss": 0.6877,
|
| 295 |
+
"step": 390
|
| 296 |
+
},
|
| 297 |
+
{
|
| 298 |
+
"epoch": 0.05119672340970178,
|
| 299 |
+
"grad_norm": 0.8346193432807922,
|
| 300 |
+
"learning_rate": 1.7025816087049288e-06,
|
| 301 |
+
"loss": 0.6889,
|
| 302 |
+
"step": 400
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 0.05119672340970178,
|
| 306 |
+
"eval_accuracy": 0.54043,
|
| 307 |
+
"eval_loss": 0.6827595233917236,
|
| 308 |
+
"eval_macro_f1": 0.41833840347598905,
|
| 309 |
+
"eval_precision": 0.5839205994464474,
|
| 310 |
+
"eval_recall": 0.5204152767503025,
|
| 311 |
+
"eval_runtime": 1634.6244,
|
| 312 |
+
"eval_samples_per_second": 61.176,
|
| 313 |
+
"eval_steps_per_second": 0.478,
|
| 314 |
+
"step": 400
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 0.052476641494944326,
|
| 318 |
+
"grad_norm": 0.8074935674667358,
|
| 319 |
+
"learning_rate": 1.7452528269682101e-06,
|
| 320 |
+
"loss": 0.684,
|
| 321 |
+
"step": 410
|
| 322 |
+
},
|
| 323 |
+
{
|
| 324 |
+
"epoch": 0.053756559580186866,
|
| 325 |
+
"grad_norm": 0.8965069651603699,
|
| 326 |
+
"learning_rate": 1.7879240452314913e-06,
|
| 327 |
+
"loss": 0.6922,
|
| 328 |
+
"step": 420
|
| 329 |
+
},
|
| 330 |
+
{
|
| 331 |
+
"epoch": 0.05503647766542941,
|
| 332 |
+
"grad_norm": 2.597215414047241,
|
| 333 |
+
"learning_rate": 1.8305952634947729e-06,
|
| 334 |
+
"loss": 0.6909,
|
| 335 |
+
"step": 430
|
| 336 |
+
},
|
| 337 |
+
{
|
| 338 |
+
"epoch": 0.05631639575067196,
|
| 339 |
+
"grad_norm": 0.8233373165130615,
|
| 340 |
+
"learning_rate": 1.8732664817580542e-06,
|
| 341 |
+
"loss": 0.6831,
|
| 342 |
+
"step": 440
|
| 343 |
+
},
|
| 344 |
+
{
|
| 345 |
+
"epoch": 0.0575963138359145,
|
| 346 |
+
"grad_norm": 1.0624489784240723,
|
| 347 |
+
"learning_rate": 1.915937700021336e-06,
|
| 348 |
+
"loss": 0.6817,
|
| 349 |
+
"step": 450
|
| 350 |
+
},
|
| 351 |
+
{
|
| 352 |
+
"epoch": 0.05887623192115705,
|
| 353 |
+
"grad_norm": 2.0623559951782227,
|
| 354 |
+
"learning_rate": 1.958608918284617e-06,
|
| 355 |
+
"loss": 0.6819,
|
| 356 |
+
"step": 460
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 0.06015615000639959,
|
| 360 |
+
"grad_norm": 0.912433922290802,
|
| 361 |
+
"learning_rate": 2.0012801365478988e-06,
|
| 362 |
+
"loss": 0.686,
|
| 363 |
+
"step": 470
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"epoch": 0.061436068091642135,
|
| 367 |
+
"grad_norm": 0.6926271319389343,
|
| 368 |
+
"learning_rate": 2.04395135481118e-06,
|
| 369 |
+
"loss": 0.6834,
|
| 370 |
+
"step": 480
|
| 371 |
+
},
|
| 372 |
+
{
|
| 373 |
+
"epoch": 0.06271598617688467,
|
| 374 |
+
"grad_norm": 1.498634934425354,
|
| 375 |
+
"learning_rate": 2.0866225730744615e-06,
|
| 376 |
+
"loss": 0.6805,
|
| 377 |
+
"step": 490
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 0.06399590426212723,
|
| 381 |
+
"grad_norm": 1.1818156242370605,
|
| 382 |
+
"learning_rate": 2.129293791337743e-06,
|
| 383 |
+
"loss": 0.6805,
|
| 384 |
+
"step": 500
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 0.06527582234736977,
|
| 388 |
+
"grad_norm": 1.4476051330566406,
|
| 389 |
+
"learning_rate": 2.1719650096010242e-06,
|
| 390 |
+
"loss": 0.6773,
|
| 391 |
+
"step": 510
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 0.06655574043261231,
|
| 395 |
+
"grad_norm": 1.1143743991851807,
|
| 396 |
+
"learning_rate": 2.214636227864306e-06,
|
| 397 |
+
"loss": 0.6801,
|
| 398 |
+
"step": 520
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 0.06783565851785486,
|
| 402 |
+
"grad_norm": 1.7030247449874878,
|
| 403 |
+
"learning_rate": 2.257307446127587e-06,
|
| 404 |
+
"loss": 0.6794,
|
| 405 |
+
"step": 530
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 0.0691155766030974,
|
| 409 |
+
"grad_norm": 0.8592290282249451,
|
| 410 |
+
"learning_rate": 2.2999786643908685e-06,
|
| 411 |
+
"loss": 0.6736,
|
| 412 |
+
"step": 540
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 0.07039549468833994,
|
| 416 |
+
"grad_norm": 0.8231940865516663,
|
| 417 |
+
"learning_rate": 2.34264988265415e-06,
|
| 418 |
+
"loss": 0.67,
|
| 419 |
+
"step": 550
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 0.0716754127735825,
|
| 423 |
+
"grad_norm": 0.88897705078125,
|
| 424 |
+
"learning_rate": 2.3853211009174317e-06,
|
| 425 |
+
"loss": 0.685,
|
| 426 |
+
"step": 560
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 0.07295533085882504,
|
| 430 |
+
"grad_norm": 1.6297541856765747,
|
| 431 |
+
"learning_rate": 2.427992319180713e-06,
|
| 432 |
+
"loss": 0.6797,
|
| 433 |
+
"step": 570
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 0.07423524894406758,
|
| 437 |
+
"grad_norm": 2.9297447204589844,
|
| 438 |
+
"learning_rate": 2.470663537443994e-06,
|
| 439 |
+
"loss": 0.6766,
|
| 440 |
+
"step": 580
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 0.07551516702931012,
|
| 444 |
+
"grad_norm": 0.9277703762054443,
|
| 445 |
+
"learning_rate": 2.5133347557072756e-06,
|
| 446 |
+
"loss": 0.6824,
|
| 447 |
+
"step": 590
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 0.07679508511455267,
|
| 451 |
+
"grad_norm": 0.5883333683013916,
|
| 452 |
+
"learning_rate": 2.556005973970557e-06,
|
| 453 |
+
"loss": 0.6818,
|
| 454 |
+
"step": 600
|
| 455 |
+
},
|
| 456 |
+
{
|
| 457 |
+
"epoch": 0.07679508511455267,
|
| 458 |
+
"eval_accuracy": 0.63909,
|
| 459 |
+
"eval_loss": 0.6700075268745422,
|
| 460 |
+
"eval_macro_f1": 0.621848143053056,
|
| 461 |
+
"eval_precision": 0.6524313071088819,
|
| 462 |
+
"eval_recall": 0.6305863955136095,
|
| 463 |
+
"eval_runtime": 1633.9732,
|
| 464 |
+
"eval_samples_per_second": 61.201,
|
| 465 |
+
"eval_steps_per_second": 0.479,
|
| 466 |
+
"step": 600
|
| 467 |
+
},
|
| 468 |
+
{
|
| 469 |
+
"epoch": 0.07807500319979521,
|
| 470 |
+
"grad_norm": 0.7149448394775391,
|
| 471 |
+
"learning_rate": 2.5986771922338383e-06,
|
| 472 |
+
"loss": 0.6746,
|
| 473 |
+
"step": 610
|
| 474 |
+
},
|
| 475 |
+
{
|
| 476 |
+
"epoch": 0.07935492128503775,
|
| 477 |
+
"grad_norm": 2.0167856216430664,
|
| 478 |
+
"learning_rate": 2.64134841049712e-06,
|
| 479 |
+
"loss": 0.682,
|
| 480 |
+
"step": 620
|
| 481 |
+
},
|
| 482 |
+
{
|
| 483 |
+
"epoch": 0.0806348393702803,
|
| 484 |
+
"grad_norm": 1.303001880645752,
|
| 485 |
+
"learning_rate": 2.6840196287604015e-06,
|
| 486 |
+
"loss": 0.6735,
|
| 487 |
+
"step": 630
|
| 488 |
+
},
|
| 489 |
+
{
|
| 490 |
+
"epoch": 0.08191475745552285,
|
| 491 |
+
"grad_norm": 2.0516321659088135,
|
| 492 |
+
"learning_rate": 2.7266908470236826e-06,
|
| 493 |
+
"loss": 0.6705,
|
| 494 |
+
"step": 640
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 0.08319467554076539,
|
| 498 |
+
"grad_norm": 2.242105484008789,
|
| 499 |
+
"learning_rate": 2.7693620652869642e-06,
|
| 500 |
+
"loss": 0.6798,
|
| 501 |
+
"step": 650
|
| 502 |
+
},
|
| 503 |
+
{
|
| 504 |
+
"epoch": 0.08447459362600794,
|
| 505 |
+
"grad_norm": 1.1836538314819336,
|
| 506 |
+
"learning_rate": 2.812033283550246e-06,
|
| 507 |
+
"loss": 0.6684,
|
| 508 |
+
"step": 660
|
| 509 |
+
},
|
| 510 |
+
{
|
| 511 |
+
"epoch": 0.08575451171125048,
|
| 512 |
+
"grad_norm": 0.8640874624252319,
|
| 513 |
+
"learning_rate": 2.8547045018135274e-06,
|
| 514 |
+
"loss": 0.6783,
|
| 515 |
+
"step": 670
|
| 516 |
+
},
|
| 517 |
+
{
|
| 518 |
+
"epoch": 0.08703442979649302,
|
| 519 |
+
"grad_norm": 0.6901177763938904,
|
| 520 |
+
"learning_rate": 2.897375720076808e-06,
|
| 521 |
+
"loss": 0.6724,
|
| 522 |
+
"step": 680
|
| 523 |
+
},
|
| 524 |
+
{
|
| 525 |
+
"epoch": 0.08831434788173557,
|
| 526 |
+
"grad_norm": 1.3443949222564697,
|
| 527 |
+
"learning_rate": 2.9400469383400897e-06,
|
| 528 |
+
"loss": 0.67,
|
| 529 |
+
"step": 690
|
| 530 |
+
},
|
| 531 |
+
{
|
| 532 |
+
"epoch": 0.08959426596697811,
|
| 533 |
+
"grad_norm": 0.7775189280509949,
|
| 534 |
+
"learning_rate": 2.9827181566033713e-06,
|
| 535 |
+
"loss": 0.6733,
|
| 536 |
+
"step": 700
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 0.09087418405222066,
|
| 540 |
+
"grad_norm": 1.16916823387146,
|
| 541 |
+
"learning_rate": 3.025389374866653e-06,
|
| 542 |
+
"loss": 0.6713,
|
| 543 |
+
"step": 710
|
| 544 |
+
},
|
| 545 |
+
{
|
| 546 |
+
"epoch": 0.09215410213746321,
|
| 547 |
+
"grad_norm": 1.707152009010315,
|
| 548 |
+
"learning_rate": 3.068060593129934e-06,
|
| 549 |
+
"loss": 0.6659,
|
| 550 |
+
"step": 720
|
| 551 |
+
},
|
| 552 |
+
{
|
| 553 |
+
"epoch": 0.09343402022270575,
|
| 554 |
+
"grad_norm": 0.9780622124671936,
|
| 555 |
+
"learning_rate": 3.1107318113932156e-06,
|
| 556 |
+
"loss": 0.6639,
|
| 557 |
+
"step": 730
|
| 558 |
+
},
|
| 559 |
+
{
|
| 560 |
+
"epoch": 0.09471393830794829,
|
| 561 |
+
"grad_norm": 2.3687374591827393,
|
| 562 |
+
"learning_rate": 3.153403029656497e-06,
|
| 563 |
+
"loss": 0.6666,
|
| 564 |
+
"step": 740
|
| 565 |
+
},
|
| 566 |
+
{
|
| 567 |
+
"epoch": 0.09599385639319083,
|
| 568 |
+
"grad_norm": 0.8506460189819336,
|
| 569 |
+
"learning_rate": 3.1960742479197783e-06,
|
| 570 |
+
"loss": 0.6778,
|
| 571 |
+
"step": 750
|
| 572 |
+
},
|
| 573 |
+
{
|
| 574 |
+
"epoch": 0.09727377447843338,
|
| 575 |
+
"grad_norm": 1.0102107524871826,
|
| 576 |
+
"learning_rate": 3.23874546618306e-06,
|
| 577 |
+
"loss": 0.6635,
|
| 578 |
+
"step": 760
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 0.09855369256367592,
|
| 582 |
+
"grad_norm": 1.1480194330215454,
|
| 583 |
+
"learning_rate": 3.2814166844463415e-06,
|
| 584 |
+
"loss": 0.6709,
|
| 585 |
+
"step": 770
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 0.09983361064891846,
|
| 589 |
+
"grad_norm": 1.3119179010391235,
|
| 590 |
+
"learning_rate": 3.324087902709623e-06,
|
| 591 |
+
"loss": 0.6564,
|
| 592 |
+
"step": 780
|
| 593 |
+
},
|
| 594 |
+
{
|
| 595 |
+
"epoch": 0.10111352873416102,
|
| 596 |
+
"grad_norm": 0.681377649307251,
|
| 597 |
+
"learning_rate": 3.3667591209729038e-06,
|
| 598 |
+
"loss": 0.6656,
|
| 599 |
+
"step": 790
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"epoch": 0.10239344681940356,
|
| 603 |
+
"grad_norm": 1.022257685661316,
|
| 604 |
+
"learning_rate": 3.4094303392361854e-06,
|
| 605 |
+
"loss": 0.6709,
|
| 606 |
+
"step": 800
|
| 607 |
+
},
|
| 608 |
+
{
|
| 609 |
+
"epoch": 0.10239344681940356,
|
| 610 |
+
"eval_accuracy": 0.66247,
|
| 611 |
+
"eval_loss": 0.6552017331123352,
|
| 612 |
+
"eval_macro_f1": 0.6514270557181998,
|
| 613 |
+
"eval_precision": 0.6718270655389242,
|
| 614 |
+
"eval_recall": 0.6556580720913894,
|
| 615 |
+
"eval_runtime": 1633.5478,
|
| 616 |
+
"eval_samples_per_second": 61.216,
|
| 617 |
+
"eval_steps_per_second": 0.479,
|
| 618 |
+
"step": 800
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 0.1036733649046461,
|
| 622 |
+
"grad_norm": 1.0268975496292114,
|
| 623 |
+
"learning_rate": 3.452101557499467e-06,
|
| 624 |
+
"loss": 0.6602,
|
| 625 |
+
"step": 810
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.10495328298988865,
|
| 629 |
+
"grad_norm": 2.6206603050231934,
|
| 630 |
+
"learning_rate": 3.494772775762748e-06,
|
| 631 |
+
"loss": 0.6705,
|
| 632 |
+
"step": 820
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.10623320107513119,
|
| 636 |
+
"grad_norm": 1.035157561302185,
|
| 637 |
+
"learning_rate": 3.5374439940260297e-06,
|
| 638 |
+
"loss": 0.6643,
|
| 639 |
+
"step": 830
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.10751311916037373,
|
| 643 |
+
"grad_norm": 1.736412763595581,
|
| 644 |
+
"learning_rate": 3.5801152122893113e-06,
|
| 645 |
+
"loss": 0.6612,
|
| 646 |
+
"step": 840
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 0.10879303724561629,
|
| 650 |
+
"grad_norm": 0.6906964778900146,
|
| 651 |
+
"learning_rate": 3.622786430552593e-06,
|
| 652 |
+
"loss": 0.6576,
|
| 653 |
+
"step": 850
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.11007295533085883,
|
| 657 |
+
"grad_norm": 0.8148729801177979,
|
| 658 |
+
"learning_rate": 3.665457648815874e-06,
|
| 659 |
+
"loss": 0.6561,
|
| 660 |
+
"step": 860
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 0.11135287341610137,
|
| 664 |
+
"grad_norm": 1.0217723846435547,
|
| 665 |
+
"learning_rate": 3.7081288670791556e-06,
|
| 666 |
+
"loss": 0.6581,
|
| 667 |
+
"step": 870
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.11263279150134392,
|
| 671 |
+
"grad_norm": 1.2743115425109863,
|
| 672 |
+
"learning_rate": 3.750800085342437e-06,
|
| 673 |
+
"loss": 0.6606,
|
| 674 |
+
"step": 880
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.11391270958658646,
|
| 678 |
+
"grad_norm": 0.6405361294746399,
|
| 679 |
+
"learning_rate": 3.793471303605718e-06,
|
| 680 |
+
"loss": 0.652,
|
| 681 |
+
"step": 890
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.115192627671829,
|
| 685 |
+
"grad_norm": 0.8357161283493042,
|
| 686 |
+
"learning_rate": 3.836142521869e-06,
|
| 687 |
+
"loss": 0.6558,
|
| 688 |
+
"step": 900
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 0.11647254575707154,
|
| 692 |
+
"grad_norm": 1.3505173921585083,
|
| 693 |
+
"learning_rate": 3.8788137401322815e-06,
|
| 694 |
+
"loss": 0.6576,
|
| 695 |
+
"step": 910
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.1177524638423141,
|
| 699 |
+
"grad_norm": 1.2253319025039673,
|
| 700 |
+
"learning_rate": 3.921484958395563e-06,
|
| 701 |
+
"loss": 0.6548,
|
| 702 |
+
"step": 920
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 0.11903238192755664,
|
| 706 |
+
"grad_norm": 0.6269568800926208,
|
| 707 |
+
"learning_rate": 3.964156176658844e-06,
|
| 708 |
+
"loss": 0.6562,
|
| 709 |
+
"step": 930
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.12031230001279918,
|
| 713 |
+
"grad_norm": 0.9413424134254456,
|
| 714 |
+
"learning_rate": 4.006827394922125e-06,
|
| 715 |
+
"loss": 0.6461,
|
| 716 |
+
"step": 940
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.12159221809804173,
|
| 720 |
+
"grad_norm": 1.0903513431549072,
|
| 721 |
+
"learning_rate": 4.049498613185407e-06,
|
| 722 |
+
"loss": 0.6457,
|
| 723 |
+
"step": 950
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.12287213618328427,
|
| 727 |
+
"grad_norm": 2.0023152828216553,
|
| 728 |
+
"learning_rate": 4.092169831448688e-06,
|
| 729 |
+
"loss": 0.6458,
|
| 730 |
+
"step": 960
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 0.12415205426852681,
|
| 734 |
+
"grad_norm": 0.9638109803199768,
|
| 735 |
+
"learning_rate": 4.134841049711969e-06,
|
| 736 |
+
"loss": 0.6486,
|
| 737 |
+
"step": 970
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.12543197235376935,
|
| 741 |
+
"grad_norm": 0.9894546270370483,
|
| 742 |
+
"learning_rate": 4.177512267975251e-06,
|
| 743 |
+
"loss": 0.6453,
|
| 744 |
+
"step": 980
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 0.1267118904390119,
|
| 748 |
+
"grad_norm": 1.2289057970046997,
|
| 749 |
+
"learning_rate": 4.220183486238532e-06,
|
| 750 |
+
"loss": 0.6605,
|
| 751 |
+
"step": 990
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.12799180852425446,
|
| 755 |
+
"grad_norm": 1.119125247001648,
|
| 756 |
+
"learning_rate": 4.262854704501814e-06,
|
| 757 |
+
"loss": 0.6564,
|
| 758 |
+
"step": 1000
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 0.12799180852425446,
|
| 762 |
+
"eval_accuracy": 0.69068,
|
| 763 |
+
"eval_loss": 0.6396247148513794,
|
| 764 |
+
"eval_macro_f1": 0.689588440856387,
|
| 765 |
+
"eval_precision": 0.6900076024604327,
|
| 766 |
+
"eval_recall": 0.6894114140312133,
|
| 767 |
+
"eval_runtime": 1634.4321,
|
| 768 |
+
"eval_samples_per_second": 61.183,
|
| 769 |
+
"eval_steps_per_second": 0.478,
|
| 770 |
+
"step": 1000
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 0.12927172660949698,
|
| 774 |
+
"grad_norm": 2.02966046333313,
|
| 775 |
+
"learning_rate": 4.3055259227650956e-06,
|
| 776 |
+
"loss": 0.646,
|
| 777 |
+
"step": 1010
|
| 778 |
+
},
|
| 779 |
+
{
|
| 780 |
+
"epoch": 0.13055164469473954,
|
| 781 |
+
"grad_norm": 0.7430800199508667,
|
| 782 |
+
"learning_rate": 4.348197141028377e-06,
|
| 783 |
+
"loss": 0.6453,
|
| 784 |
+
"step": 1020
|
| 785 |
+
},
|
| 786 |
+
{
|
| 787 |
+
"epoch": 0.1318315627799821,
|
| 788 |
+
"grad_norm": 1.031962275505066,
|
| 789 |
+
"learning_rate": 4.390868359291658e-06,
|
| 790 |
+
"loss": 0.6425,
|
| 791 |
+
"step": 1030
|
| 792 |
+
},
|
| 793 |
+
{
|
| 794 |
+
"epoch": 0.13311148086522462,
|
| 795 |
+
"grad_norm": 1.0249106884002686,
|
| 796 |
+
"learning_rate": 4.4335395775549394e-06,
|
| 797 |
+
"loss": 0.6426,
|
| 798 |
+
"step": 1040
|
| 799 |
+
},
|
| 800 |
+
{
|
| 801 |
+
"epoch": 0.13439139895046717,
|
| 802 |
+
"grad_norm": 0.7375448942184448,
|
| 803 |
+
"learning_rate": 4.476210795818221e-06,
|
| 804 |
+
"loss": 0.6401,
|
| 805 |
+
"step": 1050
|
| 806 |
+
},
|
| 807 |
+
{
|
| 808 |
+
"epoch": 0.13567131703570973,
|
| 809 |
+
"grad_norm": 1.295983076095581,
|
| 810 |
+
"learning_rate": 4.518882014081503e-06,
|
| 811 |
+
"loss": 0.6449,
|
| 812 |
+
"step": 1060
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 0.13695123512095225,
|
| 816 |
+
"grad_norm": 1.6978998184204102,
|
| 817 |
+
"learning_rate": 4.561553232344783e-06,
|
| 818 |
+
"loss": 0.6484,
|
| 819 |
+
"step": 1070
|
| 820 |
+
},
|
| 821 |
+
{
|
| 822 |
+
"epoch": 0.1382311532061948,
|
| 823 |
+
"grad_norm": 0.7747512459754944,
|
| 824 |
+
"learning_rate": 4.604224450608065e-06,
|
| 825 |
+
"loss": 0.6526,
|
| 826 |
+
"step": 1080
|
| 827 |
+
},
|
| 828 |
+
{
|
| 829 |
+
"epoch": 0.13951107129143736,
|
| 830 |
+
"grad_norm": 0.7895913124084473,
|
| 831 |
+
"learning_rate": 4.6468956688713465e-06,
|
| 832 |
+
"loss": 0.6311,
|
| 833 |
+
"step": 1090
|
| 834 |
+
},
|
| 835 |
+
{
|
| 836 |
+
"epoch": 0.1407909893766799,
|
| 837 |
+
"grad_norm": 0.6975072026252747,
|
| 838 |
+
"learning_rate": 4.689566887134628e-06,
|
| 839 |
+
"loss": 0.6454,
|
| 840 |
+
"step": 1100
|
| 841 |
+
},
|
| 842 |
+
{
|
| 843 |
+
"epoch": 0.14207090746192244,
|
| 844 |
+
"grad_norm": 0.8316811919212341,
|
| 845 |
+
"learning_rate": 4.73223810539791e-06,
|
| 846 |
+
"loss": 0.6447,
|
| 847 |
+
"step": 1110
|
| 848 |
+
},
|
| 849 |
+
{
|
| 850 |
+
"epoch": 0.143350825547165,
|
| 851 |
+
"grad_norm": 0.6887441277503967,
|
| 852 |
+
"learning_rate": 4.774909323661191e-06,
|
| 853 |
+
"loss": 0.6325,
|
| 854 |
+
"step": 1120
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 0.14463074363240752,
|
| 858 |
+
"grad_norm": 1.0218420028686523,
|
| 859 |
+
"learning_rate": 4.817580541924473e-06,
|
| 860 |
+
"loss": 0.6308,
|
| 861 |
+
"step": 1130
|
| 862 |
+
},
|
| 863 |
+
{
|
| 864 |
+
"epoch": 0.14591066171765008,
|
| 865 |
+
"grad_norm": 0.7663633823394775,
|
| 866 |
+
"learning_rate": 4.8602517601877535e-06,
|
| 867 |
+
"loss": 0.6227,
|
| 868 |
+
"step": 1140
|
| 869 |
+
},
|
| 870 |
+
{
|
| 871 |
+
"epoch": 0.1471905798028926,
|
| 872 |
+
"grad_norm": 1.5115224123001099,
|
| 873 |
+
"learning_rate": 4.902922978451035e-06,
|
| 874 |
+
"loss": 0.6315,
|
| 875 |
+
"step": 1150
|
| 876 |
+
},
|
| 877 |
+
{
|
| 878 |
+
"epoch": 0.14847049788813516,
|
| 879 |
+
"grad_norm": 0.8011078238487244,
|
| 880 |
+
"learning_rate": 4.945594196714317e-06,
|
| 881 |
+
"loss": 0.6428,
|
| 882 |
+
"step": 1160
|
| 883 |
+
},
|
| 884 |
+
{
|
| 885 |
+
"epoch": 0.1497504159733777,
|
| 886 |
+
"grad_norm": 1.3246116638183594,
|
| 887 |
+
"learning_rate": 4.9882654149775974e-06,
|
| 888 |
+
"loss": 0.6397,
|
| 889 |
+
"step": 1170
|
| 890 |
+
},
|
| 891 |
+
{
|
| 892 |
+
"epoch": 0.15103033405862024,
|
| 893 |
+
"grad_norm": 0.979498565196991,
|
| 894 |
+
"learning_rate": 5.030936633240879e-06,
|
| 895 |
+
"loss": 0.6308,
|
| 896 |
+
"step": 1180
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 0.1523102521438628,
|
| 900 |
+
"grad_norm": 0.8756135106086731,
|
| 901 |
+
"learning_rate": 5.073607851504161e-06,
|
| 902 |
+
"loss": 0.6373,
|
| 903 |
+
"step": 1190
|
| 904 |
+
},
|
| 905 |
+
{
|
| 906 |
+
"epoch": 0.15359017022910534,
|
| 907 |
+
"grad_norm": 0.845967710018158,
|
| 908 |
+
"learning_rate": 5.116279069767442e-06,
|
| 909 |
+
"loss": 0.6322,
|
| 910 |
+
"step": 1200
|
| 911 |
+
},
|
| 912 |
+
{
|
| 913 |
+
"epoch": 0.15359017022910534,
|
| 914 |
+
"eval_accuracy": 0.69931,
|
| 915 |
+
"eval_loss": 0.6224722862243652,
|
| 916 |
+
"eval_macro_f1": 0.6985627236986651,
|
| 917 |
+
"eval_precision": 0.6986454338713279,
|
| 918 |
+
"eval_recall": 0.6984991383538209,
|
| 919 |
+
"eval_runtime": 692.1526,
|
| 920 |
+
"eval_samples_per_second": 144.477,
|
| 921 |
+
"eval_steps_per_second": 2.258,
|
| 922 |
+
"step": 1200
|
| 923 |
+
},
|
| 924 |
+
{
|
| 925 |
+
"epoch": 0.15487008831434787,
|
| 926 |
+
"grad_norm": 2.0670483112335205,
|
| 927 |
+
"learning_rate": 1.032010243277849e-05,
|
| 928 |
+
"loss": 0.6386,
|
| 929 |
+
"step": 1210
|
| 930 |
+
},
|
| 931 |
+
{
|
| 932 |
+
"epoch": 0.15615000639959042,
|
| 933 |
+
"grad_norm": 2.482168436050415,
|
| 934 |
+
"learning_rate": 1.0405463081519421e-05,
|
| 935 |
+
"loss": 0.6281,
|
| 936 |
+
"step": 1220
|
| 937 |
+
},
|
| 938 |
+
{
|
| 939 |
+
"epoch": 0.15742992448483298,
|
| 940 |
+
"grad_norm": 1.4530706405639648,
|
| 941 |
+
"learning_rate": 1.049082373026035e-05,
|
| 942 |
+
"loss": 0.6277,
|
| 943 |
+
"step": 1230
|
| 944 |
+
},
|
| 945 |
+
{
|
| 946 |
+
"epoch": 0.1587098425700755,
|
| 947 |
+
"grad_norm": 2.9233577251434326,
|
| 948 |
+
"learning_rate": 1.0576184379001282e-05,
|
| 949 |
+
"loss": 0.6363,
|
| 950 |
+
"step": 1240
|
| 951 |
+
},
|
| 952 |
+
{
|
| 953 |
+
"epoch": 0.15998976065531806,
|
| 954 |
+
"grad_norm": 0.6975642442703247,
|
| 955 |
+
"learning_rate": 1.0661545027742212e-05,
|
| 956 |
+
"loss": 0.6344,
|
| 957 |
+
"step": 1250
|
| 958 |
+
},
|
| 959 |
+
{
|
| 960 |
+
"epoch": 0.1612696787405606,
|
| 961 |
+
"grad_norm": 2.3634731769561768,
|
| 962 |
+
"learning_rate": 1.0746905676483142e-05,
|
| 963 |
+
"loss": 0.6217,
|
| 964 |
+
"step": 1260
|
| 965 |
+
},
|
| 966 |
+
{
|
| 967 |
+
"epoch": 0.16254959682580314,
|
| 968 |
+
"grad_norm": 0.8812585473060608,
|
| 969 |
+
"learning_rate": 1.0832266325224072e-05,
|
| 970 |
+
"loss": 0.6311,
|
| 971 |
+
"step": 1270
|
| 972 |
+
},
|
| 973 |
+
{
|
| 974 |
+
"epoch": 0.1638295149110457,
|
| 975 |
+
"grad_norm": 1.8986798524856567,
|
| 976 |
+
"learning_rate": 1.0917626973965003e-05,
|
| 977 |
+
"loss": 0.6304,
|
| 978 |
+
"step": 1280
|
| 979 |
+
},
|
| 980 |
+
{
|
| 981 |
+
"epoch": 0.16510943299628825,
|
| 982 |
+
"grad_norm": 0.8174984455108643,
|
| 983 |
+
"learning_rate": 1.1002987622705935e-05,
|
| 984 |
+
"loss": 0.6176,
|
| 985 |
+
"step": 1290
|
| 986 |
+
},
|
| 987 |
+
{
|
| 988 |
+
"epoch": 0.16638935108153077,
|
| 989 |
+
"grad_norm": 1.0268642902374268,
|
| 990 |
+
"learning_rate": 1.1088348271446865e-05,
|
| 991 |
+
"loss": 0.6233,
|
| 992 |
+
"step": 1300
|
| 993 |
+
},
|
| 994 |
+
{
|
| 995 |
+
"epoch": 0.16766926916677333,
|
| 996 |
+
"grad_norm": 1.047821044921875,
|
| 997 |
+
"learning_rate": 1.1173708920187793e-05,
|
| 998 |
+
"loss": 0.6127,
|
| 999 |
+
"step": 1310
|
| 1000 |
+
},
|
| 1001 |
+
{
|
| 1002 |
+
"epoch": 0.16894918725201588,
|
| 1003 |
+
"grad_norm": 0.9875057935714722,
|
| 1004 |
+
"learning_rate": 1.1259069568928724e-05,
|
| 1005 |
+
"loss": 0.6211,
|
| 1006 |
+
"step": 1320
|
| 1007 |
+
},
|
| 1008 |
+
{
|
| 1009 |
+
"epoch": 0.1702291053372584,
|
| 1010 |
+
"grad_norm": 1.0082708597183228,
|
| 1011 |
+
"learning_rate": 1.1344430217669656e-05,
|
| 1012 |
+
"loss": 0.6202,
|
| 1013 |
+
"step": 1330
|
| 1014 |
+
},
|
| 1015 |
+
{
|
| 1016 |
+
"epoch": 0.17150902342250096,
|
| 1017 |
+
"grad_norm": 0.7229795455932617,
|
| 1018 |
+
"learning_rate": 1.1429790866410586e-05,
|
| 1019 |
+
"loss": 0.6071,
|
| 1020 |
+
"step": 1340
|
| 1021 |
+
},
|
| 1022 |
+
{
|
| 1023 |
+
"epoch": 0.17278894150774352,
|
| 1024 |
+
"grad_norm": 0.7687681317329407,
|
| 1025 |
+
"learning_rate": 1.1515151515151517e-05,
|
| 1026 |
+
"loss": 0.6218,
|
| 1027 |
+
"step": 1350
|
| 1028 |
+
},
|
| 1029 |
+
{
|
| 1030 |
+
"epoch": 0.17406885959298604,
|
| 1031 |
+
"grad_norm": 1.1643378734588623,
|
| 1032 |
+
"learning_rate": 1.1600512163892445e-05,
|
| 1033 |
+
"loss": 0.6208,
|
| 1034 |
+
"step": 1360
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 0.1753487776782286,
|
| 1038 |
+
"grad_norm": 0.7338509559631348,
|
| 1039 |
+
"learning_rate": 1.1685872812633377e-05,
|
| 1040 |
+
"loss": 0.6105,
|
| 1041 |
+
"step": 1370
|
| 1042 |
+
},
|
| 1043 |
+
{
|
| 1044 |
+
"epoch": 0.17662869576347115,
|
| 1045 |
+
"grad_norm": 0.7444656491279602,
|
| 1046 |
+
"learning_rate": 1.1771233461374307e-05,
|
| 1047 |
+
"loss": 0.6183,
|
| 1048 |
+
"step": 1380
|
| 1049 |
+
},
|
| 1050 |
+
{
|
| 1051 |
+
"epoch": 0.17790861384871368,
|
| 1052 |
+
"grad_norm": 1.1902283430099487,
|
| 1053 |
+
"learning_rate": 1.1856594110115238e-05,
|
| 1054 |
+
"loss": 0.6209,
|
| 1055 |
+
"step": 1390
|
| 1056 |
+
},
|
| 1057 |
+
{
|
| 1058 |
+
"epoch": 0.17918853193395623,
|
| 1059 |
+
"grad_norm": 1.0872228145599365,
|
| 1060 |
+
"learning_rate": 1.194195475885617e-05,
|
| 1061 |
+
"loss": 0.6172,
|
| 1062 |
+
"step": 1400
|
| 1063 |
+
},
|
| 1064 |
+
{
|
| 1065 |
+
"epoch": 0.18046845001919878,
|
| 1066 |
+
"grad_norm": 1.8439041376113892,
|
| 1067 |
+
"learning_rate": 1.2027315407597098e-05,
|
| 1068 |
+
"loss": 0.6086,
|
| 1069 |
+
"step": 1410
|
| 1070 |
+
},
|
| 1071 |
+
{
|
| 1072 |
+
"epoch": 0.1817483681044413,
|
| 1073 |
+
"grad_norm": 0.8699899911880493,
|
| 1074 |
+
"learning_rate": 1.211267605633803e-05,
|
| 1075 |
+
"loss": 0.6247,
|
| 1076 |
+
"step": 1420
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 0.18302828618968386,
|
| 1080 |
+
"grad_norm": 1.1468650102615356,
|
| 1081 |
+
"learning_rate": 1.2198036705078959e-05,
|
| 1082 |
+
"loss": 0.6114,
|
| 1083 |
+
"step": 1430
|
| 1084 |
+
},
|
| 1085 |
+
{
|
| 1086 |
+
"epoch": 0.18430820427492642,
|
| 1087 |
+
"grad_norm": 1.0204615592956543,
|
| 1088 |
+
"learning_rate": 1.228339735381989e-05,
|
| 1089 |
+
"loss": 0.6015,
|
| 1090 |
+
"step": 1440
|
| 1091 |
+
},
|
| 1092 |
+
{
|
| 1093 |
+
"epoch": 0.18558812236016894,
|
| 1094 |
+
"grad_norm": 2.660583257675171,
|
| 1095 |
+
"learning_rate": 1.236875800256082e-05,
|
| 1096 |
+
"loss": 0.6002,
|
| 1097 |
+
"step": 1450
|
| 1098 |
+
},
|
| 1099 |
+
{
|
| 1100 |
+
"epoch": 0.1868680404454115,
|
| 1101 |
+
"grad_norm": 1.1916371583938599,
|
| 1102 |
+
"learning_rate": 1.2454118651301752e-05,
|
| 1103 |
+
"loss": 0.6076,
|
| 1104 |
+
"step": 1460
|
| 1105 |
+
},
|
| 1106 |
+
{
|
| 1107 |
+
"epoch": 0.18814795853065402,
|
| 1108 |
+
"grad_norm": 0.8097987174987793,
|
| 1109 |
+
"learning_rate": 1.253947930004268e-05,
|
| 1110 |
+
"loss": 0.6126,
|
| 1111 |
+
"step": 1470
|
| 1112 |
+
},
|
| 1113 |
+
{
|
| 1114 |
+
"epoch": 0.18942787661589658,
|
| 1115 |
+
"grad_norm": 1.978747010231018,
|
| 1116 |
+
"learning_rate": 1.2624839948783612e-05,
|
| 1117 |
+
"loss": 0.5962,
|
| 1118 |
+
"step": 1480
|
| 1119 |
+
},
|
| 1120 |
+
{
|
| 1121 |
+
"epoch": 0.19070779470113913,
|
| 1122 |
+
"grad_norm": 1.0975425243377686,
|
| 1123 |
+
"learning_rate": 1.2710200597524543e-05,
|
| 1124 |
+
"loss": 0.587,
|
| 1125 |
+
"step": 1490
|
| 1126 |
+
},
|
| 1127 |
+
{
|
| 1128 |
+
"epoch": 0.19198771278638166,
|
| 1129 |
+
"grad_norm": 2.150351047515869,
|
| 1130 |
+
"learning_rate": 1.2795561246265473e-05,
|
| 1131 |
+
"loss": 0.5962,
|
| 1132 |
+
"step": 1500
|
| 1133 |
+
},
|
| 1134 |
+
{
|
| 1135 |
+
"epoch": 0.19198771278638166,
|
| 1136 |
+
"eval_accuracy": 0.72526,
|
| 1137 |
+
"eval_loss": 0.5823243856430054,
|
| 1138 |
+
"eval_macro_f1": 0.7252418839884757,
|
| 1139 |
+
"eval_precision": 0.7275813809678697,
|
| 1140 |
+
"eval_recall": 0.7271797416587706,
|
| 1141 |
+
"eval_runtime": 708.4119,
|
| 1142 |
+
"eval_samples_per_second": 141.161,
|
| 1143 |
+
"eval_steps_per_second": 1.104,
|
| 1144 |
+
"step": 1500
|
| 1145 |
+
},
|
| 1146 |
+
{
|
| 1147 |
+
"epoch": 0.1932676308716242,
|
| 1148 |
+
"grad_norm": 1.8355706930160522,
|
| 1149 |
+
"learning_rate": 1.2880921895006405e-05,
|
| 1150 |
+
"loss": 0.6133,
|
| 1151 |
+
"step": 1510
|
| 1152 |
+
},
|
| 1153 |
+
{
|
| 1154 |
+
"epoch": 0.19454754895686677,
|
| 1155 |
+
"grad_norm": 1.4349753856658936,
|
| 1156 |
+
"learning_rate": 1.2966282543747333e-05,
|
| 1157 |
+
"loss": 0.6062,
|
| 1158 |
+
"step": 1520
|
| 1159 |
+
},
|
| 1160 |
+
{
|
| 1161 |
+
"epoch": 0.1958274670421093,
|
| 1162 |
+
"grad_norm": 1.8322596549987793,
|
| 1163 |
+
"learning_rate": 1.3051643192488264e-05,
|
| 1164 |
+
"loss": 0.5982,
|
| 1165 |
+
"step": 1530
|
| 1166 |
+
},
|
| 1167 |
+
{
|
| 1168 |
+
"epoch": 0.19710738512735185,
|
| 1169 |
+
"grad_norm": 1.509338617324829,
|
| 1170 |
+
"learning_rate": 1.3137003841229194e-05,
|
| 1171 |
+
"loss": 0.5961,
|
| 1172 |
+
"step": 1540
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 0.1983873032125944,
|
| 1176 |
+
"grad_norm": 0.9844911098480225,
|
| 1177 |
+
"learning_rate": 1.3222364489970126e-05,
|
| 1178 |
+
"loss": 0.583,
|
| 1179 |
+
"step": 1550
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"epoch": 0.19966722129783693,
|
| 1183 |
+
"grad_norm": 1.035900354385376,
|
| 1184 |
+
"learning_rate": 1.3307725138711055e-05,
|
| 1185 |
+
"loss": 0.598,
|
| 1186 |
+
"step": 1560
|
| 1187 |
+
},
|
| 1188 |
+
{
|
| 1189 |
+
"epoch": 0.20094713938307948,
|
| 1190 |
+
"grad_norm": 0.7891924977302551,
|
| 1191 |
+
"learning_rate": 1.3393085787451985e-05,
|
| 1192 |
+
"loss": 0.6064,
|
| 1193 |
+
"step": 1570
|
| 1194 |
+
},
|
| 1195 |
+
{
|
| 1196 |
+
"epoch": 0.20222705746832204,
|
| 1197 |
+
"grad_norm": 1.4326112270355225,
|
| 1198 |
+
"learning_rate": 1.3478446436192915e-05,
|
| 1199 |
+
"loss": 0.5954,
|
| 1200 |
+
"step": 1580
|
| 1201 |
+
},
|
| 1202 |
+
{
|
| 1203 |
+
"epoch": 0.20350697555356456,
|
| 1204 |
+
"grad_norm": 0.9054266214370728,
|
| 1205 |
+
"learning_rate": 1.3563807084933847e-05,
|
| 1206 |
+
"loss": 0.59,
|
| 1207 |
+
"step": 1590
|
| 1208 |
+
},
|
| 1209 |
+
{
|
| 1210 |
+
"epoch": 0.20478689363880712,
|
| 1211 |
+
"grad_norm": 1.0216469764709473,
|
| 1212 |
+
"learning_rate": 1.3649167733674778e-05,
|
| 1213 |
+
"loss": 0.5845,
|
| 1214 |
+
"step": 1600
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 0.20606681172404967,
|
| 1218 |
+
"grad_norm": 1.0591177940368652,
|
| 1219 |
+
"learning_rate": 1.3734528382415708e-05,
|
| 1220 |
+
"loss": 0.599,
|
| 1221 |
+
"step": 1610
|
| 1222 |
+
},
|
| 1223 |
+
{
|
| 1224 |
+
"epoch": 0.2073467298092922,
|
| 1225 |
+
"grad_norm": 0.7826931476593018,
|
| 1226 |
+
"learning_rate": 1.3819889031156638e-05,
|
| 1227 |
+
"loss": 0.5898,
|
| 1228 |
+
"step": 1620
|
| 1229 |
+
},
|
| 1230 |
+
{
|
| 1231 |
+
"epoch": 0.20862664789453475,
|
| 1232 |
+
"grad_norm": 0.755007803440094,
|
| 1233 |
+
"learning_rate": 1.3905249679897568e-05,
|
| 1234 |
+
"loss": 0.5769,
|
| 1235 |
+
"step": 1630
|
| 1236 |
+
},
|
| 1237 |
+
{
|
| 1238 |
+
"epoch": 0.2099065659797773,
|
| 1239 |
+
"grad_norm": 1.3526090383529663,
|
| 1240 |
+
"learning_rate": 1.3990610328638499e-05,
|
| 1241 |
+
"loss": 0.5819,
|
| 1242 |
+
"step": 1640
|
| 1243 |
+
},
|
| 1244 |
+
{
|
| 1245 |
+
"epoch": 0.21118648406501983,
|
| 1246 |
+
"grad_norm": 0.7078200578689575,
|
| 1247 |
+
"learning_rate": 1.4075970977379429e-05,
|
| 1248 |
+
"loss": 0.5884,
|
| 1249 |
+
"step": 1650
|
| 1250 |
+
},
|
| 1251 |
+
{
|
| 1252 |
+
"epoch": 0.21246640215026238,
|
| 1253 |
+
"grad_norm": 0.80736243724823,
|
| 1254 |
+
"learning_rate": 1.416133162612036e-05,
|
| 1255 |
+
"loss": 0.5991,
|
| 1256 |
+
"step": 1660
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 0.21374632023550494,
|
| 1260 |
+
"grad_norm": 0.6882274150848389,
|
| 1261 |
+
"learning_rate": 1.4246692274861289e-05,
|
| 1262 |
+
"loss": 0.5825,
|
| 1263 |
+
"step": 1670
|
| 1264 |
+
},
|
| 1265 |
+
{
|
| 1266 |
+
"epoch": 0.21502623832074746,
|
| 1267 |
+
"grad_norm": 0.7223654985427856,
|
| 1268 |
+
"learning_rate": 1.433205292360222e-05,
|
| 1269 |
+
"loss": 0.5689,
|
| 1270 |
+
"step": 1680
|
| 1271 |
+
},
|
| 1272 |
+
{
|
| 1273 |
+
"epoch": 0.21630615640599002,
|
| 1274 |
+
"grad_norm": 0.6446322798728943,
|
| 1275 |
+
"learning_rate": 1.4417413572343152e-05,
|
| 1276 |
+
"loss": 0.5773,
|
| 1277 |
+
"step": 1690
|
| 1278 |
+
},
|
| 1279 |
+
{
|
| 1280 |
+
"epoch": 0.21758607449123257,
|
| 1281 |
+
"grad_norm": 1.505007266998291,
|
| 1282 |
+
"learning_rate": 1.4502774221084081e-05,
|
| 1283 |
+
"loss": 0.5677,
|
| 1284 |
+
"step": 1700
|
| 1285 |
+
},
|
| 1286 |
+
{
|
| 1287 |
+
"epoch": 0.2188659925764751,
|
| 1288 |
+
"grad_norm": 1.1939284801483154,
|
| 1289 |
+
"learning_rate": 1.4588134869825013e-05,
|
| 1290 |
+
"loss": 0.5805,
|
| 1291 |
+
"step": 1710
|
| 1292 |
+
},
|
| 1293 |
+
{
|
| 1294 |
+
"epoch": 0.22014591066171765,
|
| 1295 |
+
"grad_norm": 2.848264217376709,
|
| 1296 |
+
"learning_rate": 1.4673495518565943e-05,
|
| 1297 |
+
"loss": 0.5712,
|
| 1298 |
+
"step": 1720
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 0.2214258287469602,
|
| 1302 |
+
"grad_norm": 1.2117749452590942,
|
| 1303 |
+
"learning_rate": 1.4758856167306873e-05,
|
| 1304 |
+
"loss": 0.582,
|
| 1305 |
+
"step": 1730
|
| 1306 |
+
},
|
| 1307 |
+
{
|
| 1308 |
+
"epoch": 0.22270574683220273,
|
| 1309 |
+
"grad_norm": 2.2297441959381104,
|
| 1310 |
+
"learning_rate": 1.4844216816047802e-05,
|
| 1311 |
+
"loss": 0.5751,
|
| 1312 |
+
"step": 1740
|
| 1313 |
+
},
|
| 1314 |
+
{
|
| 1315 |
+
"epoch": 0.2239856649174453,
|
| 1316 |
+
"grad_norm": 1.211978554725647,
|
| 1317 |
+
"learning_rate": 1.4929577464788734e-05,
|
| 1318 |
+
"loss": 0.5607,
|
| 1319 |
+
"step": 1750
|
| 1320 |
+
},
|
| 1321 |
+
{
|
| 1322 |
+
"epoch": 0.22526558300268784,
|
| 1323 |
+
"grad_norm": 0.8025451898574829,
|
| 1324 |
+
"learning_rate": 1.5014938113529664e-05,
|
| 1325 |
+
"loss": 0.5803,
|
| 1326 |
+
"step": 1760
|
| 1327 |
+
},
|
| 1328 |
+
{
|
| 1329 |
+
"epoch": 0.22654550108793037,
|
| 1330 |
+
"grad_norm": 1.2300900220870972,
|
| 1331 |
+
"learning_rate": 1.5100298762270595e-05,
|
| 1332 |
+
"loss": 0.5827,
|
| 1333 |
+
"step": 1770
|
| 1334 |
+
},
|
| 1335 |
+
{
|
| 1336 |
+
"epoch": 0.22782541917317292,
|
| 1337 |
+
"grad_norm": 1.0382864475250244,
|
| 1338 |
+
"learning_rate": 1.5185659411011523e-05,
|
| 1339 |
+
"loss": 0.5916,
|
| 1340 |
+
"step": 1780
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 0.22910533725841545,
|
| 1344 |
+
"grad_norm": 1.4876656532287598,
|
| 1345 |
+
"learning_rate": 1.5271020059752453e-05,
|
| 1346 |
+
"loss": 0.5536,
|
| 1347 |
+
"step": 1790
|
| 1348 |
+
},
|
| 1349 |
+
{
|
| 1350 |
+
"epoch": 0.230385255343658,
|
| 1351 |
+
"grad_norm": 1.1409252882003784,
|
| 1352 |
+
"learning_rate": 1.5356380708493386e-05,
|
| 1353 |
+
"loss": 0.5484,
|
| 1354 |
+
"step": 1800
|
| 1355 |
+
},
|
| 1356 |
+
{
|
| 1357 |
+
"epoch": 0.23166517342890056,
|
| 1358 |
+
"grad_norm": 0.7180280089378357,
|
| 1359 |
+
"learning_rate": 1.5441741357234316e-05,
|
| 1360 |
+
"loss": 0.5712,
|
| 1361 |
+
"step": 1810
|
| 1362 |
+
},
|
| 1363 |
+
{
|
| 1364 |
+
"epoch": 0.23294509151414308,
|
| 1365 |
+
"grad_norm": 1.5890289545059204,
|
| 1366 |
+
"learning_rate": 1.5527102005975246e-05,
|
| 1367 |
+
"loss": 0.5758,
|
| 1368 |
+
"step": 1820
|
| 1369 |
+
},
|
| 1370 |
+
{
|
| 1371 |
+
"epoch": 0.23422500959938564,
|
| 1372 |
+
"grad_norm": 0.6635769605636597,
|
| 1373 |
+
"learning_rate": 1.5612462654716176e-05,
|
| 1374 |
+
"loss": 0.5543,
|
| 1375 |
+
"step": 1830
|
| 1376 |
+
},
|
| 1377 |
+
{
|
| 1378 |
+
"epoch": 0.2355049276846282,
|
| 1379 |
+
"grad_norm": 0.8864602446556091,
|
| 1380 |
+
"learning_rate": 1.5697823303457106e-05,
|
| 1381 |
+
"loss": 0.5536,
|
| 1382 |
+
"step": 1840
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 0.23678484576987072,
|
| 1386 |
+
"grad_norm": 0.8372675776481628,
|
| 1387 |
+
"learning_rate": 1.578318395219804e-05,
|
| 1388 |
+
"loss": 0.5743,
|
| 1389 |
+
"step": 1850
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"epoch": 0.23806476385511327,
|
| 1393 |
+
"grad_norm": 0.7532811164855957,
|
| 1394 |
+
"learning_rate": 1.586854460093897e-05,
|
| 1395 |
+
"loss": 0.5593,
|
| 1396 |
+
"step": 1860
|
| 1397 |
+
},
|
| 1398 |
+
{
|
| 1399 |
+
"epoch": 0.23934468194035582,
|
| 1400 |
+
"grad_norm": 0.7231799960136414,
|
| 1401 |
+
"learning_rate": 1.59539052496799e-05,
|
| 1402 |
+
"loss": 0.5831,
|
| 1403 |
+
"step": 1870
|
| 1404 |
+
},
|
| 1405 |
+
{
|
| 1406 |
+
"epoch": 0.24062460002559835,
|
| 1407 |
+
"grad_norm": 1.124228596687317,
|
| 1408 |
+
"learning_rate": 1.603926589842083e-05,
|
| 1409 |
+
"loss": 0.5936,
|
| 1410 |
+
"step": 1880
|
| 1411 |
+
},
|
| 1412 |
+
{
|
| 1413 |
+
"epoch": 0.2419045181108409,
|
| 1414 |
+
"grad_norm": 1.0107067823410034,
|
| 1415 |
+
"learning_rate": 1.612462654716176e-05,
|
| 1416 |
+
"loss": 0.5598,
|
| 1417 |
+
"step": 1890
|
| 1418 |
+
},
|
| 1419 |
+
{
|
| 1420 |
+
"epoch": 0.24318443619608346,
|
| 1421 |
+
"grad_norm": 1.6117154359817505,
|
| 1422 |
+
"learning_rate": 1.620998719590269e-05,
|
| 1423 |
+
"loss": 0.5605,
|
| 1424 |
+
"step": 1900
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 0.24446435428132599,
|
| 1428 |
+
"grad_norm": 1.5564496517181396,
|
| 1429 |
+
"learning_rate": 1.629534784464362e-05,
|
| 1430 |
+
"loss": 0.5666,
|
| 1431 |
+
"step": 1910
|
| 1432 |
+
},
|
| 1433 |
+
{
|
| 1434 |
+
"epoch": 0.24574427236656854,
|
| 1435 |
+
"grad_norm": 0.8275917768478394,
|
| 1436 |
+
"learning_rate": 1.638070849338455e-05,
|
| 1437 |
+
"loss": 0.5597,
|
| 1438 |
+
"step": 1920
|
| 1439 |
+
},
|
| 1440 |
+
{
|
| 1441 |
+
"epoch": 0.2470241904518111,
|
| 1442 |
+
"grad_norm": 0.8956324458122253,
|
| 1443 |
+
"learning_rate": 1.646606914212548e-05,
|
| 1444 |
+
"loss": 0.5453,
|
| 1445 |
+
"step": 1930
|
| 1446 |
+
},
|
| 1447 |
+
{
|
| 1448 |
+
"epoch": 0.24830410853705362,
|
| 1449 |
+
"grad_norm": 1.9644063711166382,
|
| 1450 |
+
"learning_rate": 1.655142979086641e-05,
|
| 1451 |
+
"loss": 0.5732,
|
| 1452 |
+
"step": 1940
|
| 1453 |
+
},
|
| 1454 |
+
{
|
| 1455 |
+
"epoch": 0.24958402662229617,
|
| 1456 |
+
"grad_norm": 0.7495090365409851,
|
| 1457 |
+
"learning_rate": 1.663679043960734e-05,
|
| 1458 |
+
"loss": 0.5549,
|
| 1459 |
+
"step": 1950
|
| 1460 |
+
},
|
| 1461 |
+
{
|
| 1462 |
+
"epoch": 0.2508639447075387,
|
| 1463 |
+
"grad_norm": 0.9346863627433777,
|
| 1464 |
+
"learning_rate": 1.6722151088348274e-05,
|
| 1465 |
+
"loss": 0.5803,
|
| 1466 |
+
"step": 1960
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"epoch": 0.2521438627927813,
|
| 1470 |
+
"grad_norm": 0.5847868919372559,
|
| 1471 |
+
"learning_rate": 1.6807511737089204e-05,
|
| 1472 |
+
"loss": 0.5705,
|
| 1473 |
+
"step": 1970
|
| 1474 |
+
},
|
| 1475 |
+
{
|
| 1476 |
+
"epoch": 0.2534237808780238,
|
| 1477 |
+
"grad_norm": 1.5892306566238403,
|
| 1478 |
+
"learning_rate": 1.6892872385830134e-05,
|
| 1479 |
+
"loss": 0.5371,
|
| 1480 |
+
"step": 1980
|
| 1481 |
+
},
|
| 1482 |
+
{
|
| 1483 |
+
"epoch": 0.25470369896326633,
|
| 1484 |
+
"grad_norm": 0.8782947063446045,
|
| 1485 |
+
"learning_rate": 1.6978233034571063e-05,
|
| 1486 |
+
"loss": 0.565,
|
| 1487 |
+
"step": 1990
|
| 1488 |
+
},
|
| 1489 |
+
{
|
| 1490 |
+
"epoch": 0.2559836170485089,
|
| 1491 |
+
"grad_norm": 1.0143548250198364,
|
| 1492 |
+
"learning_rate": 1.7063593683311993e-05,
|
| 1493 |
+
"loss": 0.557,
|
| 1494 |
+
"step": 2000
|
| 1495 |
+
},
|
| 1496 |
+
{
|
| 1497 |
+
"epoch": 0.2559836170485089,
|
| 1498 |
+
"eval_accuracy": 0.74427,
|
| 1499 |
+
"eval_loss": 0.5308485627174377,
|
| 1500 |
+
"eval_macro_f1": 0.743516265066894,
|
| 1501 |
+
"eval_precision": 0.7538937627554241,
|
| 1502 |
+
"eval_recall": 0.7483595020974525,
|
| 1503 |
+
"eval_runtime": 708.2272,
|
| 1504 |
+
"eval_samples_per_second": 141.198,
|
| 1505 |
+
"eval_steps_per_second": 1.104,
|
| 1506 |
+
"step": 2000
|
| 1507 |
+
},
|
| 1508 |
+
{
|
| 1509 |
+
"epoch": 0.25726353513375144,
|
| 1510 |
+
"grad_norm": 0.9080437421798706,
|
| 1511 |
+
"learning_rate": 1.7148954332052926e-05,
|
| 1512 |
+
"loss": 0.5478,
|
| 1513 |
+
"step": 2010
|
| 1514 |
+
},
|
| 1515 |
+
{
|
| 1516 |
+
"epoch": 0.25854345321899397,
|
| 1517 |
+
"grad_norm": 1.1889078617095947,
|
| 1518 |
+
"learning_rate": 1.7234314980793856e-05,
|
| 1519 |
+
"loss": 0.5394,
|
| 1520 |
+
"step": 2020
|
| 1521 |
+
},
|
| 1522 |
+
{
|
| 1523 |
+
"epoch": 0.25982337130423655,
|
| 1524 |
+
"grad_norm": 1.7290714979171753,
|
| 1525 |
+
"learning_rate": 1.7319675629534786e-05,
|
| 1526 |
+
"loss": 0.5385,
|
| 1527 |
+
"step": 2030
|
| 1528 |
+
},
|
| 1529 |
+
{
|
| 1530 |
+
"epoch": 0.2611032893894791,
|
| 1531 |
+
"grad_norm": 0.9356957077980042,
|
| 1532 |
+
"learning_rate": 1.7405036278275716e-05,
|
| 1533 |
+
"loss": 0.5392,
|
| 1534 |
+
"step": 2040
|
| 1535 |
+
},
|
| 1536 |
+
{
|
| 1537 |
+
"epoch": 0.2623832074747216,
|
| 1538 |
+
"grad_norm": 0.765798032283783,
|
| 1539 |
+
"learning_rate": 1.7490396927016646e-05,
|
| 1540 |
+
"loss": 0.5442,
|
| 1541 |
+
"step": 2050
|
| 1542 |
+
},
|
| 1543 |
+
{
|
| 1544 |
+
"epoch": 0.2636631255599642,
|
| 1545 |
+
"grad_norm": 1.6606382131576538,
|
| 1546 |
+
"learning_rate": 1.7575757575757576e-05,
|
| 1547 |
+
"loss": 0.5331,
|
| 1548 |
+
"step": 2060
|
| 1549 |
+
},
|
| 1550 |
+
{
|
| 1551 |
+
"epoch": 0.2649430436452067,
|
| 1552 |
+
"grad_norm": 0.7928575277328491,
|
| 1553 |
+
"learning_rate": 1.766111822449851e-05,
|
| 1554 |
+
"loss": 0.5483,
|
| 1555 |
+
"step": 2070
|
| 1556 |
+
},
|
| 1557 |
+
{
|
| 1558 |
+
"epoch": 0.26622296173044924,
|
| 1559 |
+
"grad_norm": 2.2261486053466797,
|
| 1560 |
+
"learning_rate": 1.774647887323944e-05,
|
| 1561 |
+
"loss": 0.5536,
|
| 1562 |
+
"step": 2080
|
| 1563 |
+
},
|
| 1564 |
+
{
|
| 1565 |
+
"epoch": 0.2675028798156918,
|
| 1566 |
+
"grad_norm": 0.9795323610305786,
|
| 1567 |
+
"learning_rate": 1.783183952198037e-05,
|
| 1568 |
+
"loss": 0.5584,
|
| 1569 |
+
"step": 2090
|
| 1570 |
+
},
|
| 1571 |
+
{
|
| 1572 |
+
"epoch": 0.26878279790093434,
|
| 1573 |
+
"grad_norm": 2.373581886291504,
|
| 1574 |
+
"learning_rate": 1.7917200170721298e-05,
|
| 1575 |
+
"loss": 0.5475,
|
| 1576 |
+
"step": 2100
|
| 1577 |
+
},
|
| 1578 |
+
{
|
| 1579 |
+
"epoch": 0.27006271598617687,
|
| 1580 |
+
"grad_norm": 0.9658742547035217,
|
| 1581 |
+
"learning_rate": 1.8002560819462228e-05,
|
| 1582 |
+
"loss": 0.5577,
|
| 1583 |
+
"step": 2110
|
| 1584 |
+
},
|
| 1585 |
+
{
|
| 1586 |
+
"epoch": 0.27134263407141945,
|
| 1587 |
+
"grad_norm": 0.8572520017623901,
|
| 1588 |
+
"learning_rate": 1.808792146820316e-05,
|
| 1589 |
+
"loss": 0.5301,
|
| 1590 |
+
"step": 2120
|
| 1591 |
+
},
|
| 1592 |
+
{
|
| 1593 |
+
"epoch": 0.272622552156662,
|
| 1594 |
+
"grad_norm": 0.7225477695465088,
|
| 1595 |
+
"learning_rate": 1.817328211694409e-05,
|
| 1596 |
+
"loss": 0.5626,
|
| 1597 |
+
"step": 2130
|
| 1598 |
+
},
|
| 1599 |
+
{
|
| 1600 |
+
"epoch": 0.2739024702419045,
|
| 1601 |
+
"grad_norm": 1.4276026487350464,
|
| 1602 |
+
"learning_rate": 1.825864276568502e-05,
|
| 1603 |
+
"loss": 0.5208,
|
| 1604 |
+
"step": 2140
|
| 1605 |
+
},
|
| 1606 |
+
{
|
| 1607 |
+
"epoch": 0.2751823883271471,
|
| 1608 |
+
"grad_norm": 1.7469464540481567,
|
| 1609 |
+
"learning_rate": 1.834400341442595e-05,
|
| 1610 |
+
"loss": 0.5669,
|
| 1611 |
+
"step": 2150
|
| 1612 |
+
},
|
| 1613 |
+
{
|
| 1614 |
+
"epoch": 0.2764623064123896,
|
| 1615 |
+
"grad_norm": 1.250931978225708,
|
| 1616 |
+
"learning_rate": 1.842936406316688e-05,
|
| 1617 |
+
"loss": 0.5577,
|
| 1618 |
+
"step": 2160
|
| 1619 |
+
},
|
| 1620 |
+
{
|
| 1621 |
+
"epoch": 0.27774222449763214,
|
| 1622 |
+
"grad_norm": 1.011313557624817,
|
| 1623 |
+
"learning_rate": 1.8514724711907814e-05,
|
| 1624 |
+
"loss": 0.5494,
|
| 1625 |
+
"step": 2170
|
| 1626 |
+
},
|
| 1627 |
+
{
|
| 1628 |
+
"epoch": 0.2790221425828747,
|
| 1629 |
+
"grad_norm": 0.7606518268585205,
|
| 1630 |
+
"learning_rate": 1.8600085360648744e-05,
|
| 1631 |
+
"loss": 0.5391,
|
| 1632 |
+
"step": 2180
|
| 1633 |
+
},
|
| 1634 |
+
{
|
| 1635 |
+
"epoch": 0.28030206066811725,
|
| 1636 |
+
"grad_norm": 0.8173381090164185,
|
| 1637 |
+
"learning_rate": 1.8685446009389673e-05,
|
| 1638 |
+
"loss": 0.5262,
|
| 1639 |
+
"step": 2190
|
| 1640 |
+
},
|
| 1641 |
+
{
|
| 1642 |
+
"epoch": 0.2815819787533598,
|
| 1643 |
+
"grad_norm": 0.7178583145141602,
|
| 1644 |
+
"learning_rate": 1.8770806658130603e-05,
|
| 1645 |
+
"loss": 0.5182,
|
| 1646 |
+
"step": 2200
|
| 1647 |
+
},
|
| 1648 |
+
{
|
| 1649 |
+
"epoch": 0.28286189683860236,
|
| 1650 |
+
"grad_norm": 0.6852149367332458,
|
| 1651 |
+
"learning_rate": 1.8856167306871533e-05,
|
| 1652 |
+
"loss": 0.528,
|
| 1653 |
+
"step": 2210
|
| 1654 |
+
},
|
| 1655 |
+
{
|
| 1656 |
+
"epoch": 0.2841418149238449,
|
| 1657 |
+
"grad_norm": 1.9807002544403076,
|
| 1658 |
+
"learning_rate": 1.8941527955612463e-05,
|
| 1659 |
+
"loss": 0.5315,
|
| 1660 |
+
"step": 2220
|
| 1661 |
+
},
|
| 1662 |
+
{
|
| 1663 |
+
"epoch": 0.2854217330090874,
|
| 1664 |
+
"grad_norm": 0.7938176989555359,
|
| 1665 |
+
"learning_rate": 1.9026888604353396e-05,
|
| 1666 |
+
"loss": 0.5324,
|
| 1667 |
+
"step": 2230
|
| 1668 |
+
},
|
| 1669 |
+
{
|
| 1670 |
+
"epoch": 0.28670165109433,
|
| 1671 |
+
"grad_norm": 0.8181535005569458,
|
| 1672 |
+
"learning_rate": 1.9112249253094326e-05,
|
| 1673 |
+
"loss": 0.51,
|
| 1674 |
+
"step": 2240
|
| 1675 |
+
},
|
| 1676 |
+
{
|
| 1677 |
+
"epoch": 0.2879815691795725,
|
| 1678 |
+
"grad_norm": 0.8722071051597595,
|
| 1679 |
+
"learning_rate": 1.9197609901835256e-05,
|
| 1680 |
+
"loss": 0.5373,
|
| 1681 |
+
"step": 2250
|
| 1682 |
+
},
|
| 1683 |
+
{
|
| 1684 |
+
"epoch": 0.28926148726481504,
|
| 1685 |
+
"grad_norm": 1.371565818786621,
|
| 1686 |
+
"learning_rate": 1.9282970550576186e-05,
|
| 1687 |
+
"loss": 0.5227,
|
| 1688 |
+
"step": 2260
|
| 1689 |
+
},
|
| 1690 |
+
{
|
| 1691 |
+
"epoch": 0.29054140535005757,
|
| 1692 |
+
"grad_norm": 0.6769284605979919,
|
| 1693 |
+
"learning_rate": 1.9368331199317115e-05,
|
| 1694 |
+
"loss": 0.5421,
|
| 1695 |
+
"step": 2270
|
| 1696 |
+
},
|
| 1697 |
+
{
|
| 1698 |
+
"epoch": 0.29182132343530015,
|
| 1699 |
+
"grad_norm": 2.1016697883605957,
|
| 1700 |
+
"learning_rate": 1.945369184805805e-05,
|
| 1701 |
+
"loss": 0.5299,
|
| 1702 |
+
"step": 2280
|
| 1703 |
+
},
|
| 1704 |
+
{
|
| 1705 |
+
"epoch": 0.2931012415205427,
|
| 1706 |
+
"grad_norm": 1.7011477947235107,
|
| 1707 |
+
"learning_rate": 1.953905249679898e-05,
|
| 1708 |
+
"loss": 0.5403,
|
| 1709 |
+
"step": 2290
|
| 1710 |
+
},
|
| 1711 |
+
{
|
| 1712 |
+
"epoch": 0.2943811596057852,
|
| 1713 |
+
"grad_norm": 1.0961430072784424,
|
| 1714 |
+
"learning_rate": 1.962441314553991e-05,
|
| 1715 |
+
"loss": 0.534,
|
| 1716 |
+
"step": 2300
|
| 1717 |
+
},
|
| 1718 |
+
{
|
| 1719 |
+
"epoch": 0.2956610776910278,
|
| 1720 |
+
"grad_norm": 0.8536306023597717,
|
| 1721 |
+
"learning_rate": 1.9709773794280838e-05,
|
| 1722 |
+
"loss": 0.5562,
|
| 1723 |
+
"step": 2310
|
| 1724 |
+
},
|
| 1725 |
+
{
|
| 1726 |
+
"epoch": 0.2969409957762703,
|
| 1727 |
+
"grad_norm": 1.046201467514038,
|
| 1728 |
+
"learning_rate": 1.9795134443021768e-05,
|
| 1729 |
+
"loss": 0.5418,
|
| 1730 |
+
"step": 2320
|
| 1731 |
+
},
|
| 1732 |
+
{
|
| 1733 |
+
"epoch": 0.29822091386151284,
|
| 1734 |
+
"grad_norm": 1.38905930519104,
|
| 1735 |
+
"learning_rate": 1.9880495091762698e-05,
|
| 1736 |
+
"loss": 0.5557,
|
| 1737 |
+
"step": 2330
|
| 1738 |
+
},
|
| 1739 |
+
{
|
| 1740 |
+
"epoch": 0.2995008319467554,
|
| 1741 |
+
"grad_norm": 2.44215989112854,
|
| 1742 |
+
"learning_rate": 1.996585574050363e-05,
|
| 1743 |
+
"loss": 0.5285,
|
| 1744 |
+
"step": 2340
|
| 1745 |
+
},
|
| 1746 |
+
{
|
| 1747 |
+
"epoch": 0.30078075003199795,
|
| 1748 |
+
"grad_norm": 1.4736480712890625,
|
| 1749 |
+
"learning_rate": 1.9994311717861208e-05,
|
| 1750 |
+
"loss": 0.513,
|
| 1751 |
+
"step": 2350
|
| 1752 |
+
},
|
| 1753 |
+
{
|
| 1754 |
+
"epoch": 0.30206066811724047,
|
| 1755 |
+
"grad_norm": 0.7968178987503052,
|
| 1756 |
+
"learning_rate": 1.9984831247629883e-05,
|
| 1757 |
+
"loss": 0.5249,
|
| 1758 |
+
"step": 2360
|
| 1759 |
+
},
|
| 1760 |
+
{
|
| 1761 |
+
"epoch": 0.30334058620248305,
|
| 1762 |
+
"grad_norm": 1.6530416011810303,
|
| 1763 |
+
"learning_rate": 1.997535077739856e-05,
|
| 1764 |
+
"loss": 0.5329,
|
| 1765 |
+
"step": 2370
|
| 1766 |
+
},
|
| 1767 |
+
{
|
| 1768 |
+
"epoch": 0.3046205042877256,
|
| 1769 |
+
"grad_norm": 1.0949759483337402,
|
| 1770 |
+
"learning_rate": 1.9965870307167238e-05,
|
| 1771 |
+
"loss": 0.5164,
|
| 1772 |
+
"step": 2380
|
| 1773 |
+
},
|
| 1774 |
+
{
|
| 1775 |
+
"epoch": 0.3059004223729681,
|
| 1776 |
+
"grad_norm": 1.2724019289016724,
|
| 1777 |
+
"learning_rate": 1.9956389836935916e-05,
|
| 1778 |
+
"loss": 0.5293,
|
| 1779 |
+
"step": 2390
|
| 1780 |
+
},
|
| 1781 |
+
{
|
| 1782 |
+
"epoch": 0.3071803404582107,
|
| 1783 |
+
"grad_norm": 1.8785524368286133,
|
| 1784 |
+
"learning_rate": 1.994690936670459e-05,
|
| 1785 |
+
"loss": 0.5286,
|
| 1786 |
+
"step": 2400
|
| 1787 |
+
},
|
| 1788 |
+
{
|
| 1789 |
+
"epoch": 0.3084602585434532,
|
| 1790 |
+
"grad_norm": 2.1517200469970703,
|
| 1791 |
+
"learning_rate": 1.9937428896473267e-05,
|
| 1792 |
+
"loss": 0.5273,
|
| 1793 |
+
"step": 2410
|
| 1794 |
+
},
|
| 1795 |
+
{
|
| 1796 |
+
"epoch": 0.30974017662869574,
|
| 1797 |
+
"grad_norm": 0.999515175819397,
|
| 1798 |
+
"learning_rate": 1.9927948426241942e-05,
|
| 1799 |
+
"loss": 0.5115,
|
| 1800 |
+
"step": 2420
|
| 1801 |
+
},
|
| 1802 |
+
{
|
| 1803 |
+
"epoch": 0.3110200947139383,
|
| 1804 |
+
"grad_norm": 2.214357852935791,
|
| 1805 |
+
"learning_rate": 1.991846795601062e-05,
|
| 1806 |
+
"loss": 0.5145,
|
| 1807 |
+
"step": 2430
|
| 1808 |
+
},
|
| 1809 |
+
{
|
| 1810 |
+
"epoch": 0.31230001279918085,
|
| 1811 |
+
"grad_norm": 1.8795071840286255,
|
| 1812 |
+
"learning_rate": 1.9908987485779297e-05,
|
| 1813 |
+
"loss": 0.509,
|
| 1814 |
+
"step": 2440
|
| 1815 |
+
},
|
| 1816 |
+
{
|
| 1817 |
+
"epoch": 0.3135799308844234,
|
| 1818 |
+
"grad_norm": 0.9374184608459473,
|
| 1819 |
+
"learning_rate": 1.989950701554797e-05,
|
| 1820 |
+
"loss": 0.5112,
|
| 1821 |
+
"step": 2450
|
| 1822 |
+
},
|
| 1823 |
+
{
|
| 1824 |
+
"epoch": 0.31485984896966596,
|
| 1825 |
+
"grad_norm": 0.7281107306480408,
|
| 1826 |
+
"learning_rate": 1.989002654531665e-05,
|
| 1827 |
+
"loss": 0.5029,
|
| 1828 |
+
"step": 2460
|
| 1829 |
+
},
|
| 1830 |
+
{
|
| 1831 |
+
"epoch": 0.3161397670549085,
|
| 1832 |
+
"grad_norm": 2.6096737384796143,
|
| 1833 |
+
"learning_rate": 1.9880546075085327e-05,
|
| 1834 |
+
"loss": 0.5302,
|
| 1835 |
+
"step": 2470
|
| 1836 |
+
},
|
| 1837 |
+
{
|
| 1838 |
+
"epoch": 0.317419685140151,
|
| 1839 |
+
"grad_norm": 0.7463436126708984,
|
| 1840 |
+
"learning_rate": 1.9871065604854004e-05,
|
| 1841 |
+
"loss": 0.5249,
|
| 1842 |
+
"step": 2480
|
| 1843 |
+
},
|
| 1844 |
+
{
|
| 1845 |
+
"epoch": 0.3186996032253936,
|
| 1846 |
+
"grad_norm": 0.7995405197143555,
|
| 1847 |
+
"learning_rate": 1.986158513462268e-05,
|
| 1848 |
+
"loss": 0.5171,
|
| 1849 |
+
"step": 2490
|
| 1850 |
+
},
|
| 1851 |
+
{
|
| 1852 |
+
"epoch": 0.3199795213106361,
|
| 1853 |
+
"grad_norm": 0.873103678226471,
|
| 1854 |
+
"learning_rate": 1.9852104664391356e-05,
|
| 1855 |
+
"loss": 0.5125,
|
| 1856 |
+
"step": 2500
|
| 1857 |
+
},
|
| 1858 |
+
{
|
| 1859 |
+
"epoch": 0.32125943939587864,
|
| 1860 |
+
"grad_norm": 1.5319750308990479,
|
| 1861 |
+
"learning_rate": 1.984262419416003e-05,
|
| 1862 |
+
"loss": 0.527,
|
| 1863 |
+
"step": 2510
|
| 1864 |
+
},
|
| 1865 |
+
{
|
| 1866 |
+
"epoch": 0.3225393574811212,
|
| 1867 |
+
"grad_norm": 1.717421531677246,
|
| 1868 |
+
"learning_rate": 1.983314372392871e-05,
|
| 1869 |
+
"loss": 0.5415,
|
| 1870 |
+
"step": 2520
|
| 1871 |
+
},
|
| 1872 |
+
{
|
| 1873 |
+
"epoch": 0.32381927556636375,
|
| 1874 |
+
"grad_norm": 0.7133586406707764,
|
| 1875 |
+
"learning_rate": 1.9823663253697386e-05,
|
| 1876 |
+
"loss": 0.5251,
|
| 1877 |
+
"step": 2530
|
| 1878 |
+
},
|
| 1879 |
+
{
|
| 1880 |
+
"epoch": 0.3250991936516063,
|
| 1881 |
+
"grad_norm": 0.7959030270576477,
|
| 1882 |
+
"learning_rate": 1.981418278346606e-05,
|
| 1883 |
+
"loss": 0.5127,
|
| 1884 |
+
"step": 2540
|
| 1885 |
+
},
|
| 1886 |
+
{
|
| 1887 |
+
"epoch": 0.32637911173684886,
|
| 1888 |
+
"grad_norm": 2.4905223846435547,
|
| 1889 |
+
"learning_rate": 1.9804702313234738e-05,
|
| 1890 |
+
"loss": 0.4944,
|
| 1891 |
+
"step": 2550
|
| 1892 |
+
},
|
| 1893 |
+
{
|
| 1894 |
+
"epoch": 0.3276590298220914,
|
| 1895 |
+
"grad_norm": 1.0039845705032349,
|
| 1896 |
+
"learning_rate": 1.9795221843003412e-05,
|
| 1897 |
+
"loss": 0.534,
|
| 1898 |
+
"step": 2560
|
| 1899 |
+
},
|
| 1900 |
+
{
|
| 1901 |
+
"epoch": 0.3289389479073339,
|
| 1902 |
+
"grad_norm": 1.2181288003921509,
|
| 1903 |
+
"learning_rate": 1.978574137277209e-05,
|
| 1904 |
+
"loss": 0.5198,
|
| 1905 |
+
"step": 2570
|
| 1906 |
+
},
|
| 1907 |
+
{
|
| 1908 |
+
"epoch": 0.3302188659925765,
|
| 1909 |
+
"grad_norm": 0.9166068434715271,
|
| 1910 |
+
"learning_rate": 1.9776260902540768e-05,
|
| 1911 |
+
"loss": 0.5373,
|
| 1912 |
+
"step": 2580
|
| 1913 |
+
},
|
| 1914 |
+
{
|
| 1915 |
+
"epoch": 0.331498784077819,
|
| 1916 |
+
"grad_norm": 0.938994288444519,
|
| 1917 |
+
"learning_rate": 1.9766780432309445e-05,
|
| 1918 |
+
"loss": 0.5179,
|
| 1919 |
+
"step": 2590
|
| 1920 |
+
},
|
| 1921 |
+
{
|
| 1922 |
+
"epoch": 0.33277870216306155,
|
| 1923 |
+
"grad_norm": 0.8827023506164551,
|
| 1924 |
+
"learning_rate": 1.975729996207812e-05,
|
| 1925 |
+
"loss": 0.5249,
|
| 1926 |
+
"step": 2600
|
| 1927 |
+
},
|
| 1928 |
+
{
|
| 1929 |
+
"epoch": 0.33405862024830413,
|
| 1930 |
+
"grad_norm": 0.8544501066207886,
|
| 1931 |
+
"learning_rate": 1.9747819491846797e-05,
|
| 1932 |
+
"loss": 0.5172,
|
| 1933 |
+
"step": 2610
|
| 1934 |
+
},
|
| 1935 |
+
{
|
| 1936 |
+
"epoch": 0.33533853833354665,
|
| 1937 |
+
"grad_norm": 0.7890387177467346,
|
| 1938 |
+
"learning_rate": 1.9738339021615475e-05,
|
| 1939 |
+
"loss": 0.5207,
|
| 1940 |
+
"step": 2620
|
| 1941 |
+
},
|
| 1942 |
+
{
|
| 1943 |
+
"epoch": 0.3366184564187892,
|
| 1944 |
+
"grad_norm": 0.7724694013595581,
|
| 1945 |
+
"learning_rate": 1.9728858551384153e-05,
|
| 1946 |
+
"loss": 0.5065,
|
| 1947 |
+
"step": 2630
|
| 1948 |
+
},
|
| 1949 |
+
{
|
| 1950 |
+
"epoch": 0.33789837450403176,
|
| 1951 |
+
"grad_norm": 0.7064672708511353,
|
| 1952 |
+
"learning_rate": 1.9719378081152827e-05,
|
| 1953 |
+
"loss": 0.4828,
|
| 1954 |
+
"step": 2640
|
| 1955 |
+
},
|
| 1956 |
+
{
|
| 1957 |
+
"epoch": 0.3391782925892743,
|
| 1958 |
+
"grad_norm": 0.9625025391578674,
|
| 1959 |
+
"learning_rate": 1.97098976109215e-05,
|
| 1960 |
+
"loss": 0.5108,
|
| 1961 |
+
"step": 2650
|
| 1962 |
+
},
|
| 1963 |
+
{
|
| 1964 |
+
"epoch": 0.3404582106745168,
|
| 1965 |
+
"grad_norm": 2.1604244709014893,
|
| 1966 |
+
"learning_rate": 1.970041714069018e-05,
|
| 1967 |
+
"loss": 0.5129,
|
| 1968 |
+
"step": 2660
|
| 1969 |
+
},
|
| 1970 |
+
{
|
| 1971 |
+
"epoch": 0.3417381287597594,
|
| 1972 |
+
"grad_norm": 0.8561355471611023,
|
| 1973 |
+
"learning_rate": 1.9690936670458856e-05,
|
| 1974 |
+
"loss": 0.521,
|
| 1975 |
+
"step": 2670
|
| 1976 |
+
},
|
| 1977 |
+
{
|
| 1978 |
+
"epoch": 0.3430180468450019,
|
| 1979 |
+
"grad_norm": 0.8122753500938416,
|
| 1980 |
+
"learning_rate": 1.9681456200227534e-05,
|
| 1981 |
+
"loss": 0.5039,
|
| 1982 |
+
"step": 2680
|
| 1983 |
+
},
|
| 1984 |
+
{
|
| 1985 |
+
"epoch": 0.34429796493024445,
|
| 1986 |
+
"grad_norm": 0.848838210105896,
|
| 1987 |
+
"learning_rate": 1.967197572999621e-05,
|
| 1988 |
+
"loss": 0.5559,
|
| 1989 |
+
"step": 2690
|
| 1990 |
+
},
|
| 1991 |
+
{
|
| 1992 |
+
"epoch": 0.34557788301548703,
|
| 1993 |
+
"grad_norm": 0.8424274325370789,
|
| 1994 |
+
"learning_rate": 1.9662495259764886e-05,
|
| 1995 |
+
"loss": 0.5299,
|
| 1996 |
+
"step": 2700
|
| 1997 |
+
},
|
| 1998 |
+
{
|
| 1999 |
+
"epoch": 0.34685780110072956,
|
| 2000 |
+
"grad_norm": 1.0475687980651855,
|
| 2001 |
+
"learning_rate": 1.9653014789533564e-05,
|
| 2002 |
+
"loss": 0.5003,
|
| 2003 |
+
"step": 2710
|
| 2004 |
+
},
|
| 2005 |
+
{
|
| 2006 |
+
"epoch": 0.3481377191859721,
|
| 2007 |
+
"grad_norm": 0.9720590710639954,
|
| 2008 |
+
"learning_rate": 1.964353431930224e-05,
|
| 2009 |
+
"loss": 0.4976,
|
| 2010 |
+
"step": 2720
|
| 2011 |
+
},
|
| 2012 |
+
{
|
| 2013 |
+
"epoch": 0.34941763727121466,
|
| 2014 |
+
"grad_norm": 1.1562892198562622,
|
| 2015 |
+
"learning_rate": 1.9634053849070916e-05,
|
| 2016 |
+
"loss": 0.4728,
|
| 2017 |
+
"step": 2730
|
| 2018 |
+
},
|
| 2019 |
+
{
|
| 2020 |
+
"epoch": 0.3506975553564572,
|
| 2021 |
+
"grad_norm": 1.2779773473739624,
|
| 2022 |
+
"learning_rate": 1.9624573378839593e-05,
|
| 2023 |
+
"loss": 0.4933,
|
| 2024 |
+
"step": 2740
|
| 2025 |
+
},
|
| 2026 |
+
{
|
| 2027 |
+
"epoch": 0.3519774734416997,
|
| 2028 |
+
"grad_norm": 1.047945499420166,
|
| 2029 |
+
"learning_rate": 1.9615092908608268e-05,
|
| 2030 |
+
"loss": 0.518,
|
| 2031 |
+
"step": 2750
|
| 2032 |
+
},
|
| 2033 |
+
{
|
| 2034 |
+
"epoch": 0.3532573915269423,
|
| 2035 |
+
"grad_norm": 0.7196894884109497,
|
| 2036 |
+
"learning_rate": 1.9605612438376945e-05,
|
| 2037 |
+
"loss": 0.5006,
|
| 2038 |
+
"step": 2760
|
| 2039 |
+
},
|
| 2040 |
+
{
|
| 2041 |
+
"epoch": 0.3545373096121848,
|
| 2042 |
+
"grad_norm": 0.9644359350204468,
|
| 2043 |
+
"learning_rate": 1.9596131968145623e-05,
|
| 2044 |
+
"loss": 0.5083,
|
| 2045 |
+
"step": 2770
|
| 2046 |
+
},
|
| 2047 |
+
{
|
| 2048 |
+
"epoch": 0.35581722769742735,
|
| 2049 |
+
"grad_norm": 0.9720232486724854,
|
| 2050 |
+
"learning_rate": 1.9586651497914297e-05,
|
| 2051 |
+
"loss": 0.5194,
|
| 2052 |
+
"step": 2780
|
| 2053 |
+
},
|
| 2054 |
+
{
|
| 2055 |
+
"epoch": 0.35709714578266993,
|
| 2056 |
+
"grad_norm": 0.8277772665023804,
|
| 2057 |
+
"learning_rate": 1.9577171027682975e-05,
|
| 2058 |
+
"loss": 0.5121,
|
| 2059 |
+
"step": 2790
|
| 2060 |
+
},
|
| 2061 |
+
{
|
| 2062 |
+
"epoch": 0.35837706386791246,
|
| 2063 |
+
"grad_norm": 0.7811515927314758,
|
| 2064 |
+
"learning_rate": 1.956769055745165e-05,
|
| 2065 |
+
"loss": 0.5106,
|
| 2066 |
+
"step": 2800
|
| 2067 |
+
},
|
| 2068 |
+
{
|
| 2069 |
+
"epoch": 0.359656981953155,
|
| 2070 |
+
"grad_norm": 0.7438198328018188,
|
| 2071 |
+
"learning_rate": 1.9558210087220327e-05,
|
| 2072 |
+
"loss": 0.5183,
|
| 2073 |
+
"step": 2810
|
| 2074 |
+
},
|
| 2075 |
+
{
|
| 2076 |
+
"epoch": 0.36093690003839757,
|
| 2077 |
+
"grad_norm": 1.1424624919891357,
|
| 2078 |
+
"learning_rate": 1.9548729616989005e-05,
|
| 2079 |
+
"loss": 0.4973,
|
| 2080 |
+
"step": 2820
|
| 2081 |
+
},
|
| 2082 |
+
{
|
| 2083 |
+
"epoch": 0.3622168181236401,
|
| 2084 |
+
"grad_norm": 1.0706868171691895,
|
| 2085 |
+
"learning_rate": 1.9539249146757682e-05,
|
| 2086 |
+
"loss": 0.5063,
|
| 2087 |
+
"step": 2830
|
| 2088 |
+
},
|
| 2089 |
+
{
|
| 2090 |
+
"epoch": 0.3634967362088826,
|
| 2091 |
+
"grad_norm": 1.1213792562484741,
|
| 2092 |
+
"learning_rate": 1.9529768676526357e-05,
|
| 2093 |
+
"loss": 0.4991,
|
| 2094 |
+
"step": 2840
|
| 2095 |
+
},
|
| 2096 |
+
{
|
| 2097 |
+
"epoch": 0.3647766542941252,
|
| 2098 |
+
"grad_norm": 1.0018831491470337,
|
| 2099 |
+
"learning_rate": 1.9520288206295034e-05,
|
| 2100 |
+
"loss": 0.5004,
|
| 2101 |
+
"step": 2850
|
| 2102 |
+
},
|
| 2103 |
+
{
|
| 2104 |
+
"epoch": 0.36605657237936773,
|
| 2105 |
+
"grad_norm": 0.8369762897491455,
|
| 2106 |
+
"learning_rate": 1.9510807736063712e-05,
|
| 2107 |
+
"loss": 0.5341,
|
| 2108 |
+
"step": 2860
|
| 2109 |
+
},
|
| 2110 |
+
{
|
| 2111 |
+
"epoch": 0.36733649046461025,
|
| 2112 |
+
"grad_norm": 0.8711100220680237,
|
| 2113 |
+
"learning_rate": 1.9501327265832386e-05,
|
| 2114 |
+
"loss": 0.4788,
|
| 2115 |
+
"step": 2870
|
| 2116 |
+
},
|
| 2117 |
+
{
|
| 2118 |
+
"epoch": 0.36861640854985284,
|
| 2119 |
+
"grad_norm": 1.0853824615478516,
|
| 2120 |
+
"learning_rate": 1.9491846795601064e-05,
|
| 2121 |
+
"loss": 0.5419,
|
| 2122 |
+
"step": 2880
|
| 2123 |
+
},
|
| 2124 |
+
{
|
| 2125 |
+
"epoch": 0.36989632663509536,
|
| 2126 |
+
"grad_norm": 1.4816603660583496,
|
| 2127 |
+
"learning_rate": 1.9482366325369738e-05,
|
| 2128 |
+
"loss": 0.5252,
|
| 2129 |
+
"step": 2890
|
| 2130 |
+
},
|
| 2131 |
+
{
|
| 2132 |
+
"epoch": 0.3711762447203379,
|
| 2133 |
+
"grad_norm": 1.3375904560089111,
|
| 2134 |
+
"learning_rate": 1.9472885855138416e-05,
|
| 2135 |
+
"loss": 0.5127,
|
| 2136 |
+
"step": 2900
|
| 2137 |
+
},
|
| 2138 |
+
{
|
| 2139 |
+
"epoch": 0.3724561628055804,
|
| 2140 |
+
"grad_norm": 0.7361944913864136,
|
| 2141 |
+
"learning_rate": 1.9463405384907093e-05,
|
| 2142 |
+
"loss": 0.5278,
|
| 2143 |
+
"step": 2910
|
| 2144 |
+
},
|
| 2145 |
+
{
|
| 2146 |
+
"epoch": 0.373736080890823,
|
| 2147 |
+
"grad_norm": 1.1056461334228516,
|
| 2148 |
+
"learning_rate": 1.945392491467577e-05,
|
| 2149 |
+
"loss": 0.5072,
|
| 2150 |
+
"step": 2920
|
| 2151 |
+
},
|
| 2152 |
+
{
|
| 2153 |
+
"epoch": 0.3750159989760655,
|
| 2154 |
+
"grad_norm": 0.921783447265625,
|
| 2155 |
+
"learning_rate": 1.9444444444444445e-05,
|
| 2156 |
+
"loss": 0.5161,
|
| 2157 |
+
"step": 2930
|
| 2158 |
+
},
|
| 2159 |
+
{
|
| 2160 |
+
"epoch": 0.37629591706130805,
|
| 2161 |
+
"grad_norm": 1.15033757686615,
|
| 2162 |
+
"learning_rate": 1.9434963974213123e-05,
|
| 2163 |
+
"loss": 0.5329,
|
| 2164 |
+
"step": 2940
|
| 2165 |
+
},
|
| 2166 |
+
{
|
| 2167 |
+
"epoch": 0.37757583514655063,
|
| 2168 |
+
"grad_norm": 1.3983162641525269,
|
| 2169 |
+
"learning_rate": 1.94254835039818e-05,
|
| 2170 |
+
"loss": 0.5034,
|
| 2171 |
+
"step": 2950
|
| 2172 |
+
},
|
| 2173 |
+
{
|
| 2174 |
+
"epoch": 0.37885575323179316,
|
| 2175 |
+
"grad_norm": 1.0419155359268188,
|
| 2176 |
+
"learning_rate": 1.9416003033750475e-05,
|
| 2177 |
+
"loss": 0.5281,
|
| 2178 |
+
"step": 2960
|
| 2179 |
+
},
|
| 2180 |
+
{
|
| 2181 |
+
"epoch": 0.3801356713170357,
|
| 2182 |
+
"grad_norm": 2.478982448577881,
|
| 2183 |
+
"learning_rate": 1.9406522563519153e-05,
|
| 2184 |
+
"loss": 0.4843,
|
| 2185 |
+
"step": 2970
|
| 2186 |
+
},
|
| 2187 |
+
{
|
| 2188 |
+
"epoch": 0.38141558940227827,
|
| 2189 |
+
"grad_norm": 0.7638375163078308,
|
| 2190 |
+
"learning_rate": 1.9397042093287827e-05,
|
| 2191 |
+
"loss": 0.5214,
|
| 2192 |
+
"step": 2980
|
| 2193 |
+
},
|
| 2194 |
+
{
|
| 2195 |
+
"epoch": 0.3826955074875208,
|
| 2196 |
+
"grad_norm": 2.18105411529541,
|
| 2197 |
+
"learning_rate": 1.9387561623056505e-05,
|
| 2198 |
+
"loss": 0.5373,
|
| 2199 |
+
"step": 2990
|
| 2200 |
+
},
|
| 2201 |
+
{
|
| 2202 |
+
"epoch": 0.3839754255727633,
|
| 2203 |
+
"grad_norm": 1.0936481952667236,
|
| 2204 |
+
"learning_rate": 1.9378081152825182e-05,
|
| 2205 |
+
"loss": 0.5187,
|
| 2206 |
+
"step": 3000
|
| 2207 |
+
},
|
| 2208 |
+
{
|
| 2209 |
+
"epoch": 0.3852553436580059,
|
| 2210 |
+
"grad_norm": 1.0057384967803955,
|
| 2211 |
+
"learning_rate": 1.936860068259386e-05,
|
| 2212 |
+
"loss": 0.5179,
|
| 2213 |
+
"step": 3010
|
| 2214 |
+
},
|
| 2215 |
+
{
|
| 2216 |
+
"epoch": 0.3865352617432484,
|
| 2217 |
+
"grad_norm": 1.9525461196899414,
|
| 2218 |
+
"learning_rate": 1.9359120212362534e-05,
|
| 2219 |
+
"loss": 0.4964,
|
| 2220 |
+
"step": 3020
|
| 2221 |
+
},
|
| 2222 |
+
{
|
| 2223 |
+
"epoch": 0.38781517982849095,
|
| 2224 |
+
"grad_norm": 0.9308017492294312,
|
| 2225 |
+
"learning_rate": 1.9349639742131212e-05,
|
| 2226 |
+
"loss": 0.4751,
|
| 2227 |
+
"step": 3030
|
| 2228 |
+
},
|
| 2229 |
+
{
|
| 2230 |
+
"epoch": 0.38909509791373353,
|
| 2231 |
+
"grad_norm": 0.6847271919250488,
|
| 2232 |
+
"learning_rate": 1.9340159271899886e-05,
|
| 2233 |
+
"loss": 0.5049,
|
| 2234 |
+
"step": 3040
|
| 2235 |
+
},
|
| 2236 |
+
{
|
| 2237 |
+
"epoch": 0.39037501599897606,
|
| 2238 |
+
"grad_norm": 0.9826173186302185,
|
| 2239 |
+
"learning_rate": 1.9330678801668564e-05,
|
| 2240 |
+
"loss": 0.5089,
|
| 2241 |
+
"step": 3050
|
| 2242 |
+
},
|
| 2243 |
+
{
|
| 2244 |
+
"epoch": 0.3916549340842186,
|
| 2245 |
+
"grad_norm": 1.1526622772216797,
|
| 2246 |
+
"learning_rate": 1.932119833143724e-05,
|
| 2247 |
+
"loss": 0.5263,
|
| 2248 |
+
"step": 3060
|
| 2249 |
+
},
|
| 2250 |
+
{
|
| 2251 |
+
"epoch": 0.39293485216946117,
|
| 2252 |
+
"grad_norm": 1.3890351057052612,
|
| 2253 |
+
"learning_rate": 1.9311717861205916e-05,
|
| 2254 |
+
"loss": 0.5293,
|
| 2255 |
+
"step": 3070
|
| 2256 |
+
},
|
| 2257 |
+
{
|
| 2258 |
+
"epoch": 0.3942147702547037,
|
| 2259 |
+
"grad_norm": 1.8960402011871338,
|
| 2260 |
+
"learning_rate": 1.9302237390974593e-05,
|
| 2261 |
+
"loss": 0.4944,
|
| 2262 |
+
"step": 3080
|
| 2263 |
+
},
|
| 2264 |
+
{
|
| 2265 |
+
"epoch": 0.3954946883399462,
|
| 2266 |
+
"grad_norm": 3.4239537715911865,
|
| 2267 |
+
"learning_rate": 1.929275692074327e-05,
|
| 2268 |
+
"loss": 0.5281,
|
| 2269 |
+
"step": 3090
|
| 2270 |
+
},
|
| 2271 |
+
{
|
| 2272 |
+
"epoch": 0.3967746064251888,
|
| 2273 |
+
"grad_norm": 0.8984873294830322,
|
| 2274 |
+
"learning_rate": 1.928327645051195e-05,
|
| 2275 |
+
"loss": 0.5169,
|
| 2276 |
+
"step": 3100
|
| 2277 |
+
},
|
| 2278 |
+
{
|
| 2279 |
+
"epoch": 0.39805452451043133,
|
| 2280 |
+
"grad_norm": 1.3759665489196777,
|
| 2281 |
+
"learning_rate": 1.9273795980280623e-05,
|
| 2282 |
+
"loss": 0.5086,
|
| 2283 |
+
"step": 3110
|
| 2284 |
+
},
|
| 2285 |
+
{
|
| 2286 |
+
"epoch": 0.39933444259567386,
|
| 2287 |
+
"grad_norm": 1.0449821949005127,
|
| 2288 |
+
"learning_rate": 1.92643155100493e-05,
|
| 2289 |
+
"loss": 0.4921,
|
| 2290 |
+
"step": 3120
|
| 2291 |
+
},
|
| 2292 |
+
{
|
| 2293 |
+
"epoch": 0.40061436068091644,
|
| 2294 |
+
"grad_norm": 0.9045126438140869,
|
| 2295 |
+
"learning_rate": 1.9254835039817975e-05,
|
| 2296 |
+
"loss": 0.4888,
|
| 2297 |
+
"step": 3130
|
| 2298 |
+
},
|
| 2299 |
+
{
|
| 2300 |
+
"epoch": 0.40189427876615896,
|
| 2301 |
+
"grad_norm": 0.8578636050224304,
|
| 2302 |
+
"learning_rate": 1.9245354569586653e-05,
|
| 2303 |
+
"loss": 0.4956,
|
| 2304 |
+
"step": 3140
|
| 2305 |
+
},
|
| 2306 |
+
{
|
| 2307 |
+
"epoch": 0.4031741968514015,
|
| 2308 |
+
"grad_norm": 1.2654173374176025,
|
| 2309 |
+
"learning_rate": 1.923587409935533e-05,
|
| 2310 |
+
"loss": 0.4942,
|
| 2311 |
+
"step": 3150
|
| 2312 |
+
},
|
| 2313 |
+
{
|
| 2314 |
+
"epoch": 0.40445411493664407,
|
| 2315 |
+
"grad_norm": 1.6073803901672363,
|
| 2316 |
+
"learning_rate": 1.9226393629124008e-05,
|
| 2317 |
+
"loss": 0.5037,
|
| 2318 |
+
"step": 3160
|
| 2319 |
+
},
|
| 2320 |
+
{
|
| 2321 |
+
"epoch": 0.4057340330218866,
|
| 2322 |
+
"grad_norm": 0.9294481873512268,
|
| 2323 |
+
"learning_rate": 1.9216913158892682e-05,
|
| 2324 |
+
"loss": 0.4836,
|
| 2325 |
+
"step": 3170
|
| 2326 |
+
},
|
| 2327 |
+
{
|
| 2328 |
+
"epoch": 0.4070139511071291,
|
| 2329 |
+
"grad_norm": 1.1646478176116943,
|
| 2330 |
+
"learning_rate": 1.9207432688661357e-05,
|
| 2331 |
+
"loss": 0.5189,
|
| 2332 |
+
"step": 3180
|
| 2333 |
+
},
|
| 2334 |
+
{
|
| 2335 |
+
"epoch": 0.4082938691923717,
|
| 2336 |
+
"grad_norm": 0.8620973825454712,
|
| 2337 |
+
"learning_rate": 1.9197952218430034e-05,
|
| 2338 |
+
"loss": 0.5002,
|
| 2339 |
+
"step": 3190
|
| 2340 |
+
},
|
| 2341 |
+
{
|
| 2342 |
+
"epoch": 0.40957378727761423,
|
| 2343 |
+
"grad_norm": 0.7730439305305481,
|
| 2344 |
+
"learning_rate": 1.9188471748198712e-05,
|
| 2345 |
+
"loss": 0.4802,
|
| 2346 |
+
"step": 3200
|
| 2347 |
+
},
|
| 2348 |
+
{
|
| 2349 |
+
"epoch": 0.41085370536285676,
|
| 2350 |
+
"grad_norm": 0.706767201423645,
|
| 2351 |
+
"learning_rate": 1.917899127796739e-05,
|
| 2352 |
+
"loss": 0.535,
|
| 2353 |
+
"step": 3210
|
| 2354 |
+
},
|
| 2355 |
+
{
|
| 2356 |
+
"epoch": 0.41213362344809934,
|
| 2357 |
+
"grad_norm": 1.1379191875457764,
|
| 2358 |
+
"learning_rate": 1.9169510807736064e-05,
|
| 2359 |
+
"loss": 0.5186,
|
| 2360 |
+
"step": 3220
|
| 2361 |
+
},
|
| 2362 |
+
{
|
| 2363 |
+
"epoch": 0.41341354153334187,
|
| 2364 |
+
"grad_norm": 0.6369794011116028,
|
| 2365 |
+
"learning_rate": 1.916003033750474e-05,
|
| 2366 |
+
"loss": 0.4896,
|
| 2367 |
+
"step": 3230
|
| 2368 |
+
},
|
| 2369 |
+
{
|
| 2370 |
+
"epoch": 0.4146934596185844,
|
| 2371 |
+
"grad_norm": 3.0110867023468018,
|
| 2372 |
+
"learning_rate": 1.915054986727342e-05,
|
| 2373 |
+
"loss": 0.5275,
|
| 2374 |
+
"step": 3240
|
| 2375 |
+
},
|
| 2376 |
+
{
|
| 2377 |
+
"epoch": 0.415973377703827,
|
| 2378 |
+
"grad_norm": 1.3241641521453857,
|
| 2379 |
+
"learning_rate": 1.9141069397042097e-05,
|
| 2380 |
+
"loss": 0.5078,
|
| 2381 |
+
"step": 3250
|
| 2382 |
+
},
|
| 2383 |
+
{
|
| 2384 |
+
"epoch": 0.4172532957890695,
|
| 2385 |
+
"grad_norm": 0.942676305770874,
|
| 2386 |
+
"learning_rate": 1.913158892681077e-05,
|
| 2387 |
+
"loss": 0.4901,
|
| 2388 |
+
"step": 3260
|
| 2389 |
+
},
|
| 2390 |
+
{
|
| 2391 |
+
"epoch": 0.418533213874312,
|
| 2392 |
+
"grad_norm": 2.2048139572143555,
|
| 2393 |
+
"learning_rate": 1.9122108456579446e-05,
|
| 2394 |
+
"loss": 0.5272,
|
| 2395 |
+
"step": 3270
|
| 2396 |
+
},
|
| 2397 |
+
{
|
| 2398 |
+
"epoch": 0.4198131319595546,
|
| 2399 |
+
"grad_norm": 1.8954252004623413,
|
| 2400 |
+
"learning_rate": 1.9112627986348123e-05,
|
| 2401 |
+
"loss": 0.4853,
|
| 2402 |
+
"step": 3280
|
| 2403 |
+
},
|
| 2404 |
+
{
|
| 2405 |
+
"epoch": 0.42109305004479713,
|
| 2406 |
+
"grad_norm": 0.8221827745437622,
|
| 2407 |
+
"learning_rate": 1.91031475161168e-05,
|
| 2408 |
+
"loss": 0.4919,
|
| 2409 |
+
"step": 3290
|
| 2410 |
+
},
|
| 2411 |
+
{
|
| 2412 |
+
"epoch": 0.42237296813003966,
|
| 2413 |
+
"grad_norm": 0.742501437664032,
|
| 2414 |
+
"learning_rate": 1.909366704588548e-05,
|
| 2415 |
+
"loss": 0.5139,
|
| 2416 |
+
"step": 3300
|
| 2417 |
+
},
|
| 2418 |
+
{
|
| 2419 |
+
"epoch": 0.42365288621528224,
|
| 2420 |
+
"grad_norm": 1.9924613237380981,
|
| 2421 |
+
"learning_rate": 1.9084186575654153e-05,
|
| 2422 |
+
"loss": 0.4803,
|
| 2423 |
+
"step": 3310
|
| 2424 |
+
},
|
| 2425 |
+
{
|
| 2426 |
+
"epoch": 0.42493280430052477,
|
| 2427 |
+
"grad_norm": 0.9771905541419983,
|
| 2428 |
+
"learning_rate": 1.907470610542283e-05,
|
| 2429 |
+
"loss": 0.475,
|
| 2430 |
+
"step": 3320
|
| 2431 |
+
},
|
| 2432 |
+
{
|
| 2433 |
+
"epoch": 0.4262127223857673,
|
| 2434 |
+
"grad_norm": 1.5012576580047607,
|
| 2435 |
+
"learning_rate": 1.9065225635191508e-05,
|
| 2436 |
+
"loss": 0.4941,
|
| 2437 |
+
"step": 3330
|
| 2438 |
+
},
|
| 2439 |
+
{
|
| 2440 |
+
"epoch": 0.4274926404710099,
|
| 2441 |
+
"grad_norm": 0.6949377059936523,
|
| 2442 |
+
"learning_rate": 1.9055745164960186e-05,
|
| 2443 |
+
"loss": 0.512,
|
| 2444 |
+
"step": 3340
|
| 2445 |
+
},
|
| 2446 |
+
{
|
| 2447 |
+
"epoch": 0.4287725585562524,
|
| 2448 |
+
"grad_norm": 0.9960921406745911,
|
| 2449 |
+
"learning_rate": 1.904626469472886e-05,
|
| 2450 |
+
"loss": 0.4825,
|
| 2451 |
+
"step": 3350
|
| 2452 |
+
},
|
| 2453 |
+
{
|
| 2454 |
+
"epoch": 0.43005247664149493,
|
| 2455 |
+
"grad_norm": 1.0516977310180664,
|
| 2456 |
+
"learning_rate": 1.9036784224497538e-05,
|
| 2457 |
+
"loss": 0.5117,
|
| 2458 |
+
"step": 3360
|
| 2459 |
+
},
|
| 2460 |
+
{
|
| 2461 |
+
"epoch": 0.4313323947267375,
|
| 2462 |
+
"grad_norm": 0.9433191418647766,
|
| 2463 |
+
"learning_rate": 1.9027303754266212e-05,
|
| 2464 |
+
"loss": 0.484,
|
| 2465 |
+
"step": 3370
|
| 2466 |
+
},
|
| 2467 |
+
{
|
| 2468 |
+
"epoch": 0.43261231281198004,
|
| 2469 |
+
"grad_norm": 2.7105135917663574,
|
| 2470 |
+
"learning_rate": 1.901782328403489e-05,
|
| 2471 |
+
"loss": 0.4986,
|
| 2472 |
+
"step": 3380
|
| 2473 |
+
},
|
| 2474 |
+
{
|
| 2475 |
+
"epoch": 0.43389223089722256,
|
| 2476 |
+
"grad_norm": 0.8293517827987671,
|
| 2477 |
+
"learning_rate": 1.9008342813803567e-05,
|
| 2478 |
+
"loss": 0.4797,
|
| 2479 |
+
"step": 3390
|
| 2480 |
+
},
|
| 2481 |
+
{
|
| 2482 |
+
"epoch": 0.43517214898246515,
|
| 2483 |
+
"grad_norm": 1.3104497194290161,
|
| 2484 |
+
"learning_rate": 1.899886234357224e-05,
|
| 2485 |
+
"loss": 0.4971,
|
| 2486 |
+
"step": 3400
|
| 2487 |
+
},
|
| 2488 |
+
{
|
| 2489 |
+
"epoch": 0.43645206706770767,
|
| 2490 |
+
"grad_norm": 2.3978216648101807,
|
| 2491 |
+
"learning_rate": 1.898938187334092e-05,
|
| 2492 |
+
"loss": 0.5012,
|
| 2493 |
+
"step": 3410
|
| 2494 |
+
},
|
| 2495 |
+
{
|
| 2496 |
+
"epoch": 0.4377319851529502,
|
| 2497 |
+
"grad_norm": 0.838647186756134,
|
| 2498 |
+
"learning_rate": 1.8979901403109594e-05,
|
| 2499 |
+
"loss": 0.5042,
|
| 2500 |
+
"step": 3420
|
| 2501 |
+
},
|
| 2502 |
+
{
|
| 2503 |
+
"epoch": 0.4390119032381928,
|
| 2504 |
+
"grad_norm": 0.7617325782775879,
|
| 2505 |
+
"learning_rate": 1.897042093287827e-05,
|
| 2506 |
+
"loss": 0.5094,
|
| 2507 |
+
"step": 3430
|
| 2508 |
+
},
|
| 2509 |
+
{
|
| 2510 |
+
"epoch": 0.4402918213234353,
|
| 2511 |
+
"grad_norm": 0.8923324942588806,
|
| 2512 |
+
"learning_rate": 1.896094046264695e-05,
|
| 2513 |
+
"loss": 0.487,
|
| 2514 |
+
"step": 3440
|
| 2515 |
+
},
|
| 2516 |
+
{
|
| 2517 |
+
"epoch": 0.44157173940867783,
|
| 2518 |
+
"grad_norm": 1.7616841793060303,
|
| 2519 |
+
"learning_rate": 1.8951459992415627e-05,
|
| 2520 |
+
"loss": 0.4929,
|
| 2521 |
+
"step": 3450
|
| 2522 |
+
},
|
| 2523 |
+
{
|
| 2524 |
+
"epoch": 0.4428516574939204,
|
| 2525 |
+
"grad_norm": 0.7590444087982178,
|
| 2526 |
+
"learning_rate": 1.89419795221843e-05,
|
| 2527 |
+
"loss": 0.4844,
|
| 2528 |
+
"step": 3460
|
| 2529 |
+
},
|
| 2530 |
+
{
|
| 2531 |
+
"epoch": 0.44413157557916294,
|
| 2532 |
+
"grad_norm": 0.8714098930358887,
|
| 2533 |
+
"learning_rate": 1.893249905195298e-05,
|
| 2534 |
+
"loss": 0.4736,
|
| 2535 |
+
"step": 3470
|
| 2536 |
+
},
|
| 2537 |
+
{
|
| 2538 |
+
"epoch": 0.44541149366440547,
|
| 2539 |
+
"grad_norm": 1.3062705993652344,
|
| 2540 |
+
"learning_rate": 1.8923018581721656e-05,
|
| 2541 |
+
"loss": 0.501,
|
| 2542 |
+
"step": 3480
|
| 2543 |
+
},
|
| 2544 |
+
{
|
| 2545 |
+
"epoch": 0.44669141174964805,
|
| 2546 |
+
"grad_norm": 1.3274645805358887,
|
| 2547 |
+
"learning_rate": 1.891353811149033e-05,
|
| 2548 |
+
"loss": 0.506,
|
| 2549 |
+
"step": 3490
|
| 2550 |
+
},
|
| 2551 |
+
{
|
| 2552 |
+
"epoch": 0.4479713298348906,
|
| 2553 |
+
"grad_norm": 3.321751832962036,
|
| 2554 |
+
"learning_rate": 1.8904057641259008e-05,
|
| 2555 |
+
"loss": 0.5244,
|
| 2556 |
+
"step": 3500
|
| 2557 |
+
},
|
| 2558 |
+
{
|
| 2559 |
+
"epoch": 0.4492512479201331,
|
| 2560 |
+
"grad_norm": 0.8489649891853333,
|
| 2561 |
+
"learning_rate": 1.8894577171027683e-05,
|
| 2562 |
+
"loss": 0.5003,
|
| 2563 |
+
"step": 3510
|
| 2564 |
+
},
|
| 2565 |
+
{
|
| 2566 |
+
"epoch": 0.4505311660053757,
|
| 2567 |
+
"grad_norm": 1.0105081796646118,
|
| 2568 |
+
"learning_rate": 1.888509670079636e-05,
|
| 2569 |
+
"loss": 0.4748,
|
| 2570 |
+
"step": 3520
|
| 2571 |
+
},
|
| 2572 |
+
{
|
| 2573 |
+
"epoch": 0.4518110840906182,
|
| 2574 |
+
"grad_norm": 1.452003836631775,
|
| 2575 |
+
"learning_rate": 1.8875616230565038e-05,
|
| 2576 |
+
"loss": 0.4842,
|
| 2577 |
+
"step": 3530
|
| 2578 |
+
},
|
| 2579 |
+
{
|
| 2580 |
+
"epoch": 0.45309100217586074,
|
| 2581 |
+
"grad_norm": 0.8805971741676331,
|
| 2582 |
+
"learning_rate": 1.8866135760333716e-05,
|
| 2583 |
+
"loss": 0.4923,
|
| 2584 |
+
"step": 3540
|
| 2585 |
+
},
|
| 2586 |
+
{
|
| 2587 |
+
"epoch": 0.4543709202611033,
|
| 2588 |
+
"grad_norm": 2.563197135925293,
|
| 2589 |
+
"learning_rate": 1.885665529010239e-05,
|
| 2590 |
+
"loss": 0.5021,
|
| 2591 |
+
"step": 3550
|
| 2592 |
+
},
|
| 2593 |
+
{
|
| 2594 |
+
"epoch": 0.45565083834634584,
|
| 2595 |
+
"grad_norm": 0.7883086204528809,
|
| 2596 |
+
"learning_rate": 1.8847174819871067e-05,
|
| 2597 |
+
"loss": 0.4891,
|
| 2598 |
+
"step": 3560
|
| 2599 |
+
},
|
| 2600 |
+
{
|
| 2601 |
+
"epoch": 0.45693075643158837,
|
| 2602 |
+
"grad_norm": 1.5375005006790161,
|
| 2603 |
+
"learning_rate": 1.8837694349639745e-05,
|
| 2604 |
+
"loss": 0.486,
|
| 2605 |
+
"step": 3570
|
| 2606 |
+
},
|
| 2607 |
+
{
|
| 2608 |
+
"epoch": 0.4582106745168309,
|
| 2609 |
+
"grad_norm": 1.0281457901000977,
|
| 2610 |
+
"learning_rate": 1.8828213879408423e-05,
|
| 2611 |
+
"loss": 0.5057,
|
| 2612 |
+
"step": 3580
|
| 2613 |
+
},
|
| 2614 |
+
{
|
| 2615 |
+
"epoch": 0.4594905926020735,
|
| 2616 |
+
"grad_norm": 1.281043291091919,
|
| 2617 |
+
"learning_rate": 1.8818733409177097e-05,
|
| 2618 |
+
"loss": 0.4998,
|
| 2619 |
+
"step": 3590
|
| 2620 |
+
},
|
| 2621 |
+
{
|
| 2622 |
+
"epoch": 0.460770510687316,
|
| 2623 |
+
"grad_norm": 0.9816468954086304,
|
| 2624 |
+
"learning_rate": 1.880925293894577e-05,
|
| 2625 |
+
"loss": 0.4944,
|
| 2626 |
+
"step": 3600
|
| 2627 |
+
},
|
| 2628 |
+
{
|
| 2629 |
+
"epoch": 0.46205042877255853,
|
| 2630 |
+
"grad_norm": 0.9405584335327148,
|
| 2631 |
+
"learning_rate": 1.879977246871445e-05,
|
| 2632 |
+
"loss": 0.5036,
|
| 2633 |
+
"step": 3610
|
| 2634 |
+
},
|
| 2635 |
+
{
|
| 2636 |
+
"epoch": 0.4633303468578011,
|
| 2637 |
+
"grad_norm": 1.3431020975112915,
|
| 2638 |
+
"learning_rate": 1.8790291998483127e-05,
|
| 2639 |
+
"loss": 0.5004,
|
| 2640 |
+
"step": 3620
|
| 2641 |
+
},
|
| 2642 |
+
{
|
| 2643 |
+
"epoch": 0.46461026494304364,
|
| 2644 |
+
"grad_norm": 1.1362318992614746,
|
| 2645 |
+
"learning_rate": 1.8780811528251804e-05,
|
| 2646 |
+
"loss": 0.494,
|
| 2647 |
+
"step": 3630
|
| 2648 |
+
},
|
| 2649 |
+
{
|
| 2650 |
+
"epoch": 0.46589018302828616,
|
| 2651 |
+
"grad_norm": 1.2532789707183838,
|
| 2652 |
+
"learning_rate": 1.877133105802048e-05,
|
| 2653 |
+
"loss": 0.4748,
|
| 2654 |
+
"step": 3640
|
| 2655 |
+
},
|
| 2656 |
+
{
|
| 2657 |
+
"epoch": 0.46717010111352875,
|
| 2658 |
+
"grad_norm": 0.729250431060791,
|
| 2659 |
+
"learning_rate": 1.8761850587789156e-05,
|
| 2660 |
+
"loss": 0.5086,
|
| 2661 |
+
"step": 3650
|
| 2662 |
+
},
|
| 2663 |
+
{
|
| 2664 |
+
"epoch": 0.4684500191987713,
|
| 2665 |
+
"grad_norm": 0.6976643800735474,
|
| 2666 |
+
"learning_rate": 1.875237011755783e-05,
|
| 2667 |
+
"loss": 0.4781,
|
| 2668 |
+
"step": 3660
|
| 2669 |
+
},
|
| 2670 |
+
{
|
| 2671 |
+
"epoch": 0.4697299372840138,
|
| 2672 |
+
"grad_norm": 0.9878288507461548,
|
| 2673 |
+
"learning_rate": 1.8742889647326508e-05,
|
| 2674 |
+
"loss": 0.4809,
|
| 2675 |
+
"step": 3670
|
| 2676 |
+
},
|
| 2677 |
+
{
|
| 2678 |
+
"epoch": 0.4710098553692564,
|
| 2679 |
+
"grad_norm": 0.9907483458518982,
|
| 2680 |
+
"learning_rate": 1.8733409177095186e-05,
|
| 2681 |
+
"loss": 0.4914,
|
| 2682 |
+
"step": 3680
|
| 2683 |
+
},
|
| 2684 |
+
{
|
| 2685 |
+
"epoch": 0.4722897734544989,
|
| 2686 |
+
"grad_norm": 1.046175479888916,
|
| 2687 |
+
"learning_rate": 1.872392870686386e-05,
|
| 2688 |
+
"loss": 0.5048,
|
| 2689 |
+
"step": 3690
|
| 2690 |
+
},
|
| 2691 |
+
{
|
| 2692 |
+
"epoch": 0.47356969153974143,
|
| 2693 |
+
"grad_norm": 1.8168991804122925,
|
| 2694 |
+
"learning_rate": 1.8714448236632538e-05,
|
| 2695 |
+
"loss": 0.4528,
|
| 2696 |
+
"step": 3700
|
| 2697 |
+
},
|
| 2698 |
+
{
|
| 2699 |
+
"epoch": 0.474849609624984,
|
| 2700 |
+
"grad_norm": 2.9127612113952637,
|
| 2701 |
+
"learning_rate": 1.8704967766401216e-05,
|
| 2702 |
+
"loss": 0.5226,
|
| 2703 |
+
"step": 3710
|
| 2704 |
+
},
|
| 2705 |
+
{
|
| 2706 |
+
"epoch": 0.47612952771022654,
|
| 2707 |
+
"grad_norm": 1.4959512948989868,
|
| 2708 |
+
"learning_rate": 1.8695487296169893e-05,
|
| 2709 |
+
"loss": 0.4727,
|
| 2710 |
+
"step": 3720
|
| 2711 |
+
},
|
| 2712 |
+
{
|
| 2713 |
+
"epoch": 0.47740944579546907,
|
| 2714 |
+
"grad_norm": 0.6817557215690613,
|
| 2715 |
+
"learning_rate": 1.8686006825938568e-05,
|
| 2716 |
+
"loss": 0.468,
|
| 2717 |
+
"step": 3730
|
| 2718 |
+
},
|
| 2719 |
+
{
|
| 2720 |
+
"epoch": 0.47868936388071165,
|
| 2721 |
+
"grad_norm": 1.5274559259414673,
|
| 2722 |
+
"learning_rate": 1.8676526355707245e-05,
|
| 2723 |
+
"loss": 0.5158,
|
| 2724 |
+
"step": 3740
|
| 2725 |
+
},
|
| 2726 |
+
{
|
| 2727 |
+
"epoch": 0.4799692819659542,
|
| 2728 |
+
"grad_norm": 1.8979755640029907,
|
| 2729 |
+
"learning_rate": 1.866704588547592e-05,
|
| 2730 |
+
"loss": 0.4628,
|
| 2731 |
+
"step": 3750
|
| 2732 |
+
},
|
| 2733 |
+
{
|
| 2734 |
+
"epoch": 0.4812492000511967,
|
| 2735 |
+
"grad_norm": 0.7936620712280273,
|
| 2736 |
+
"learning_rate": 1.8657565415244597e-05,
|
| 2737 |
+
"loss": 0.5159,
|
| 2738 |
+
"step": 3760
|
| 2739 |
+
},
|
| 2740 |
+
{
|
| 2741 |
+
"epoch": 0.4825291181364393,
|
| 2742 |
+
"grad_norm": 1.2421464920043945,
|
| 2743 |
+
"learning_rate": 1.8648084945013275e-05,
|
| 2744 |
+
"loss": 0.495,
|
| 2745 |
+
"step": 3770
|
| 2746 |
+
},
|
| 2747 |
+
{
|
| 2748 |
+
"epoch": 0.4838090362216818,
|
| 2749 |
+
"grad_norm": 0.8110264539718628,
|
| 2750 |
+
"learning_rate": 1.8638604474781953e-05,
|
| 2751 |
+
"loss": 0.514,
|
| 2752 |
+
"step": 3780
|
| 2753 |
+
},
|
| 2754 |
+
{
|
| 2755 |
+
"epoch": 0.48508895430692434,
|
| 2756 |
+
"grad_norm": 1.168480634689331,
|
| 2757 |
+
"learning_rate": 1.8629124004550627e-05,
|
| 2758 |
+
"loss": 0.4842,
|
| 2759 |
+
"step": 3790
|
| 2760 |
+
},
|
| 2761 |
+
{
|
| 2762 |
+
"epoch": 0.4863688723921669,
|
| 2763 |
+
"grad_norm": 1.4188495874404907,
|
| 2764 |
+
"learning_rate": 1.86196435343193e-05,
|
| 2765 |
+
"loss": 0.4855,
|
| 2766 |
+
"step": 3800
|
| 2767 |
+
}
|
| 2768 |
+
],
|
| 2769 |
+
"logging_steps": 10,
|
| 2770 |
+
"max_steps": 23439,
|
| 2771 |
+
"num_input_tokens_seen": 0,
|
| 2772 |
+
"num_train_epochs": 3,
|
| 2773 |
+
"save_steps": 200,
|
| 2774 |
+
"stateful_callbacks": {
|
| 2775 |
+
"EarlyStoppingCallback": {
|
| 2776 |
+
"args": {
|
| 2777 |
+
"early_stopping_patience": 3,
|
| 2778 |
+
"early_stopping_threshold": 0.0
|
| 2779 |
+
},
|
| 2780 |
+
"attributes": {
|
| 2781 |
+
"early_stopping_patience_counter": 0
|
| 2782 |
+
}
|
| 2783 |
+
},
|
| 2784 |
+
"TrainerControl": {
|
| 2785 |
+
"args": {
|
| 2786 |
+
"should_epoch_stop": false,
|
| 2787 |
+
"should_evaluate": false,
|
| 2788 |
+
"should_log": false,
|
| 2789 |
+
"should_save": true,
|
| 2790 |
+
"should_training_stop": false
|
| 2791 |
+
},
|
| 2792 |
+
"attributes": {}
|
| 2793 |
+
}
|
| 2794 |
+
},
|
| 2795 |
+
"total_flos": 6.3988608663552e+16,
|
| 2796 |
+
"train_batch_size": 64,
|
| 2797 |
+
"trial_name": null,
|
| 2798 |
+
"trial_params": null
|
| 2799 |
+
}
|
codebert-base-cross-entropy/checkpoint-3800/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:dbdc2f4c9cd8ad671ca740ab491512a481c34d40fafc15bd45db45cd63388bf1
|
| 3 |
+
size 5841
|
codebert-base-cross-entropy/checkpoint-3800/vocab.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
codebert-base-cross-entropy/checkpoint-4000/config.json
ADDED
|
@@ -0,0 +1,27 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"RobertaForSequenceClassification"
|
| 4 |
+
],
|
| 5 |
+
"attention_probs_dropout_prob": 0.1,
|
| 6 |
+
"bos_token_id": 0,
|
| 7 |
+
"classifier_dropout": null,
|
| 8 |
+
"dtype": "float32",
|
| 9 |
+
"eos_token_id": 2,
|
| 10 |
+
"hidden_act": "gelu",
|
| 11 |
+
"hidden_dropout_prob": 0.1,
|
| 12 |
+
"hidden_size": 768,
|
| 13 |
+
"initializer_range": 0.02,
|
| 14 |
+
"intermediate_size": 3072,
|
| 15 |
+
"layer_norm_eps": 1e-05,
|
| 16 |
+
"max_position_embeddings": 514,
|
| 17 |
+
"model_type": "roberta",
|
| 18 |
+
"num_attention_heads": 12,
|
| 19 |
+
"num_hidden_layers": 12,
|
| 20 |
+
"output_past": true,
|
| 21 |
+
"pad_token_id": 1,
|
| 22 |
+
"position_embedding_type": "absolute",
|
| 23 |
+
"transformers_version": "4.56.0",
|
| 24 |
+
"type_vocab_size": 1,
|
| 25 |
+
"use_cache": true,
|
| 26 |
+
"vocab_size": 50265
|
| 27 |
+
}
|
codebert-base-cross-entropy/checkpoint-4000/merges.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
codebert-base-cross-entropy/checkpoint-4000/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2800d2a3c0e50047644191270159eb624f6b52017ba169c0c5ec647b10b9e14b
|
| 3 |
+
size 498612824
|
codebert-base-cross-entropy/checkpoint-4000/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f7c3f30249d8e1bfbfb962de7737378143dd95355634be8f01ab5d26730903e8
|
| 3 |
+
size 4741859
|
codebert-base-cross-entropy/checkpoint-4000/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:938845322a52de09d3ff8726e960c273dbdbc92ab02c1b71947488421ebcf7d3
|
| 3 |
+
size 14645
|
codebert-base-cross-entropy/checkpoint-4000/scaler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:68e7755116e5a328e3b0591c5c7b4e2b055e71a1d3865d416b158778f49a17e2
|
| 3 |
+
size 1383
|
codebert-base-cross-entropy/checkpoint-4000/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bb881d17d2d6e34ad832fd96935eb15df03a1e272d372fbeaa04a3db4422df26
|
| 3 |
+
size 1465
|
codebert-base-cross-entropy/checkpoint-4000/special_tokens_map.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": true,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"cls_token": {
|
| 10 |
+
"content": "<s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": true,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"eos_token": {
|
| 17 |
+
"content": "</s>",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": true,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"mask_token": {
|
| 24 |
+
"content": "<mask>",
|
| 25 |
+
"lstrip": true,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"pad_token": {
|
| 31 |
+
"content": "<pad>",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": true,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
},
|
| 37 |
+
"sep_token": {
|
| 38 |
+
"content": "</s>",
|
| 39 |
+
"lstrip": false,
|
| 40 |
+
"normalized": true,
|
| 41 |
+
"rstrip": false,
|
| 42 |
+
"single_word": false
|
| 43 |
+
},
|
| 44 |
+
"unk_token": {
|
| 45 |
+
"content": "<unk>",
|
| 46 |
+
"lstrip": false,
|
| 47 |
+
"normalized": true,
|
| 48 |
+
"rstrip": false,
|
| 49 |
+
"single_word": false
|
| 50 |
+
}
|
| 51 |
+
}
|
codebert-base-cross-entropy/checkpoint-4000/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
codebert-base-cross-entropy/checkpoint-4000/tokenizer_config.json
ADDED
|
@@ -0,0 +1,58 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"added_tokens_decoder": {
|
| 4 |
+
"0": {
|
| 5 |
+
"content": "<s>",
|
| 6 |
+
"lstrip": false,
|
| 7 |
+
"normalized": true,
|
| 8 |
+
"rstrip": false,
|
| 9 |
+
"single_word": false,
|
| 10 |
+
"special": true
|
| 11 |
+
},
|
| 12 |
+
"1": {
|
| 13 |
+
"content": "<pad>",
|
| 14 |
+
"lstrip": false,
|
| 15 |
+
"normalized": true,
|
| 16 |
+
"rstrip": false,
|
| 17 |
+
"single_word": false,
|
| 18 |
+
"special": true
|
| 19 |
+
},
|
| 20 |
+
"2": {
|
| 21 |
+
"content": "</s>",
|
| 22 |
+
"lstrip": false,
|
| 23 |
+
"normalized": true,
|
| 24 |
+
"rstrip": false,
|
| 25 |
+
"single_word": false,
|
| 26 |
+
"special": true
|
| 27 |
+
},
|
| 28 |
+
"3": {
|
| 29 |
+
"content": "<unk>",
|
| 30 |
+
"lstrip": false,
|
| 31 |
+
"normalized": true,
|
| 32 |
+
"rstrip": false,
|
| 33 |
+
"single_word": false,
|
| 34 |
+
"special": true
|
| 35 |
+
},
|
| 36 |
+
"50264": {
|
| 37 |
+
"content": "<mask>",
|
| 38 |
+
"lstrip": true,
|
| 39 |
+
"normalized": false,
|
| 40 |
+
"rstrip": false,
|
| 41 |
+
"single_word": false,
|
| 42 |
+
"special": true
|
| 43 |
+
}
|
| 44 |
+
},
|
| 45 |
+
"bos_token": "<s>",
|
| 46 |
+
"clean_up_tokenization_spaces": false,
|
| 47 |
+
"cls_token": "<s>",
|
| 48 |
+
"eos_token": "</s>",
|
| 49 |
+
"errors": "replace",
|
| 50 |
+
"extra_special_tokens": {},
|
| 51 |
+
"mask_token": "<mask>",
|
| 52 |
+
"model_max_length": 512,
|
| 53 |
+
"pad_token": "<pad>",
|
| 54 |
+
"sep_token": "</s>",
|
| 55 |
+
"tokenizer_class": "RobertaTokenizer",
|
| 56 |
+
"trim_offsets": true,
|
| 57 |
+
"unk_token": "<unk>"
|
| 58 |
+
}
|
codebert-base-cross-entropy/checkpoint-4000/trainer_state.json
ADDED
|
@@ -0,0 +1,2951 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 4000,
|
| 3 |
+
"best_metric": 0.7899137546751205,
|
| 4 |
+
"best_model_checkpoint": "./output_checkpoints/codebert-base-cross-entropy/checkpoint-4000",
|
| 5 |
+
"epoch": 0.5119672340970178,
|
| 6 |
+
"eval_steps": 2000,
|
| 7 |
+
"global_step": 4000,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.0012799180852425445,
|
| 14 |
+
"grad_norm": 3.6954128742218018,
|
| 15 |
+
"learning_rate": 3.840409643695328e-08,
|
| 16 |
+
"loss": 0.7763,
|
| 17 |
+
"step": 10
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.002559836170485089,
|
| 21 |
+
"grad_norm": 3.3660645484924316,
|
| 22 |
+
"learning_rate": 8.10753147002347e-08,
|
| 23 |
+
"loss": 0.7763,
|
| 24 |
+
"step": 20
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.0038397542557276334,
|
| 28 |
+
"grad_norm": 4.491067886352539,
|
| 29 |
+
"learning_rate": 1.2374653296351612e-07,
|
| 30 |
+
"loss": 0.7889,
|
| 31 |
+
"step": 30
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.005119672340970178,
|
| 35 |
+
"grad_norm": 1.5920521020889282,
|
| 36 |
+
"learning_rate": 1.6641775122679754e-07,
|
| 37 |
+
"loss": 0.7888,
|
| 38 |
+
"step": 40
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.006399590426212722,
|
| 42 |
+
"grad_norm": 4.410898208618164,
|
| 43 |
+
"learning_rate": 2.0908896949007894e-07,
|
| 44 |
+
"loss": 0.7692,
|
| 45 |
+
"step": 50
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.007679508511455267,
|
| 49 |
+
"grad_norm": 2.5669264793395996,
|
| 50 |
+
"learning_rate": 2.517601877533604e-07,
|
| 51 |
+
"loss": 0.7735,
|
| 52 |
+
"step": 60
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.008959426596697812,
|
| 56 |
+
"grad_norm": 6.190986156463623,
|
| 57 |
+
"learning_rate": 2.944314060166418e-07,
|
| 58 |
+
"loss": 0.7679,
|
| 59 |
+
"step": 70
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 0.010239344681940356,
|
| 63 |
+
"grad_norm": 5.736413478851318,
|
| 64 |
+
"learning_rate": 3.371026242799232e-07,
|
| 65 |
+
"loss": 0.744,
|
| 66 |
+
"step": 80
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"epoch": 0.011519262767182901,
|
| 70 |
+
"grad_norm": 4.278866767883301,
|
| 71 |
+
"learning_rate": 3.7977384254320464e-07,
|
| 72 |
+
"loss": 0.7542,
|
| 73 |
+
"step": 90
|
| 74 |
+
},
|
| 75 |
+
{
|
| 76 |
+
"epoch": 0.012799180852425445,
|
| 77 |
+
"grad_norm": 5.565455436706543,
|
| 78 |
+
"learning_rate": 4.22445060806486e-07,
|
| 79 |
+
"loss": 0.752,
|
| 80 |
+
"step": 100
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.01407909893766799,
|
| 84 |
+
"grad_norm": 4.031781196594238,
|
| 85 |
+
"learning_rate": 4.651162790697675e-07,
|
| 86 |
+
"loss": 0.7614,
|
| 87 |
+
"step": 110
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.015359017022910534,
|
| 91 |
+
"grad_norm": 2.484126329421997,
|
| 92 |
+
"learning_rate": 5.077874973330489e-07,
|
| 93 |
+
"loss": 0.7664,
|
| 94 |
+
"step": 120
|
| 95 |
+
},
|
| 96 |
+
{
|
| 97 |
+
"epoch": 0.016638935108153077,
|
| 98 |
+
"grad_norm": 2.7525155544281006,
|
| 99 |
+
"learning_rate": 5.504587155963304e-07,
|
| 100 |
+
"loss": 0.7521,
|
| 101 |
+
"step": 130
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"epoch": 0.017918853193395624,
|
| 105 |
+
"grad_norm": 1.6374385356903076,
|
| 106 |
+
"learning_rate": 5.931299338596117e-07,
|
| 107 |
+
"loss": 0.7436,
|
| 108 |
+
"step": 140
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"epoch": 0.019198771278638168,
|
| 112 |
+
"grad_norm": 2.370774269104004,
|
| 113 |
+
"learning_rate": 6.358011521228932e-07,
|
| 114 |
+
"loss": 0.7429,
|
| 115 |
+
"step": 150
|
| 116 |
+
},
|
| 117 |
+
{
|
| 118 |
+
"epoch": 0.02047868936388071,
|
| 119 |
+
"grad_norm": 2.8948566913604736,
|
| 120 |
+
"learning_rate": 6.784723703861745e-07,
|
| 121 |
+
"loss": 0.7333,
|
| 122 |
+
"step": 160
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.021758607449123255,
|
| 126 |
+
"grad_norm": 2.7481186389923096,
|
| 127 |
+
"learning_rate": 7.21143588649456e-07,
|
| 128 |
+
"loss": 0.7294,
|
| 129 |
+
"step": 170
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.023038525534365802,
|
| 133 |
+
"grad_norm": 2.0159404277801514,
|
| 134 |
+
"learning_rate": 7.638148069127374e-07,
|
| 135 |
+
"loss": 0.731,
|
| 136 |
+
"step": 180
|
| 137 |
+
},
|
| 138 |
+
{
|
| 139 |
+
"epoch": 0.024318443619608346,
|
| 140 |
+
"grad_norm": 3.6898131370544434,
|
| 141 |
+
"learning_rate": 8.064860251760189e-07,
|
| 142 |
+
"loss": 0.7153,
|
| 143 |
+
"step": 190
|
| 144 |
+
},
|
| 145 |
+
{
|
| 146 |
+
"epoch": 0.02559836170485089,
|
| 147 |
+
"grad_norm": 3.2842342853546143,
|
| 148 |
+
"learning_rate": 8.491572434393003e-07,
|
| 149 |
+
"loss": 0.7245,
|
| 150 |
+
"step": 200
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"epoch": 0.02559836170485089,
|
| 154 |
+
"eval_accuracy": 0.47695,
|
| 155 |
+
"eval_loss": 0.7148504853248596,
|
| 156 |
+
"eval_macro_f1": 0.3229290091066048,
|
| 157 |
+
"eval_precision": 0.238475,
|
| 158 |
+
"eval_recall": 0.5,
|
| 159 |
+
"eval_runtime": 1641.324,
|
| 160 |
+
"eval_samples_per_second": 60.926,
|
| 161 |
+
"eval_steps_per_second": 0.476,
|
| 162 |
+
"step": 200
|
| 163 |
+
},
|
| 164 |
+
{
|
| 165 |
+
"epoch": 0.026878279790093433,
|
| 166 |
+
"grad_norm": 2.294558048248291,
|
| 167 |
+
"learning_rate": 8.918284617025817e-07,
|
| 168 |
+
"loss": 0.718,
|
| 169 |
+
"step": 210
|
| 170 |
+
},
|
| 171 |
+
{
|
| 172 |
+
"epoch": 0.02815819787533598,
|
| 173 |
+
"grad_norm": 2.9467175006866455,
|
| 174 |
+
"learning_rate": 9.344996799658632e-07,
|
| 175 |
+
"loss": 0.7076,
|
| 176 |
+
"step": 220
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.029438115960578524,
|
| 180 |
+
"grad_norm": 2.6889193058013916,
|
| 181 |
+
"learning_rate": 9.771708982291445e-07,
|
| 182 |
+
"loss": 0.7019,
|
| 183 |
+
"step": 230
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 0.030718034045821067,
|
| 187 |
+
"grad_norm": 0.7491384744644165,
|
| 188 |
+
"learning_rate": 1.0198421164924258e-06,
|
| 189 |
+
"loss": 0.7046,
|
| 190 |
+
"step": 240
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 0.031997952131063614,
|
| 194 |
+
"grad_norm": 2.0363433361053467,
|
| 195 |
+
"learning_rate": 1.0625133347557074e-06,
|
| 196 |
+
"loss": 0.7085,
|
| 197 |
+
"step": 250
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 0.033277870216306155,
|
| 201 |
+
"grad_norm": 1.1431337594985962,
|
| 202 |
+
"learning_rate": 1.1051845530189888e-06,
|
| 203 |
+
"loss": 0.7008,
|
| 204 |
+
"step": 260
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 0.0345577883015487,
|
| 208 |
+
"grad_norm": 3.7735233306884766,
|
| 209 |
+
"learning_rate": 1.1478557712822702e-06,
|
| 210 |
+
"loss": 0.7015,
|
| 211 |
+
"step": 270
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 0.03583770638679125,
|
| 215 |
+
"grad_norm": 0.7889328002929688,
|
| 216 |
+
"learning_rate": 1.1905269895455517e-06,
|
| 217 |
+
"loss": 0.699,
|
| 218 |
+
"step": 280
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 0.03711762447203379,
|
| 222 |
+
"grad_norm": 0.8796226382255554,
|
| 223 |
+
"learning_rate": 1.233198207808833e-06,
|
| 224 |
+
"loss": 0.6922,
|
| 225 |
+
"step": 290
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 0.038397542557276336,
|
| 229 |
+
"grad_norm": 0.7966652512550354,
|
| 230 |
+
"learning_rate": 1.2758694260721145e-06,
|
| 231 |
+
"loss": 0.6949,
|
| 232 |
+
"step": 300
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 0.039677460642518876,
|
| 236 |
+
"grad_norm": 0.7029069066047668,
|
| 237 |
+
"learning_rate": 1.318540644335396e-06,
|
| 238 |
+
"loss": 0.695,
|
| 239 |
+
"step": 310
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 0.04095737872776142,
|
| 243 |
+
"grad_norm": 2.04327392578125,
|
| 244 |
+
"learning_rate": 1.3612118625986772e-06,
|
| 245 |
+
"loss": 0.6944,
|
| 246 |
+
"step": 320
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 0.04223729681300397,
|
| 250 |
+
"grad_norm": 0.918025553226471,
|
| 251 |
+
"learning_rate": 1.4038830808619588e-06,
|
| 252 |
+
"loss": 0.6863,
|
| 253 |
+
"step": 330
|
| 254 |
+
},
|
| 255 |
+
{
|
| 256 |
+
"epoch": 0.04351721489824651,
|
| 257 |
+
"grad_norm": 0.9528706073760986,
|
| 258 |
+
"learning_rate": 1.4465542991252401e-06,
|
| 259 |
+
"loss": 0.6907,
|
| 260 |
+
"step": 340
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 0.04479713298348906,
|
| 264 |
+
"grad_norm": 1.071359634399414,
|
| 265 |
+
"learning_rate": 1.4892255173885215e-06,
|
| 266 |
+
"loss": 0.6873,
|
| 267 |
+
"step": 350
|
| 268 |
+
},
|
| 269 |
+
{
|
| 270 |
+
"epoch": 0.046077051068731605,
|
| 271 |
+
"grad_norm": 1.9875625371932983,
|
| 272 |
+
"learning_rate": 1.531896735651803e-06,
|
| 273 |
+
"loss": 0.6942,
|
| 274 |
+
"step": 360
|
| 275 |
+
},
|
| 276 |
+
{
|
| 277 |
+
"epoch": 0.047356969153974145,
|
| 278 |
+
"grad_norm": 2.0222439765930176,
|
| 279 |
+
"learning_rate": 1.5745679539150842e-06,
|
| 280 |
+
"loss": 0.6898,
|
| 281 |
+
"step": 370
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 0.04863688723921669,
|
| 285 |
+
"grad_norm": 1.7827857732772827,
|
| 286 |
+
"learning_rate": 1.6172391721783658e-06,
|
| 287 |
+
"loss": 0.6909,
|
| 288 |
+
"step": 380
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"epoch": 0.04991680532445923,
|
| 292 |
+
"grad_norm": 1.065674901008606,
|
| 293 |
+
"learning_rate": 1.6599103904416472e-06,
|
| 294 |
+
"loss": 0.6877,
|
| 295 |
+
"step": 390
|
| 296 |
+
},
|
| 297 |
+
{
|
| 298 |
+
"epoch": 0.05119672340970178,
|
| 299 |
+
"grad_norm": 0.8346193432807922,
|
| 300 |
+
"learning_rate": 1.7025816087049288e-06,
|
| 301 |
+
"loss": 0.6889,
|
| 302 |
+
"step": 400
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 0.05119672340970178,
|
| 306 |
+
"eval_accuracy": 0.54043,
|
| 307 |
+
"eval_loss": 0.6827595233917236,
|
| 308 |
+
"eval_macro_f1": 0.41833840347598905,
|
| 309 |
+
"eval_precision": 0.5839205994464474,
|
| 310 |
+
"eval_recall": 0.5204152767503025,
|
| 311 |
+
"eval_runtime": 1634.6244,
|
| 312 |
+
"eval_samples_per_second": 61.176,
|
| 313 |
+
"eval_steps_per_second": 0.478,
|
| 314 |
+
"step": 400
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 0.052476641494944326,
|
| 318 |
+
"grad_norm": 0.8074935674667358,
|
| 319 |
+
"learning_rate": 1.7452528269682101e-06,
|
| 320 |
+
"loss": 0.684,
|
| 321 |
+
"step": 410
|
| 322 |
+
},
|
| 323 |
+
{
|
| 324 |
+
"epoch": 0.053756559580186866,
|
| 325 |
+
"grad_norm": 0.8965069651603699,
|
| 326 |
+
"learning_rate": 1.7879240452314913e-06,
|
| 327 |
+
"loss": 0.6922,
|
| 328 |
+
"step": 420
|
| 329 |
+
},
|
| 330 |
+
{
|
| 331 |
+
"epoch": 0.05503647766542941,
|
| 332 |
+
"grad_norm": 2.597215414047241,
|
| 333 |
+
"learning_rate": 1.8305952634947729e-06,
|
| 334 |
+
"loss": 0.6909,
|
| 335 |
+
"step": 430
|
| 336 |
+
},
|
| 337 |
+
{
|
| 338 |
+
"epoch": 0.05631639575067196,
|
| 339 |
+
"grad_norm": 0.8233373165130615,
|
| 340 |
+
"learning_rate": 1.8732664817580542e-06,
|
| 341 |
+
"loss": 0.6831,
|
| 342 |
+
"step": 440
|
| 343 |
+
},
|
| 344 |
+
{
|
| 345 |
+
"epoch": 0.0575963138359145,
|
| 346 |
+
"grad_norm": 1.0624489784240723,
|
| 347 |
+
"learning_rate": 1.915937700021336e-06,
|
| 348 |
+
"loss": 0.6817,
|
| 349 |
+
"step": 450
|
| 350 |
+
},
|
| 351 |
+
{
|
| 352 |
+
"epoch": 0.05887623192115705,
|
| 353 |
+
"grad_norm": 2.0623559951782227,
|
| 354 |
+
"learning_rate": 1.958608918284617e-06,
|
| 355 |
+
"loss": 0.6819,
|
| 356 |
+
"step": 460
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 0.06015615000639959,
|
| 360 |
+
"grad_norm": 0.912433922290802,
|
| 361 |
+
"learning_rate": 2.0012801365478988e-06,
|
| 362 |
+
"loss": 0.686,
|
| 363 |
+
"step": 470
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"epoch": 0.061436068091642135,
|
| 367 |
+
"grad_norm": 0.6926271319389343,
|
| 368 |
+
"learning_rate": 2.04395135481118e-06,
|
| 369 |
+
"loss": 0.6834,
|
| 370 |
+
"step": 480
|
| 371 |
+
},
|
| 372 |
+
{
|
| 373 |
+
"epoch": 0.06271598617688467,
|
| 374 |
+
"grad_norm": 1.498634934425354,
|
| 375 |
+
"learning_rate": 2.0866225730744615e-06,
|
| 376 |
+
"loss": 0.6805,
|
| 377 |
+
"step": 490
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 0.06399590426212723,
|
| 381 |
+
"grad_norm": 1.1818156242370605,
|
| 382 |
+
"learning_rate": 2.129293791337743e-06,
|
| 383 |
+
"loss": 0.6805,
|
| 384 |
+
"step": 500
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 0.06527582234736977,
|
| 388 |
+
"grad_norm": 1.4476051330566406,
|
| 389 |
+
"learning_rate": 2.1719650096010242e-06,
|
| 390 |
+
"loss": 0.6773,
|
| 391 |
+
"step": 510
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 0.06655574043261231,
|
| 395 |
+
"grad_norm": 1.1143743991851807,
|
| 396 |
+
"learning_rate": 2.214636227864306e-06,
|
| 397 |
+
"loss": 0.6801,
|
| 398 |
+
"step": 520
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 0.06783565851785486,
|
| 402 |
+
"grad_norm": 1.7030247449874878,
|
| 403 |
+
"learning_rate": 2.257307446127587e-06,
|
| 404 |
+
"loss": 0.6794,
|
| 405 |
+
"step": 530
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 0.0691155766030974,
|
| 409 |
+
"grad_norm": 0.8592290282249451,
|
| 410 |
+
"learning_rate": 2.2999786643908685e-06,
|
| 411 |
+
"loss": 0.6736,
|
| 412 |
+
"step": 540
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 0.07039549468833994,
|
| 416 |
+
"grad_norm": 0.8231940865516663,
|
| 417 |
+
"learning_rate": 2.34264988265415e-06,
|
| 418 |
+
"loss": 0.67,
|
| 419 |
+
"step": 550
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 0.0716754127735825,
|
| 423 |
+
"grad_norm": 0.88897705078125,
|
| 424 |
+
"learning_rate": 2.3853211009174317e-06,
|
| 425 |
+
"loss": 0.685,
|
| 426 |
+
"step": 560
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 0.07295533085882504,
|
| 430 |
+
"grad_norm": 1.6297541856765747,
|
| 431 |
+
"learning_rate": 2.427992319180713e-06,
|
| 432 |
+
"loss": 0.6797,
|
| 433 |
+
"step": 570
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 0.07423524894406758,
|
| 437 |
+
"grad_norm": 2.9297447204589844,
|
| 438 |
+
"learning_rate": 2.470663537443994e-06,
|
| 439 |
+
"loss": 0.6766,
|
| 440 |
+
"step": 580
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 0.07551516702931012,
|
| 444 |
+
"grad_norm": 0.9277703762054443,
|
| 445 |
+
"learning_rate": 2.5133347557072756e-06,
|
| 446 |
+
"loss": 0.6824,
|
| 447 |
+
"step": 590
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 0.07679508511455267,
|
| 451 |
+
"grad_norm": 0.5883333683013916,
|
| 452 |
+
"learning_rate": 2.556005973970557e-06,
|
| 453 |
+
"loss": 0.6818,
|
| 454 |
+
"step": 600
|
| 455 |
+
},
|
| 456 |
+
{
|
| 457 |
+
"epoch": 0.07679508511455267,
|
| 458 |
+
"eval_accuracy": 0.63909,
|
| 459 |
+
"eval_loss": 0.6700075268745422,
|
| 460 |
+
"eval_macro_f1": 0.621848143053056,
|
| 461 |
+
"eval_precision": 0.6524313071088819,
|
| 462 |
+
"eval_recall": 0.6305863955136095,
|
| 463 |
+
"eval_runtime": 1633.9732,
|
| 464 |
+
"eval_samples_per_second": 61.201,
|
| 465 |
+
"eval_steps_per_second": 0.479,
|
| 466 |
+
"step": 600
|
| 467 |
+
},
|
| 468 |
+
{
|
| 469 |
+
"epoch": 0.07807500319979521,
|
| 470 |
+
"grad_norm": 0.7149448394775391,
|
| 471 |
+
"learning_rate": 2.5986771922338383e-06,
|
| 472 |
+
"loss": 0.6746,
|
| 473 |
+
"step": 610
|
| 474 |
+
},
|
| 475 |
+
{
|
| 476 |
+
"epoch": 0.07935492128503775,
|
| 477 |
+
"grad_norm": 2.0167856216430664,
|
| 478 |
+
"learning_rate": 2.64134841049712e-06,
|
| 479 |
+
"loss": 0.682,
|
| 480 |
+
"step": 620
|
| 481 |
+
},
|
| 482 |
+
{
|
| 483 |
+
"epoch": 0.0806348393702803,
|
| 484 |
+
"grad_norm": 1.303001880645752,
|
| 485 |
+
"learning_rate": 2.6840196287604015e-06,
|
| 486 |
+
"loss": 0.6735,
|
| 487 |
+
"step": 630
|
| 488 |
+
},
|
| 489 |
+
{
|
| 490 |
+
"epoch": 0.08191475745552285,
|
| 491 |
+
"grad_norm": 2.0516321659088135,
|
| 492 |
+
"learning_rate": 2.7266908470236826e-06,
|
| 493 |
+
"loss": 0.6705,
|
| 494 |
+
"step": 640
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 0.08319467554076539,
|
| 498 |
+
"grad_norm": 2.242105484008789,
|
| 499 |
+
"learning_rate": 2.7693620652869642e-06,
|
| 500 |
+
"loss": 0.6798,
|
| 501 |
+
"step": 650
|
| 502 |
+
},
|
| 503 |
+
{
|
| 504 |
+
"epoch": 0.08447459362600794,
|
| 505 |
+
"grad_norm": 1.1836538314819336,
|
| 506 |
+
"learning_rate": 2.812033283550246e-06,
|
| 507 |
+
"loss": 0.6684,
|
| 508 |
+
"step": 660
|
| 509 |
+
},
|
| 510 |
+
{
|
| 511 |
+
"epoch": 0.08575451171125048,
|
| 512 |
+
"grad_norm": 0.8640874624252319,
|
| 513 |
+
"learning_rate": 2.8547045018135274e-06,
|
| 514 |
+
"loss": 0.6783,
|
| 515 |
+
"step": 670
|
| 516 |
+
},
|
| 517 |
+
{
|
| 518 |
+
"epoch": 0.08703442979649302,
|
| 519 |
+
"grad_norm": 0.6901177763938904,
|
| 520 |
+
"learning_rate": 2.897375720076808e-06,
|
| 521 |
+
"loss": 0.6724,
|
| 522 |
+
"step": 680
|
| 523 |
+
},
|
| 524 |
+
{
|
| 525 |
+
"epoch": 0.08831434788173557,
|
| 526 |
+
"grad_norm": 1.3443949222564697,
|
| 527 |
+
"learning_rate": 2.9400469383400897e-06,
|
| 528 |
+
"loss": 0.67,
|
| 529 |
+
"step": 690
|
| 530 |
+
},
|
| 531 |
+
{
|
| 532 |
+
"epoch": 0.08959426596697811,
|
| 533 |
+
"grad_norm": 0.7775189280509949,
|
| 534 |
+
"learning_rate": 2.9827181566033713e-06,
|
| 535 |
+
"loss": 0.6733,
|
| 536 |
+
"step": 700
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 0.09087418405222066,
|
| 540 |
+
"grad_norm": 1.16916823387146,
|
| 541 |
+
"learning_rate": 3.025389374866653e-06,
|
| 542 |
+
"loss": 0.6713,
|
| 543 |
+
"step": 710
|
| 544 |
+
},
|
| 545 |
+
{
|
| 546 |
+
"epoch": 0.09215410213746321,
|
| 547 |
+
"grad_norm": 1.707152009010315,
|
| 548 |
+
"learning_rate": 3.068060593129934e-06,
|
| 549 |
+
"loss": 0.6659,
|
| 550 |
+
"step": 720
|
| 551 |
+
},
|
| 552 |
+
{
|
| 553 |
+
"epoch": 0.09343402022270575,
|
| 554 |
+
"grad_norm": 0.9780622124671936,
|
| 555 |
+
"learning_rate": 3.1107318113932156e-06,
|
| 556 |
+
"loss": 0.6639,
|
| 557 |
+
"step": 730
|
| 558 |
+
},
|
| 559 |
+
{
|
| 560 |
+
"epoch": 0.09471393830794829,
|
| 561 |
+
"grad_norm": 2.3687374591827393,
|
| 562 |
+
"learning_rate": 3.153403029656497e-06,
|
| 563 |
+
"loss": 0.6666,
|
| 564 |
+
"step": 740
|
| 565 |
+
},
|
| 566 |
+
{
|
| 567 |
+
"epoch": 0.09599385639319083,
|
| 568 |
+
"grad_norm": 0.8506460189819336,
|
| 569 |
+
"learning_rate": 3.1960742479197783e-06,
|
| 570 |
+
"loss": 0.6778,
|
| 571 |
+
"step": 750
|
| 572 |
+
},
|
| 573 |
+
{
|
| 574 |
+
"epoch": 0.09727377447843338,
|
| 575 |
+
"grad_norm": 1.0102107524871826,
|
| 576 |
+
"learning_rate": 3.23874546618306e-06,
|
| 577 |
+
"loss": 0.6635,
|
| 578 |
+
"step": 760
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 0.09855369256367592,
|
| 582 |
+
"grad_norm": 1.1480194330215454,
|
| 583 |
+
"learning_rate": 3.2814166844463415e-06,
|
| 584 |
+
"loss": 0.6709,
|
| 585 |
+
"step": 770
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 0.09983361064891846,
|
| 589 |
+
"grad_norm": 1.3119179010391235,
|
| 590 |
+
"learning_rate": 3.324087902709623e-06,
|
| 591 |
+
"loss": 0.6564,
|
| 592 |
+
"step": 780
|
| 593 |
+
},
|
| 594 |
+
{
|
| 595 |
+
"epoch": 0.10111352873416102,
|
| 596 |
+
"grad_norm": 0.681377649307251,
|
| 597 |
+
"learning_rate": 3.3667591209729038e-06,
|
| 598 |
+
"loss": 0.6656,
|
| 599 |
+
"step": 790
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"epoch": 0.10239344681940356,
|
| 603 |
+
"grad_norm": 1.022257685661316,
|
| 604 |
+
"learning_rate": 3.4094303392361854e-06,
|
| 605 |
+
"loss": 0.6709,
|
| 606 |
+
"step": 800
|
| 607 |
+
},
|
| 608 |
+
{
|
| 609 |
+
"epoch": 0.10239344681940356,
|
| 610 |
+
"eval_accuracy": 0.66247,
|
| 611 |
+
"eval_loss": 0.6552017331123352,
|
| 612 |
+
"eval_macro_f1": 0.6514270557181998,
|
| 613 |
+
"eval_precision": 0.6718270655389242,
|
| 614 |
+
"eval_recall": 0.6556580720913894,
|
| 615 |
+
"eval_runtime": 1633.5478,
|
| 616 |
+
"eval_samples_per_second": 61.216,
|
| 617 |
+
"eval_steps_per_second": 0.479,
|
| 618 |
+
"step": 800
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 0.1036733649046461,
|
| 622 |
+
"grad_norm": 1.0268975496292114,
|
| 623 |
+
"learning_rate": 3.452101557499467e-06,
|
| 624 |
+
"loss": 0.6602,
|
| 625 |
+
"step": 810
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.10495328298988865,
|
| 629 |
+
"grad_norm": 2.6206603050231934,
|
| 630 |
+
"learning_rate": 3.494772775762748e-06,
|
| 631 |
+
"loss": 0.6705,
|
| 632 |
+
"step": 820
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.10623320107513119,
|
| 636 |
+
"grad_norm": 1.035157561302185,
|
| 637 |
+
"learning_rate": 3.5374439940260297e-06,
|
| 638 |
+
"loss": 0.6643,
|
| 639 |
+
"step": 830
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.10751311916037373,
|
| 643 |
+
"grad_norm": 1.736412763595581,
|
| 644 |
+
"learning_rate": 3.5801152122893113e-06,
|
| 645 |
+
"loss": 0.6612,
|
| 646 |
+
"step": 840
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 0.10879303724561629,
|
| 650 |
+
"grad_norm": 0.6906964778900146,
|
| 651 |
+
"learning_rate": 3.622786430552593e-06,
|
| 652 |
+
"loss": 0.6576,
|
| 653 |
+
"step": 850
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.11007295533085883,
|
| 657 |
+
"grad_norm": 0.8148729801177979,
|
| 658 |
+
"learning_rate": 3.665457648815874e-06,
|
| 659 |
+
"loss": 0.6561,
|
| 660 |
+
"step": 860
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 0.11135287341610137,
|
| 664 |
+
"grad_norm": 1.0217723846435547,
|
| 665 |
+
"learning_rate": 3.7081288670791556e-06,
|
| 666 |
+
"loss": 0.6581,
|
| 667 |
+
"step": 870
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.11263279150134392,
|
| 671 |
+
"grad_norm": 1.2743115425109863,
|
| 672 |
+
"learning_rate": 3.750800085342437e-06,
|
| 673 |
+
"loss": 0.6606,
|
| 674 |
+
"step": 880
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.11391270958658646,
|
| 678 |
+
"grad_norm": 0.6405361294746399,
|
| 679 |
+
"learning_rate": 3.793471303605718e-06,
|
| 680 |
+
"loss": 0.652,
|
| 681 |
+
"step": 890
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.115192627671829,
|
| 685 |
+
"grad_norm": 0.8357161283493042,
|
| 686 |
+
"learning_rate": 3.836142521869e-06,
|
| 687 |
+
"loss": 0.6558,
|
| 688 |
+
"step": 900
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 0.11647254575707154,
|
| 692 |
+
"grad_norm": 1.3505173921585083,
|
| 693 |
+
"learning_rate": 3.8788137401322815e-06,
|
| 694 |
+
"loss": 0.6576,
|
| 695 |
+
"step": 910
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.1177524638423141,
|
| 699 |
+
"grad_norm": 1.2253319025039673,
|
| 700 |
+
"learning_rate": 3.921484958395563e-06,
|
| 701 |
+
"loss": 0.6548,
|
| 702 |
+
"step": 920
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 0.11903238192755664,
|
| 706 |
+
"grad_norm": 0.6269568800926208,
|
| 707 |
+
"learning_rate": 3.964156176658844e-06,
|
| 708 |
+
"loss": 0.6562,
|
| 709 |
+
"step": 930
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.12031230001279918,
|
| 713 |
+
"grad_norm": 0.9413424134254456,
|
| 714 |
+
"learning_rate": 4.006827394922125e-06,
|
| 715 |
+
"loss": 0.6461,
|
| 716 |
+
"step": 940
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.12159221809804173,
|
| 720 |
+
"grad_norm": 1.0903513431549072,
|
| 721 |
+
"learning_rate": 4.049498613185407e-06,
|
| 722 |
+
"loss": 0.6457,
|
| 723 |
+
"step": 950
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.12287213618328427,
|
| 727 |
+
"grad_norm": 2.0023152828216553,
|
| 728 |
+
"learning_rate": 4.092169831448688e-06,
|
| 729 |
+
"loss": 0.6458,
|
| 730 |
+
"step": 960
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 0.12415205426852681,
|
| 734 |
+
"grad_norm": 0.9638109803199768,
|
| 735 |
+
"learning_rate": 4.134841049711969e-06,
|
| 736 |
+
"loss": 0.6486,
|
| 737 |
+
"step": 970
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.12543197235376935,
|
| 741 |
+
"grad_norm": 0.9894546270370483,
|
| 742 |
+
"learning_rate": 4.177512267975251e-06,
|
| 743 |
+
"loss": 0.6453,
|
| 744 |
+
"step": 980
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 0.1267118904390119,
|
| 748 |
+
"grad_norm": 1.2289057970046997,
|
| 749 |
+
"learning_rate": 4.220183486238532e-06,
|
| 750 |
+
"loss": 0.6605,
|
| 751 |
+
"step": 990
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.12799180852425446,
|
| 755 |
+
"grad_norm": 1.119125247001648,
|
| 756 |
+
"learning_rate": 4.262854704501814e-06,
|
| 757 |
+
"loss": 0.6564,
|
| 758 |
+
"step": 1000
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 0.12799180852425446,
|
| 762 |
+
"eval_accuracy": 0.69068,
|
| 763 |
+
"eval_loss": 0.6396247148513794,
|
| 764 |
+
"eval_macro_f1": 0.689588440856387,
|
| 765 |
+
"eval_precision": 0.6900076024604327,
|
| 766 |
+
"eval_recall": 0.6894114140312133,
|
| 767 |
+
"eval_runtime": 1634.4321,
|
| 768 |
+
"eval_samples_per_second": 61.183,
|
| 769 |
+
"eval_steps_per_second": 0.478,
|
| 770 |
+
"step": 1000
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 0.12927172660949698,
|
| 774 |
+
"grad_norm": 2.02966046333313,
|
| 775 |
+
"learning_rate": 4.3055259227650956e-06,
|
| 776 |
+
"loss": 0.646,
|
| 777 |
+
"step": 1010
|
| 778 |
+
},
|
| 779 |
+
{
|
| 780 |
+
"epoch": 0.13055164469473954,
|
| 781 |
+
"grad_norm": 0.7430800199508667,
|
| 782 |
+
"learning_rate": 4.348197141028377e-06,
|
| 783 |
+
"loss": 0.6453,
|
| 784 |
+
"step": 1020
|
| 785 |
+
},
|
| 786 |
+
{
|
| 787 |
+
"epoch": 0.1318315627799821,
|
| 788 |
+
"grad_norm": 1.031962275505066,
|
| 789 |
+
"learning_rate": 4.390868359291658e-06,
|
| 790 |
+
"loss": 0.6425,
|
| 791 |
+
"step": 1030
|
| 792 |
+
},
|
| 793 |
+
{
|
| 794 |
+
"epoch": 0.13311148086522462,
|
| 795 |
+
"grad_norm": 1.0249106884002686,
|
| 796 |
+
"learning_rate": 4.4335395775549394e-06,
|
| 797 |
+
"loss": 0.6426,
|
| 798 |
+
"step": 1040
|
| 799 |
+
},
|
| 800 |
+
{
|
| 801 |
+
"epoch": 0.13439139895046717,
|
| 802 |
+
"grad_norm": 0.7375448942184448,
|
| 803 |
+
"learning_rate": 4.476210795818221e-06,
|
| 804 |
+
"loss": 0.6401,
|
| 805 |
+
"step": 1050
|
| 806 |
+
},
|
| 807 |
+
{
|
| 808 |
+
"epoch": 0.13567131703570973,
|
| 809 |
+
"grad_norm": 1.295983076095581,
|
| 810 |
+
"learning_rate": 4.518882014081503e-06,
|
| 811 |
+
"loss": 0.6449,
|
| 812 |
+
"step": 1060
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 0.13695123512095225,
|
| 816 |
+
"grad_norm": 1.6978998184204102,
|
| 817 |
+
"learning_rate": 4.561553232344783e-06,
|
| 818 |
+
"loss": 0.6484,
|
| 819 |
+
"step": 1070
|
| 820 |
+
},
|
| 821 |
+
{
|
| 822 |
+
"epoch": 0.1382311532061948,
|
| 823 |
+
"grad_norm": 0.7747512459754944,
|
| 824 |
+
"learning_rate": 4.604224450608065e-06,
|
| 825 |
+
"loss": 0.6526,
|
| 826 |
+
"step": 1080
|
| 827 |
+
},
|
| 828 |
+
{
|
| 829 |
+
"epoch": 0.13951107129143736,
|
| 830 |
+
"grad_norm": 0.7895913124084473,
|
| 831 |
+
"learning_rate": 4.6468956688713465e-06,
|
| 832 |
+
"loss": 0.6311,
|
| 833 |
+
"step": 1090
|
| 834 |
+
},
|
| 835 |
+
{
|
| 836 |
+
"epoch": 0.1407909893766799,
|
| 837 |
+
"grad_norm": 0.6975072026252747,
|
| 838 |
+
"learning_rate": 4.689566887134628e-06,
|
| 839 |
+
"loss": 0.6454,
|
| 840 |
+
"step": 1100
|
| 841 |
+
},
|
| 842 |
+
{
|
| 843 |
+
"epoch": 0.14207090746192244,
|
| 844 |
+
"grad_norm": 0.8316811919212341,
|
| 845 |
+
"learning_rate": 4.73223810539791e-06,
|
| 846 |
+
"loss": 0.6447,
|
| 847 |
+
"step": 1110
|
| 848 |
+
},
|
| 849 |
+
{
|
| 850 |
+
"epoch": 0.143350825547165,
|
| 851 |
+
"grad_norm": 0.6887441277503967,
|
| 852 |
+
"learning_rate": 4.774909323661191e-06,
|
| 853 |
+
"loss": 0.6325,
|
| 854 |
+
"step": 1120
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 0.14463074363240752,
|
| 858 |
+
"grad_norm": 1.0218420028686523,
|
| 859 |
+
"learning_rate": 4.817580541924473e-06,
|
| 860 |
+
"loss": 0.6308,
|
| 861 |
+
"step": 1130
|
| 862 |
+
},
|
| 863 |
+
{
|
| 864 |
+
"epoch": 0.14591066171765008,
|
| 865 |
+
"grad_norm": 0.7663633823394775,
|
| 866 |
+
"learning_rate": 4.8602517601877535e-06,
|
| 867 |
+
"loss": 0.6227,
|
| 868 |
+
"step": 1140
|
| 869 |
+
},
|
| 870 |
+
{
|
| 871 |
+
"epoch": 0.1471905798028926,
|
| 872 |
+
"grad_norm": 1.5115224123001099,
|
| 873 |
+
"learning_rate": 4.902922978451035e-06,
|
| 874 |
+
"loss": 0.6315,
|
| 875 |
+
"step": 1150
|
| 876 |
+
},
|
| 877 |
+
{
|
| 878 |
+
"epoch": 0.14847049788813516,
|
| 879 |
+
"grad_norm": 0.8011078238487244,
|
| 880 |
+
"learning_rate": 4.945594196714317e-06,
|
| 881 |
+
"loss": 0.6428,
|
| 882 |
+
"step": 1160
|
| 883 |
+
},
|
| 884 |
+
{
|
| 885 |
+
"epoch": 0.1497504159733777,
|
| 886 |
+
"grad_norm": 1.3246116638183594,
|
| 887 |
+
"learning_rate": 4.9882654149775974e-06,
|
| 888 |
+
"loss": 0.6397,
|
| 889 |
+
"step": 1170
|
| 890 |
+
},
|
| 891 |
+
{
|
| 892 |
+
"epoch": 0.15103033405862024,
|
| 893 |
+
"grad_norm": 0.979498565196991,
|
| 894 |
+
"learning_rate": 5.030936633240879e-06,
|
| 895 |
+
"loss": 0.6308,
|
| 896 |
+
"step": 1180
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 0.1523102521438628,
|
| 900 |
+
"grad_norm": 0.8756135106086731,
|
| 901 |
+
"learning_rate": 5.073607851504161e-06,
|
| 902 |
+
"loss": 0.6373,
|
| 903 |
+
"step": 1190
|
| 904 |
+
},
|
| 905 |
+
{
|
| 906 |
+
"epoch": 0.15359017022910534,
|
| 907 |
+
"grad_norm": 0.845967710018158,
|
| 908 |
+
"learning_rate": 5.116279069767442e-06,
|
| 909 |
+
"loss": 0.6322,
|
| 910 |
+
"step": 1200
|
| 911 |
+
},
|
| 912 |
+
{
|
| 913 |
+
"epoch": 0.15359017022910534,
|
| 914 |
+
"eval_accuracy": 0.69931,
|
| 915 |
+
"eval_loss": 0.6224722862243652,
|
| 916 |
+
"eval_macro_f1": 0.6985627236986651,
|
| 917 |
+
"eval_precision": 0.6986454338713279,
|
| 918 |
+
"eval_recall": 0.6984991383538209,
|
| 919 |
+
"eval_runtime": 692.1526,
|
| 920 |
+
"eval_samples_per_second": 144.477,
|
| 921 |
+
"eval_steps_per_second": 2.258,
|
| 922 |
+
"step": 1200
|
| 923 |
+
},
|
| 924 |
+
{
|
| 925 |
+
"epoch": 0.15487008831434787,
|
| 926 |
+
"grad_norm": 2.0670483112335205,
|
| 927 |
+
"learning_rate": 1.032010243277849e-05,
|
| 928 |
+
"loss": 0.6386,
|
| 929 |
+
"step": 1210
|
| 930 |
+
},
|
| 931 |
+
{
|
| 932 |
+
"epoch": 0.15615000639959042,
|
| 933 |
+
"grad_norm": 2.482168436050415,
|
| 934 |
+
"learning_rate": 1.0405463081519421e-05,
|
| 935 |
+
"loss": 0.6281,
|
| 936 |
+
"step": 1220
|
| 937 |
+
},
|
| 938 |
+
{
|
| 939 |
+
"epoch": 0.15742992448483298,
|
| 940 |
+
"grad_norm": 1.4530706405639648,
|
| 941 |
+
"learning_rate": 1.049082373026035e-05,
|
| 942 |
+
"loss": 0.6277,
|
| 943 |
+
"step": 1230
|
| 944 |
+
},
|
| 945 |
+
{
|
| 946 |
+
"epoch": 0.1587098425700755,
|
| 947 |
+
"grad_norm": 2.9233577251434326,
|
| 948 |
+
"learning_rate": 1.0576184379001282e-05,
|
| 949 |
+
"loss": 0.6363,
|
| 950 |
+
"step": 1240
|
| 951 |
+
},
|
| 952 |
+
{
|
| 953 |
+
"epoch": 0.15998976065531806,
|
| 954 |
+
"grad_norm": 0.6975642442703247,
|
| 955 |
+
"learning_rate": 1.0661545027742212e-05,
|
| 956 |
+
"loss": 0.6344,
|
| 957 |
+
"step": 1250
|
| 958 |
+
},
|
| 959 |
+
{
|
| 960 |
+
"epoch": 0.1612696787405606,
|
| 961 |
+
"grad_norm": 2.3634731769561768,
|
| 962 |
+
"learning_rate": 1.0746905676483142e-05,
|
| 963 |
+
"loss": 0.6217,
|
| 964 |
+
"step": 1260
|
| 965 |
+
},
|
| 966 |
+
{
|
| 967 |
+
"epoch": 0.16254959682580314,
|
| 968 |
+
"grad_norm": 0.8812585473060608,
|
| 969 |
+
"learning_rate": 1.0832266325224072e-05,
|
| 970 |
+
"loss": 0.6311,
|
| 971 |
+
"step": 1270
|
| 972 |
+
},
|
| 973 |
+
{
|
| 974 |
+
"epoch": 0.1638295149110457,
|
| 975 |
+
"grad_norm": 1.8986798524856567,
|
| 976 |
+
"learning_rate": 1.0917626973965003e-05,
|
| 977 |
+
"loss": 0.6304,
|
| 978 |
+
"step": 1280
|
| 979 |
+
},
|
| 980 |
+
{
|
| 981 |
+
"epoch": 0.16510943299628825,
|
| 982 |
+
"grad_norm": 0.8174984455108643,
|
| 983 |
+
"learning_rate": 1.1002987622705935e-05,
|
| 984 |
+
"loss": 0.6176,
|
| 985 |
+
"step": 1290
|
| 986 |
+
},
|
| 987 |
+
{
|
| 988 |
+
"epoch": 0.16638935108153077,
|
| 989 |
+
"grad_norm": 1.0268642902374268,
|
| 990 |
+
"learning_rate": 1.1088348271446865e-05,
|
| 991 |
+
"loss": 0.6233,
|
| 992 |
+
"step": 1300
|
| 993 |
+
},
|
| 994 |
+
{
|
| 995 |
+
"epoch": 0.16766926916677333,
|
| 996 |
+
"grad_norm": 1.047821044921875,
|
| 997 |
+
"learning_rate": 1.1173708920187793e-05,
|
| 998 |
+
"loss": 0.6127,
|
| 999 |
+
"step": 1310
|
| 1000 |
+
},
|
| 1001 |
+
{
|
| 1002 |
+
"epoch": 0.16894918725201588,
|
| 1003 |
+
"grad_norm": 0.9875057935714722,
|
| 1004 |
+
"learning_rate": 1.1259069568928724e-05,
|
| 1005 |
+
"loss": 0.6211,
|
| 1006 |
+
"step": 1320
|
| 1007 |
+
},
|
| 1008 |
+
{
|
| 1009 |
+
"epoch": 0.1702291053372584,
|
| 1010 |
+
"grad_norm": 1.0082708597183228,
|
| 1011 |
+
"learning_rate": 1.1344430217669656e-05,
|
| 1012 |
+
"loss": 0.6202,
|
| 1013 |
+
"step": 1330
|
| 1014 |
+
},
|
| 1015 |
+
{
|
| 1016 |
+
"epoch": 0.17150902342250096,
|
| 1017 |
+
"grad_norm": 0.7229795455932617,
|
| 1018 |
+
"learning_rate": 1.1429790866410586e-05,
|
| 1019 |
+
"loss": 0.6071,
|
| 1020 |
+
"step": 1340
|
| 1021 |
+
},
|
| 1022 |
+
{
|
| 1023 |
+
"epoch": 0.17278894150774352,
|
| 1024 |
+
"grad_norm": 0.7687681317329407,
|
| 1025 |
+
"learning_rate": 1.1515151515151517e-05,
|
| 1026 |
+
"loss": 0.6218,
|
| 1027 |
+
"step": 1350
|
| 1028 |
+
},
|
| 1029 |
+
{
|
| 1030 |
+
"epoch": 0.17406885959298604,
|
| 1031 |
+
"grad_norm": 1.1643378734588623,
|
| 1032 |
+
"learning_rate": 1.1600512163892445e-05,
|
| 1033 |
+
"loss": 0.6208,
|
| 1034 |
+
"step": 1360
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 0.1753487776782286,
|
| 1038 |
+
"grad_norm": 0.7338509559631348,
|
| 1039 |
+
"learning_rate": 1.1685872812633377e-05,
|
| 1040 |
+
"loss": 0.6105,
|
| 1041 |
+
"step": 1370
|
| 1042 |
+
},
|
| 1043 |
+
{
|
| 1044 |
+
"epoch": 0.17662869576347115,
|
| 1045 |
+
"grad_norm": 0.7444656491279602,
|
| 1046 |
+
"learning_rate": 1.1771233461374307e-05,
|
| 1047 |
+
"loss": 0.6183,
|
| 1048 |
+
"step": 1380
|
| 1049 |
+
},
|
| 1050 |
+
{
|
| 1051 |
+
"epoch": 0.17790861384871368,
|
| 1052 |
+
"grad_norm": 1.1902283430099487,
|
| 1053 |
+
"learning_rate": 1.1856594110115238e-05,
|
| 1054 |
+
"loss": 0.6209,
|
| 1055 |
+
"step": 1390
|
| 1056 |
+
},
|
| 1057 |
+
{
|
| 1058 |
+
"epoch": 0.17918853193395623,
|
| 1059 |
+
"grad_norm": 1.0872228145599365,
|
| 1060 |
+
"learning_rate": 1.194195475885617e-05,
|
| 1061 |
+
"loss": 0.6172,
|
| 1062 |
+
"step": 1400
|
| 1063 |
+
},
|
| 1064 |
+
{
|
| 1065 |
+
"epoch": 0.18046845001919878,
|
| 1066 |
+
"grad_norm": 1.8439041376113892,
|
| 1067 |
+
"learning_rate": 1.2027315407597098e-05,
|
| 1068 |
+
"loss": 0.6086,
|
| 1069 |
+
"step": 1410
|
| 1070 |
+
},
|
| 1071 |
+
{
|
| 1072 |
+
"epoch": 0.1817483681044413,
|
| 1073 |
+
"grad_norm": 0.8699899911880493,
|
| 1074 |
+
"learning_rate": 1.211267605633803e-05,
|
| 1075 |
+
"loss": 0.6247,
|
| 1076 |
+
"step": 1420
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 0.18302828618968386,
|
| 1080 |
+
"grad_norm": 1.1468650102615356,
|
| 1081 |
+
"learning_rate": 1.2198036705078959e-05,
|
| 1082 |
+
"loss": 0.6114,
|
| 1083 |
+
"step": 1430
|
| 1084 |
+
},
|
| 1085 |
+
{
|
| 1086 |
+
"epoch": 0.18430820427492642,
|
| 1087 |
+
"grad_norm": 1.0204615592956543,
|
| 1088 |
+
"learning_rate": 1.228339735381989e-05,
|
| 1089 |
+
"loss": 0.6015,
|
| 1090 |
+
"step": 1440
|
| 1091 |
+
},
|
| 1092 |
+
{
|
| 1093 |
+
"epoch": 0.18558812236016894,
|
| 1094 |
+
"grad_norm": 2.660583257675171,
|
| 1095 |
+
"learning_rate": 1.236875800256082e-05,
|
| 1096 |
+
"loss": 0.6002,
|
| 1097 |
+
"step": 1450
|
| 1098 |
+
},
|
| 1099 |
+
{
|
| 1100 |
+
"epoch": 0.1868680404454115,
|
| 1101 |
+
"grad_norm": 1.1916371583938599,
|
| 1102 |
+
"learning_rate": 1.2454118651301752e-05,
|
| 1103 |
+
"loss": 0.6076,
|
| 1104 |
+
"step": 1460
|
| 1105 |
+
},
|
| 1106 |
+
{
|
| 1107 |
+
"epoch": 0.18814795853065402,
|
| 1108 |
+
"grad_norm": 0.8097987174987793,
|
| 1109 |
+
"learning_rate": 1.253947930004268e-05,
|
| 1110 |
+
"loss": 0.6126,
|
| 1111 |
+
"step": 1470
|
| 1112 |
+
},
|
| 1113 |
+
{
|
| 1114 |
+
"epoch": 0.18942787661589658,
|
| 1115 |
+
"grad_norm": 1.978747010231018,
|
| 1116 |
+
"learning_rate": 1.2624839948783612e-05,
|
| 1117 |
+
"loss": 0.5962,
|
| 1118 |
+
"step": 1480
|
| 1119 |
+
},
|
| 1120 |
+
{
|
| 1121 |
+
"epoch": 0.19070779470113913,
|
| 1122 |
+
"grad_norm": 1.0975425243377686,
|
| 1123 |
+
"learning_rate": 1.2710200597524543e-05,
|
| 1124 |
+
"loss": 0.587,
|
| 1125 |
+
"step": 1490
|
| 1126 |
+
},
|
| 1127 |
+
{
|
| 1128 |
+
"epoch": 0.19198771278638166,
|
| 1129 |
+
"grad_norm": 2.150351047515869,
|
| 1130 |
+
"learning_rate": 1.2795561246265473e-05,
|
| 1131 |
+
"loss": 0.5962,
|
| 1132 |
+
"step": 1500
|
| 1133 |
+
},
|
| 1134 |
+
{
|
| 1135 |
+
"epoch": 0.19198771278638166,
|
| 1136 |
+
"eval_accuracy": 0.72526,
|
| 1137 |
+
"eval_loss": 0.5823243856430054,
|
| 1138 |
+
"eval_macro_f1": 0.7252418839884757,
|
| 1139 |
+
"eval_precision": 0.7275813809678697,
|
| 1140 |
+
"eval_recall": 0.7271797416587706,
|
| 1141 |
+
"eval_runtime": 708.4119,
|
| 1142 |
+
"eval_samples_per_second": 141.161,
|
| 1143 |
+
"eval_steps_per_second": 1.104,
|
| 1144 |
+
"step": 1500
|
| 1145 |
+
},
|
| 1146 |
+
{
|
| 1147 |
+
"epoch": 0.1932676308716242,
|
| 1148 |
+
"grad_norm": 1.8355706930160522,
|
| 1149 |
+
"learning_rate": 1.2880921895006405e-05,
|
| 1150 |
+
"loss": 0.6133,
|
| 1151 |
+
"step": 1510
|
| 1152 |
+
},
|
| 1153 |
+
{
|
| 1154 |
+
"epoch": 0.19454754895686677,
|
| 1155 |
+
"grad_norm": 1.4349753856658936,
|
| 1156 |
+
"learning_rate": 1.2966282543747333e-05,
|
| 1157 |
+
"loss": 0.6062,
|
| 1158 |
+
"step": 1520
|
| 1159 |
+
},
|
| 1160 |
+
{
|
| 1161 |
+
"epoch": 0.1958274670421093,
|
| 1162 |
+
"grad_norm": 1.8322596549987793,
|
| 1163 |
+
"learning_rate": 1.3051643192488264e-05,
|
| 1164 |
+
"loss": 0.5982,
|
| 1165 |
+
"step": 1530
|
| 1166 |
+
},
|
| 1167 |
+
{
|
| 1168 |
+
"epoch": 0.19710738512735185,
|
| 1169 |
+
"grad_norm": 1.509338617324829,
|
| 1170 |
+
"learning_rate": 1.3137003841229194e-05,
|
| 1171 |
+
"loss": 0.5961,
|
| 1172 |
+
"step": 1540
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 0.1983873032125944,
|
| 1176 |
+
"grad_norm": 0.9844911098480225,
|
| 1177 |
+
"learning_rate": 1.3222364489970126e-05,
|
| 1178 |
+
"loss": 0.583,
|
| 1179 |
+
"step": 1550
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"epoch": 0.19966722129783693,
|
| 1183 |
+
"grad_norm": 1.035900354385376,
|
| 1184 |
+
"learning_rate": 1.3307725138711055e-05,
|
| 1185 |
+
"loss": 0.598,
|
| 1186 |
+
"step": 1560
|
| 1187 |
+
},
|
| 1188 |
+
{
|
| 1189 |
+
"epoch": 0.20094713938307948,
|
| 1190 |
+
"grad_norm": 0.7891924977302551,
|
| 1191 |
+
"learning_rate": 1.3393085787451985e-05,
|
| 1192 |
+
"loss": 0.6064,
|
| 1193 |
+
"step": 1570
|
| 1194 |
+
},
|
| 1195 |
+
{
|
| 1196 |
+
"epoch": 0.20222705746832204,
|
| 1197 |
+
"grad_norm": 1.4326112270355225,
|
| 1198 |
+
"learning_rate": 1.3478446436192915e-05,
|
| 1199 |
+
"loss": 0.5954,
|
| 1200 |
+
"step": 1580
|
| 1201 |
+
},
|
| 1202 |
+
{
|
| 1203 |
+
"epoch": 0.20350697555356456,
|
| 1204 |
+
"grad_norm": 0.9054266214370728,
|
| 1205 |
+
"learning_rate": 1.3563807084933847e-05,
|
| 1206 |
+
"loss": 0.59,
|
| 1207 |
+
"step": 1590
|
| 1208 |
+
},
|
| 1209 |
+
{
|
| 1210 |
+
"epoch": 0.20478689363880712,
|
| 1211 |
+
"grad_norm": 1.0216469764709473,
|
| 1212 |
+
"learning_rate": 1.3649167733674778e-05,
|
| 1213 |
+
"loss": 0.5845,
|
| 1214 |
+
"step": 1600
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 0.20606681172404967,
|
| 1218 |
+
"grad_norm": 1.0591177940368652,
|
| 1219 |
+
"learning_rate": 1.3734528382415708e-05,
|
| 1220 |
+
"loss": 0.599,
|
| 1221 |
+
"step": 1610
|
| 1222 |
+
},
|
| 1223 |
+
{
|
| 1224 |
+
"epoch": 0.2073467298092922,
|
| 1225 |
+
"grad_norm": 0.7826931476593018,
|
| 1226 |
+
"learning_rate": 1.3819889031156638e-05,
|
| 1227 |
+
"loss": 0.5898,
|
| 1228 |
+
"step": 1620
|
| 1229 |
+
},
|
| 1230 |
+
{
|
| 1231 |
+
"epoch": 0.20862664789453475,
|
| 1232 |
+
"grad_norm": 0.755007803440094,
|
| 1233 |
+
"learning_rate": 1.3905249679897568e-05,
|
| 1234 |
+
"loss": 0.5769,
|
| 1235 |
+
"step": 1630
|
| 1236 |
+
},
|
| 1237 |
+
{
|
| 1238 |
+
"epoch": 0.2099065659797773,
|
| 1239 |
+
"grad_norm": 1.3526090383529663,
|
| 1240 |
+
"learning_rate": 1.3990610328638499e-05,
|
| 1241 |
+
"loss": 0.5819,
|
| 1242 |
+
"step": 1640
|
| 1243 |
+
},
|
| 1244 |
+
{
|
| 1245 |
+
"epoch": 0.21118648406501983,
|
| 1246 |
+
"grad_norm": 0.7078200578689575,
|
| 1247 |
+
"learning_rate": 1.4075970977379429e-05,
|
| 1248 |
+
"loss": 0.5884,
|
| 1249 |
+
"step": 1650
|
| 1250 |
+
},
|
| 1251 |
+
{
|
| 1252 |
+
"epoch": 0.21246640215026238,
|
| 1253 |
+
"grad_norm": 0.80736243724823,
|
| 1254 |
+
"learning_rate": 1.416133162612036e-05,
|
| 1255 |
+
"loss": 0.5991,
|
| 1256 |
+
"step": 1660
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 0.21374632023550494,
|
| 1260 |
+
"grad_norm": 0.6882274150848389,
|
| 1261 |
+
"learning_rate": 1.4246692274861289e-05,
|
| 1262 |
+
"loss": 0.5825,
|
| 1263 |
+
"step": 1670
|
| 1264 |
+
},
|
| 1265 |
+
{
|
| 1266 |
+
"epoch": 0.21502623832074746,
|
| 1267 |
+
"grad_norm": 0.7223654985427856,
|
| 1268 |
+
"learning_rate": 1.433205292360222e-05,
|
| 1269 |
+
"loss": 0.5689,
|
| 1270 |
+
"step": 1680
|
| 1271 |
+
},
|
| 1272 |
+
{
|
| 1273 |
+
"epoch": 0.21630615640599002,
|
| 1274 |
+
"grad_norm": 0.6446322798728943,
|
| 1275 |
+
"learning_rate": 1.4417413572343152e-05,
|
| 1276 |
+
"loss": 0.5773,
|
| 1277 |
+
"step": 1690
|
| 1278 |
+
},
|
| 1279 |
+
{
|
| 1280 |
+
"epoch": 0.21758607449123257,
|
| 1281 |
+
"grad_norm": 1.505007266998291,
|
| 1282 |
+
"learning_rate": 1.4502774221084081e-05,
|
| 1283 |
+
"loss": 0.5677,
|
| 1284 |
+
"step": 1700
|
| 1285 |
+
},
|
| 1286 |
+
{
|
| 1287 |
+
"epoch": 0.2188659925764751,
|
| 1288 |
+
"grad_norm": 1.1939284801483154,
|
| 1289 |
+
"learning_rate": 1.4588134869825013e-05,
|
| 1290 |
+
"loss": 0.5805,
|
| 1291 |
+
"step": 1710
|
| 1292 |
+
},
|
| 1293 |
+
{
|
| 1294 |
+
"epoch": 0.22014591066171765,
|
| 1295 |
+
"grad_norm": 2.848264217376709,
|
| 1296 |
+
"learning_rate": 1.4673495518565943e-05,
|
| 1297 |
+
"loss": 0.5712,
|
| 1298 |
+
"step": 1720
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 0.2214258287469602,
|
| 1302 |
+
"grad_norm": 1.2117749452590942,
|
| 1303 |
+
"learning_rate": 1.4758856167306873e-05,
|
| 1304 |
+
"loss": 0.582,
|
| 1305 |
+
"step": 1730
|
| 1306 |
+
},
|
| 1307 |
+
{
|
| 1308 |
+
"epoch": 0.22270574683220273,
|
| 1309 |
+
"grad_norm": 2.2297441959381104,
|
| 1310 |
+
"learning_rate": 1.4844216816047802e-05,
|
| 1311 |
+
"loss": 0.5751,
|
| 1312 |
+
"step": 1740
|
| 1313 |
+
},
|
| 1314 |
+
{
|
| 1315 |
+
"epoch": 0.2239856649174453,
|
| 1316 |
+
"grad_norm": 1.211978554725647,
|
| 1317 |
+
"learning_rate": 1.4929577464788734e-05,
|
| 1318 |
+
"loss": 0.5607,
|
| 1319 |
+
"step": 1750
|
| 1320 |
+
},
|
| 1321 |
+
{
|
| 1322 |
+
"epoch": 0.22526558300268784,
|
| 1323 |
+
"grad_norm": 0.8025451898574829,
|
| 1324 |
+
"learning_rate": 1.5014938113529664e-05,
|
| 1325 |
+
"loss": 0.5803,
|
| 1326 |
+
"step": 1760
|
| 1327 |
+
},
|
| 1328 |
+
{
|
| 1329 |
+
"epoch": 0.22654550108793037,
|
| 1330 |
+
"grad_norm": 1.2300900220870972,
|
| 1331 |
+
"learning_rate": 1.5100298762270595e-05,
|
| 1332 |
+
"loss": 0.5827,
|
| 1333 |
+
"step": 1770
|
| 1334 |
+
},
|
| 1335 |
+
{
|
| 1336 |
+
"epoch": 0.22782541917317292,
|
| 1337 |
+
"grad_norm": 1.0382864475250244,
|
| 1338 |
+
"learning_rate": 1.5185659411011523e-05,
|
| 1339 |
+
"loss": 0.5916,
|
| 1340 |
+
"step": 1780
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 0.22910533725841545,
|
| 1344 |
+
"grad_norm": 1.4876656532287598,
|
| 1345 |
+
"learning_rate": 1.5271020059752453e-05,
|
| 1346 |
+
"loss": 0.5536,
|
| 1347 |
+
"step": 1790
|
| 1348 |
+
},
|
| 1349 |
+
{
|
| 1350 |
+
"epoch": 0.230385255343658,
|
| 1351 |
+
"grad_norm": 1.1409252882003784,
|
| 1352 |
+
"learning_rate": 1.5356380708493386e-05,
|
| 1353 |
+
"loss": 0.5484,
|
| 1354 |
+
"step": 1800
|
| 1355 |
+
},
|
| 1356 |
+
{
|
| 1357 |
+
"epoch": 0.23166517342890056,
|
| 1358 |
+
"grad_norm": 0.7180280089378357,
|
| 1359 |
+
"learning_rate": 1.5441741357234316e-05,
|
| 1360 |
+
"loss": 0.5712,
|
| 1361 |
+
"step": 1810
|
| 1362 |
+
},
|
| 1363 |
+
{
|
| 1364 |
+
"epoch": 0.23294509151414308,
|
| 1365 |
+
"grad_norm": 1.5890289545059204,
|
| 1366 |
+
"learning_rate": 1.5527102005975246e-05,
|
| 1367 |
+
"loss": 0.5758,
|
| 1368 |
+
"step": 1820
|
| 1369 |
+
},
|
| 1370 |
+
{
|
| 1371 |
+
"epoch": 0.23422500959938564,
|
| 1372 |
+
"grad_norm": 0.6635769605636597,
|
| 1373 |
+
"learning_rate": 1.5612462654716176e-05,
|
| 1374 |
+
"loss": 0.5543,
|
| 1375 |
+
"step": 1830
|
| 1376 |
+
},
|
| 1377 |
+
{
|
| 1378 |
+
"epoch": 0.2355049276846282,
|
| 1379 |
+
"grad_norm": 0.8864602446556091,
|
| 1380 |
+
"learning_rate": 1.5697823303457106e-05,
|
| 1381 |
+
"loss": 0.5536,
|
| 1382 |
+
"step": 1840
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 0.23678484576987072,
|
| 1386 |
+
"grad_norm": 0.8372675776481628,
|
| 1387 |
+
"learning_rate": 1.578318395219804e-05,
|
| 1388 |
+
"loss": 0.5743,
|
| 1389 |
+
"step": 1850
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"epoch": 0.23806476385511327,
|
| 1393 |
+
"grad_norm": 0.7532811164855957,
|
| 1394 |
+
"learning_rate": 1.586854460093897e-05,
|
| 1395 |
+
"loss": 0.5593,
|
| 1396 |
+
"step": 1860
|
| 1397 |
+
},
|
| 1398 |
+
{
|
| 1399 |
+
"epoch": 0.23934468194035582,
|
| 1400 |
+
"grad_norm": 0.7231799960136414,
|
| 1401 |
+
"learning_rate": 1.59539052496799e-05,
|
| 1402 |
+
"loss": 0.5831,
|
| 1403 |
+
"step": 1870
|
| 1404 |
+
},
|
| 1405 |
+
{
|
| 1406 |
+
"epoch": 0.24062460002559835,
|
| 1407 |
+
"grad_norm": 1.124228596687317,
|
| 1408 |
+
"learning_rate": 1.603926589842083e-05,
|
| 1409 |
+
"loss": 0.5936,
|
| 1410 |
+
"step": 1880
|
| 1411 |
+
},
|
| 1412 |
+
{
|
| 1413 |
+
"epoch": 0.2419045181108409,
|
| 1414 |
+
"grad_norm": 1.0107067823410034,
|
| 1415 |
+
"learning_rate": 1.612462654716176e-05,
|
| 1416 |
+
"loss": 0.5598,
|
| 1417 |
+
"step": 1890
|
| 1418 |
+
},
|
| 1419 |
+
{
|
| 1420 |
+
"epoch": 0.24318443619608346,
|
| 1421 |
+
"grad_norm": 1.6117154359817505,
|
| 1422 |
+
"learning_rate": 1.620998719590269e-05,
|
| 1423 |
+
"loss": 0.5605,
|
| 1424 |
+
"step": 1900
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 0.24446435428132599,
|
| 1428 |
+
"grad_norm": 1.5564496517181396,
|
| 1429 |
+
"learning_rate": 1.629534784464362e-05,
|
| 1430 |
+
"loss": 0.5666,
|
| 1431 |
+
"step": 1910
|
| 1432 |
+
},
|
| 1433 |
+
{
|
| 1434 |
+
"epoch": 0.24574427236656854,
|
| 1435 |
+
"grad_norm": 0.8275917768478394,
|
| 1436 |
+
"learning_rate": 1.638070849338455e-05,
|
| 1437 |
+
"loss": 0.5597,
|
| 1438 |
+
"step": 1920
|
| 1439 |
+
},
|
| 1440 |
+
{
|
| 1441 |
+
"epoch": 0.2470241904518111,
|
| 1442 |
+
"grad_norm": 0.8956324458122253,
|
| 1443 |
+
"learning_rate": 1.646606914212548e-05,
|
| 1444 |
+
"loss": 0.5453,
|
| 1445 |
+
"step": 1930
|
| 1446 |
+
},
|
| 1447 |
+
{
|
| 1448 |
+
"epoch": 0.24830410853705362,
|
| 1449 |
+
"grad_norm": 1.9644063711166382,
|
| 1450 |
+
"learning_rate": 1.655142979086641e-05,
|
| 1451 |
+
"loss": 0.5732,
|
| 1452 |
+
"step": 1940
|
| 1453 |
+
},
|
| 1454 |
+
{
|
| 1455 |
+
"epoch": 0.24958402662229617,
|
| 1456 |
+
"grad_norm": 0.7495090365409851,
|
| 1457 |
+
"learning_rate": 1.663679043960734e-05,
|
| 1458 |
+
"loss": 0.5549,
|
| 1459 |
+
"step": 1950
|
| 1460 |
+
},
|
| 1461 |
+
{
|
| 1462 |
+
"epoch": 0.2508639447075387,
|
| 1463 |
+
"grad_norm": 0.9346863627433777,
|
| 1464 |
+
"learning_rate": 1.6722151088348274e-05,
|
| 1465 |
+
"loss": 0.5803,
|
| 1466 |
+
"step": 1960
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"epoch": 0.2521438627927813,
|
| 1470 |
+
"grad_norm": 0.5847868919372559,
|
| 1471 |
+
"learning_rate": 1.6807511737089204e-05,
|
| 1472 |
+
"loss": 0.5705,
|
| 1473 |
+
"step": 1970
|
| 1474 |
+
},
|
| 1475 |
+
{
|
| 1476 |
+
"epoch": 0.2534237808780238,
|
| 1477 |
+
"grad_norm": 1.5892306566238403,
|
| 1478 |
+
"learning_rate": 1.6892872385830134e-05,
|
| 1479 |
+
"loss": 0.5371,
|
| 1480 |
+
"step": 1980
|
| 1481 |
+
},
|
| 1482 |
+
{
|
| 1483 |
+
"epoch": 0.25470369896326633,
|
| 1484 |
+
"grad_norm": 0.8782947063446045,
|
| 1485 |
+
"learning_rate": 1.6978233034571063e-05,
|
| 1486 |
+
"loss": 0.565,
|
| 1487 |
+
"step": 1990
|
| 1488 |
+
},
|
| 1489 |
+
{
|
| 1490 |
+
"epoch": 0.2559836170485089,
|
| 1491 |
+
"grad_norm": 1.0143548250198364,
|
| 1492 |
+
"learning_rate": 1.7063593683311993e-05,
|
| 1493 |
+
"loss": 0.557,
|
| 1494 |
+
"step": 2000
|
| 1495 |
+
},
|
| 1496 |
+
{
|
| 1497 |
+
"epoch": 0.2559836170485089,
|
| 1498 |
+
"eval_accuracy": 0.74427,
|
| 1499 |
+
"eval_loss": 0.5308485627174377,
|
| 1500 |
+
"eval_macro_f1": 0.743516265066894,
|
| 1501 |
+
"eval_precision": 0.7538937627554241,
|
| 1502 |
+
"eval_recall": 0.7483595020974525,
|
| 1503 |
+
"eval_runtime": 708.2272,
|
| 1504 |
+
"eval_samples_per_second": 141.198,
|
| 1505 |
+
"eval_steps_per_second": 1.104,
|
| 1506 |
+
"step": 2000
|
| 1507 |
+
},
|
| 1508 |
+
{
|
| 1509 |
+
"epoch": 0.25726353513375144,
|
| 1510 |
+
"grad_norm": 0.9080437421798706,
|
| 1511 |
+
"learning_rate": 1.7148954332052926e-05,
|
| 1512 |
+
"loss": 0.5478,
|
| 1513 |
+
"step": 2010
|
| 1514 |
+
},
|
| 1515 |
+
{
|
| 1516 |
+
"epoch": 0.25854345321899397,
|
| 1517 |
+
"grad_norm": 1.1889078617095947,
|
| 1518 |
+
"learning_rate": 1.7234314980793856e-05,
|
| 1519 |
+
"loss": 0.5394,
|
| 1520 |
+
"step": 2020
|
| 1521 |
+
},
|
| 1522 |
+
{
|
| 1523 |
+
"epoch": 0.25982337130423655,
|
| 1524 |
+
"grad_norm": 1.7290714979171753,
|
| 1525 |
+
"learning_rate": 1.7319675629534786e-05,
|
| 1526 |
+
"loss": 0.5385,
|
| 1527 |
+
"step": 2030
|
| 1528 |
+
},
|
| 1529 |
+
{
|
| 1530 |
+
"epoch": 0.2611032893894791,
|
| 1531 |
+
"grad_norm": 0.9356957077980042,
|
| 1532 |
+
"learning_rate": 1.7405036278275716e-05,
|
| 1533 |
+
"loss": 0.5392,
|
| 1534 |
+
"step": 2040
|
| 1535 |
+
},
|
| 1536 |
+
{
|
| 1537 |
+
"epoch": 0.2623832074747216,
|
| 1538 |
+
"grad_norm": 0.765798032283783,
|
| 1539 |
+
"learning_rate": 1.7490396927016646e-05,
|
| 1540 |
+
"loss": 0.5442,
|
| 1541 |
+
"step": 2050
|
| 1542 |
+
},
|
| 1543 |
+
{
|
| 1544 |
+
"epoch": 0.2636631255599642,
|
| 1545 |
+
"grad_norm": 1.6606382131576538,
|
| 1546 |
+
"learning_rate": 1.7575757575757576e-05,
|
| 1547 |
+
"loss": 0.5331,
|
| 1548 |
+
"step": 2060
|
| 1549 |
+
},
|
| 1550 |
+
{
|
| 1551 |
+
"epoch": 0.2649430436452067,
|
| 1552 |
+
"grad_norm": 0.7928575277328491,
|
| 1553 |
+
"learning_rate": 1.766111822449851e-05,
|
| 1554 |
+
"loss": 0.5483,
|
| 1555 |
+
"step": 2070
|
| 1556 |
+
},
|
| 1557 |
+
{
|
| 1558 |
+
"epoch": 0.26622296173044924,
|
| 1559 |
+
"grad_norm": 2.2261486053466797,
|
| 1560 |
+
"learning_rate": 1.774647887323944e-05,
|
| 1561 |
+
"loss": 0.5536,
|
| 1562 |
+
"step": 2080
|
| 1563 |
+
},
|
| 1564 |
+
{
|
| 1565 |
+
"epoch": 0.2675028798156918,
|
| 1566 |
+
"grad_norm": 0.9795323610305786,
|
| 1567 |
+
"learning_rate": 1.783183952198037e-05,
|
| 1568 |
+
"loss": 0.5584,
|
| 1569 |
+
"step": 2090
|
| 1570 |
+
},
|
| 1571 |
+
{
|
| 1572 |
+
"epoch": 0.26878279790093434,
|
| 1573 |
+
"grad_norm": 2.373581886291504,
|
| 1574 |
+
"learning_rate": 1.7917200170721298e-05,
|
| 1575 |
+
"loss": 0.5475,
|
| 1576 |
+
"step": 2100
|
| 1577 |
+
},
|
| 1578 |
+
{
|
| 1579 |
+
"epoch": 0.27006271598617687,
|
| 1580 |
+
"grad_norm": 0.9658742547035217,
|
| 1581 |
+
"learning_rate": 1.8002560819462228e-05,
|
| 1582 |
+
"loss": 0.5577,
|
| 1583 |
+
"step": 2110
|
| 1584 |
+
},
|
| 1585 |
+
{
|
| 1586 |
+
"epoch": 0.27134263407141945,
|
| 1587 |
+
"grad_norm": 0.8572520017623901,
|
| 1588 |
+
"learning_rate": 1.808792146820316e-05,
|
| 1589 |
+
"loss": 0.5301,
|
| 1590 |
+
"step": 2120
|
| 1591 |
+
},
|
| 1592 |
+
{
|
| 1593 |
+
"epoch": 0.272622552156662,
|
| 1594 |
+
"grad_norm": 0.7225477695465088,
|
| 1595 |
+
"learning_rate": 1.817328211694409e-05,
|
| 1596 |
+
"loss": 0.5626,
|
| 1597 |
+
"step": 2130
|
| 1598 |
+
},
|
| 1599 |
+
{
|
| 1600 |
+
"epoch": 0.2739024702419045,
|
| 1601 |
+
"grad_norm": 1.4276026487350464,
|
| 1602 |
+
"learning_rate": 1.825864276568502e-05,
|
| 1603 |
+
"loss": 0.5208,
|
| 1604 |
+
"step": 2140
|
| 1605 |
+
},
|
| 1606 |
+
{
|
| 1607 |
+
"epoch": 0.2751823883271471,
|
| 1608 |
+
"grad_norm": 1.7469464540481567,
|
| 1609 |
+
"learning_rate": 1.834400341442595e-05,
|
| 1610 |
+
"loss": 0.5669,
|
| 1611 |
+
"step": 2150
|
| 1612 |
+
},
|
| 1613 |
+
{
|
| 1614 |
+
"epoch": 0.2764623064123896,
|
| 1615 |
+
"grad_norm": 1.250931978225708,
|
| 1616 |
+
"learning_rate": 1.842936406316688e-05,
|
| 1617 |
+
"loss": 0.5577,
|
| 1618 |
+
"step": 2160
|
| 1619 |
+
},
|
| 1620 |
+
{
|
| 1621 |
+
"epoch": 0.27774222449763214,
|
| 1622 |
+
"grad_norm": 1.011313557624817,
|
| 1623 |
+
"learning_rate": 1.8514724711907814e-05,
|
| 1624 |
+
"loss": 0.5494,
|
| 1625 |
+
"step": 2170
|
| 1626 |
+
},
|
| 1627 |
+
{
|
| 1628 |
+
"epoch": 0.2790221425828747,
|
| 1629 |
+
"grad_norm": 0.7606518268585205,
|
| 1630 |
+
"learning_rate": 1.8600085360648744e-05,
|
| 1631 |
+
"loss": 0.5391,
|
| 1632 |
+
"step": 2180
|
| 1633 |
+
},
|
| 1634 |
+
{
|
| 1635 |
+
"epoch": 0.28030206066811725,
|
| 1636 |
+
"grad_norm": 0.8173381090164185,
|
| 1637 |
+
"learning_rate": 1.8685446009389673e-05,
|
| 1638 |
+
"loss": 0.5262,
|
| 1639 |
+
"step": 2190
|
| 1640 |
+
},
|
| 1641 |
+
{
|
| 1642 |
+
"epoch": 0.2815819787533598,
|
| 1643 |
+
"grad_norm": 0.7178583145141602,
|
| 1644 |
+
"learning_rate": 1.8770806658130603e-05,
|
| 1645 |
+
"loss": 0.5182,
|
| 1646 |
+
"step": 2200
|
| 1647 |
+
},
|
| 1648 |
+
{
|
| 1649 |
+
"epoch": 0.28286189683860236,
|
| 1650 |
+
"grad_norm": 0.6852149367332458,
|
| 1651 |
+
"learning_rate": 1.8856167306871533e-05,
|
| 1652 |
+
"loss": 0.528,
|
| 1653 |
+
"step": 2210
|
| 1654 |
+
},
|
| 1655 |
+
{
|
| 1656 |
+
"epoch": 0.2841418149238449,
|
| 1657 |
+
"grad_norm": 1.9807002544403076,
|
| 1658 |
+
"learning_rate": 1.8941527955612463e-05,
|
| 1659 |
+
"loss": 0.5315,
|
| 1660 |
+
"step": 2220
|
| 1661 |
+
},
|
| 1662 |
+
{
|
| 1663 |
+
"epoch": 0.2854217330090874,
|
| 1664 |
+
"grad_norm": 0.7938176989555359,
|
| 1665 |
+
"learning_rate": 1.9026888604353396e-05,
|
| 1666 |
+
"loss": 0.5324,
|
| 1667 |
+
"step": 2230
|
| 1668 |
+
},
|
| 1669 |
+
{
|
| 1670 |
+
"epoch": 0.28670165109433,
|
| 1671 |
+
"grad_norm": 0.8181535005569458,
|
| 1672 |
+
"learning_rate": 1.9112249253094326e-05,
|
| 1673 |
+
"loss": 0.51,
|
| 1674 |
+
"step": 2240
|
| 1675 |
+
},
|
| 1676 |
+
{
|
| 1677 |
+
"epoch": 0.2879815691795725,
|
| 1678 |
+
"grad_norm": 0.8722071051597595,
|
| 1679 |
+
"learning_rate": 1.9197609901835256e-05,
|
| 1680 |
+
"loss": 0.5373,
|
| 1681 |
+
"step": 2250
|
| 1682 |
+
},
|
| 1683 |
+
{
|
| 1684 |
+
"epoch": 0.28926148726481504,
|
| 1685 |
+
"grad_norm": 1.371565818786621,
|
| 1686 |
+
"learning_rate": 1.9282970550576186e-05,
|
| 1687 |
+
"loss": 0.5227,
|
| 1688 |
+
"step": 2260
|
| 1689 |
+
},
|
| 1690 |
+
{
|
| 1691 |
+
"epoch": 0.29054140535005757,
|
| 1692 |
+
"grad_norm": 0.6769284605979919,
|
| 1693 |
+
"learning_rate": 1.9368331199317115e-05,
|
| 1694 |
+
"loss": 0.5421,
|
| 1695 |
+
"step": 2270
|
| 1696 |
+
},
|
| 1697 |
+
{
|
| 1698 |
+
"epoch": 0.29182132343530015,
|
| 1699 |
+
"grad_norm": 2.1016697883605957,
|
| 1700 |
+
"learning_rate": 1.945369184805805e-05,
|
| 1701 |
+
"loss": 0.5299,
|
| 1702 |
+
"step": 2280
|
| 1703 |
+
},
|
| 1704 |
+
{
|
| 1705 |
+
"epoch": 0.2931012415205427,
|
| 1706 |
+
"grad_norm": 1.7011477947235107,
|
| 1707 |
+
"learning_rate": 1.953905249679898e-05,
|
| 1708 |
+
"loss": 0.5403,
|
| 1709 |
+
"step": 2290
|
| 1710 |
+
},
|
| 1711 |
+
{
|
| 1712 |
+
"epoch": 0.2943811596057852,
|
| 1713 |
+
"grad_norm": 1.0961430072784424,
|
| 1714 |
+
"learning_rate": 1.962441314553991e-05,
|
| 1715 |
+
"loss": 0.534,
|
| 1716 |
+
"step": 2300
|
| 1717 |
+
},
|
| 1718 |
+
{
|
| 1719 |
+
"epoch": 0.2956610776910278,
|
| 1720 |
+
"grad_norm": 0.8536306023597717,
|
| 1721 |
+
"learning_rate": 1.9709773794280838e-05,
|
| 1722 |
+
"loss": 0.5562,
|
| 1723 |
+
"step": 2310
|
| 1724 |
+
},
|
| 1725 |
+
{
|
| 1726 |
+
"epoch": 0.2969409957762703,
|
| 1727 |
+
"grad_norm": 1.046201467514038,
|
| 1728 |
+
"learning_rate": 1.9795134443021768e-05,
|
| 1729 |
+
"loss": 0.5418,
|
| 1730 |
+
"step": 2320
|
| 1731 |
+
},
|
| 1732 |
+
{
|
| 1733 |
+
"epoch": 0.29822091386151284,
|
| 1734 |
+
"grad_norm": 1.38905930519104,
|
| 1735 |
+
"learning_rate": 1.9880495091762698e-05,
|
| 1736 |
+
"loss": 0.5557,
|
| 1737 |
+
"step": 2330
|
| 1738 |
+
},
|
| 1739 |
+
{
|
| 1740 |
+
"epoch": 0.2995008319467554,
|
| 1741 |
+
"grad_norm": 2.44215989112854,
|
| 1742 |
+
"learning_rate": 1.996585574050363e-05,
|
| 1743 |
+
"loss": 0.5285,
|
| 1744 |
+
"step": 2340
|
| 1745 |
+
},
|
| 1746 |
+
{
|
| 1747 |
+
"epoch": 0.30078075003199795,
|
| 1748 |
+
"grad_norm": 1.4736480712890625,
|
| 1749 |
+
"learning_rate": 1.9994311717861208e-05,
|
| 1750 |
+
"loss": 0.513,
|
| 1751 |
+
"step": 2350
|
| 1752 |
+
},
|
| 1753 |
+
{
|
| 1754 |
+
"epoch": 0.30206066811724047,
|
| 1755 |
+
"grad_norm": 0.7968178987503052,
|
| 1756 |
+
"learning_rate": 1.9984831247629883e-05,
|
| 1757 |
+
"loss": 0.5249,
|
| 1758 |
+
"step": 2360
|
| 1759 |
+
},
|
| 1760 |
+
{
|
| 1761 |
+
"epoch": 0.30334058620248305,
|
| 1762 |
+
"grad_norm": 1.6530416011810303,
|
| 1763 |
+
"learning_rate": 1.997535077739856e-05,
|
| 1764 |
+
"loss": 0.5329,
|
| 1765 |
+
"step": 2370
|
| 1766 |
+
},
|
| 1767 |
+
{
|
| 1768 |
+
"epoch": 0.3046205042877256,
|
| 1769 |
+
"grad_norm": 1.0949759483337402,
|
| 1770 |
+
"learning_rate": 1.9965870307167238e-05,
|
| 1771 |
+
"loss": 0.5164,
|
| 1772 |
+
"step": 2380
|
| 1773 |
+
},
|
| 1774 |
+
{
|
| 1775 |
+
"epoch": 0.3059004223729681,
|
| 1776 |
+
"grad_norm": 1.2724019289016724,
|
| 1777 |
+
"learning_rate": 1.9956389836935916e-05,
|
| 1778 |
+
"loss": 0.5293,
|
| 1779 |
+
"step": 2390
|
| 1780 |
+
},
|
| 1781 |
+
{
|
| 1782 |
+
"epoch": 0.3071803404582107,
|
| 1783 |
+
"grad_norm": 1.8785524368286133,
|
| 1784 |
+
"learning_rate": 1.994690936670459e-05,
|
| 1785 |
+
"loss": 0.5286,
|
| 1786 |
+
"step": 2400
|
| 1787 |
+
},
|
| 1788 |
+
{
|
| 1789 |
+
"epoch": 0.3084602585434532,
|
| 1790 |
+
"grad_norm": 2.1517200469970703,
|
| 1791 |
+
"learning_rate": 1.9937428896473267e-05,
|
| 1792 |
+
"loss": 0.5273,
|
| 1793 |
+
"step": 2410
|
| 1794 |
+
},
|
| 1795 |
+
{
|
| 1796 |
+
"epoch": 0.30974017662869574,
|
| 1797 |
+
"grad_norm": 0.999515175819397,
|
| 1798 |
+
"learning_rate": 1.9927948426241942e-05,
|
| 1799 |
+
"loss": 0.5115,
|
| 1800 |
+
"step": 2420
|
| 1801 |
+
},
|
| 1802 |
+
{
|
| 1803 |
+
"epoch": 0.3110200947139383,
|
| 1804 |
+
"grad_norm": 2.214357852935791,
|
| 1805 |
+
"learning_rate": 1.991846795601062e-05,
|
| 1806 |
+
"loss": 0.5145,
|
| 1807 |
+
"step": 2430
|
| 1808 |
+
},
|
| 1809 |
+
{
|
| 1810 |
+
"epoch": 0.31230001279918085,
|
| 1811 |
+
"grad_norm": 1.8795071840286255,
|
| 1812 |
+
"learning_rate": 1.9908987485779297e-05,
|
| 1813 |
+
"loss": 0.509,
|
| 1814 |
+
"step": 2440
|
| 1815 |
+
},
|
| 1816 |
+
{
|
| 1817 |
+
"epoch": 0.3135799308844234,
|
| 1818 |
+
"grad_norm": 0.9374184608459473,
|
| 1819 |
+
"learning_rate": 1.989950701554797e-05,
|
| 1820 |
+
"loss": 0.5112,
|
| 1821 |
+
"step": 2450
|
| 1822 |
+
},
|
| 1823 |
+
{
|
| 1824 |
+
"epoch": 0.31485984896966596,
|
| 1825 |
+
"grad_norm": 0.7281107306480408,
|
| 1826 |
+
"learning_rate": 1.989002654531665e-05,
|
| 1827 |
+
"loss": 0.5029,
|
| 1828 |
+
"step": 2460
|
| 1829 |
+
},
|
| 1830 |
+
{
|
| 1831 |
+
"epoch": 0.3161397670549085,
|
| 1832 |
+
"grad_norm": 2.6096737384796143,
|
| 1833 |
+
"learning_rate": 1.9880546075085327e-05,
|
| 1834 |
+
"loss": 0.5302,
|
| 1835 |
+
"step": 2470
|
| 1836 |
+
},
|
| 1837 |
+
{
|
| 1838 |
+
"epoch": 0.317419685140151,
|
| 1839 |
+
"grad_norm": 0.7463436126708984,
|
| 1840 |
+
"learning_rate": 1.9871065604854004e-05,
|
| 1841 |
+
"loss": 0.5249,
|
| 1842 |
+
"step": 2480
|
| 1843 |
+
},
|
| 1844 |
+
{
|
| 1845 |
+
"epoch": 0.3186996032253936,
|
| 1846 |
+
"grad_norm": 0.7995405197143555,
|
| 1847 |
+
"learning_rate": 1.986158513462268e-05,
|
| 1848 |
+
"loss": 0.5171,
|
| 1849 |
+
"step": 2490
|
| 1850 |
+
},
|
| 1851 |
+
{
|
| 1852 |
+
"epoch": 0.3199795213106361,
|
| 1853 |
+
"grad_norm": 0.873103678226471,
|
| 1854 |
+
"learning_rate": 1.9852104664391356e-05,
|
| 1855 |
+
"loss": 0.5125,
|
| 1856 |
+
"step": 2500
|
| 1857 |
+
},
|
| 1858 |
+
{
|
| 1859 |
+
"epoch": 0.32125943939587864,
|
| 1860 |
+
"grad_norm": 1.5319750308990479,
|
| 1861 |
+
"learning_rate": 1.984262419416003e-05,
|
| 1862 |
+
"loss": 0.527,
|
| 1863 |
+
"step": 2510
|
| 1864 |
+
},
|
| 1865 |
+
{
|
| 1866 |
+
"epoch": 0.3225393574811212,
|
| 1867 |
+
"grad_norm": 1.717421531677246,
|
| 1868 |
+
"learning_rate": 1.983314372392871e-05,
|
| 1869 |
+
"loss": 0.5415,
|
| 1870 |
+
"step": 2520
|
| 1871 |
+
},
|
| 1872 |
+
{
|
| 1873 |
+
"epoch": 0.32381927556636375,
|
| 1874 |
+
"grad_norm": 0.7133586406707764,
|
| 1875 |
+
"learning_rate": 1.9823663253697386e-05,
|
| 1876 |
+
"loss": 0.5251,
|
| 1877 |
+
"step": 2530
|
| 1878 |
+
},
|
| 1879 |
+
{
|
| 1880 |
+
"epoch": 0.3250991936516063,
|
| 1881 |
+
"grad_norm": 0.7959030270576477,
|
| 1882 |
+
"learning_rate": 1.981418278346606e-05,
|
| 1883 |
+
"loss": 0.5127,
|
| 1884 |
+
"step": 2540
|
| 1885 |
+
},
|
| 1886 |
+
{
|
| 1887 |
+
"epoch": 0.32637911173684886,
|
| 1888 |
+
"grad_norm": 2.4905223846435547,
|
| 1889 |
+
"learning_rate": 1.9804702313234738e-05,
|
| 1890 |
+
"loss": 0.4944,
|
| 1891 |
+
"step": 2550
|
| 1892 |
+
},
|
| 1893 |
+
{
|
| 1894 |
+
"epoch": 0.3276590298220914,
|
| 1895 |
+
"grad_norm": 1.0039845705032349,
|
| 1896 |
+
"learning_rate": 1.9795221843003412e-05,
|
| 1897 |
+
"loss": 0.534,
|
| 1898 |
+
"step": 2560
|
| 1899 |
+
},
|
| 1900 |
+
{
|
| 1901 |
+
"epoch": 0.3289389479073339,
|
| 1902 |
+
"grad_norm": 1.2181288003921509,
|
| 1903 |
+
"learning_rate": 1.978574137277209e-05,
|
| 1904 |
+
"loss": 0.5198,
|
| 1905 |
+
"step": 2570
|
| 1906 |
+
},
|
| 1907 |
+
{
|
| 1908 |
+
"epoch": 0.3302188659925765,
|
| 1909 |
+
"grad_norm": 0.9166068434715271,
|
| 1910 |
+
"learning_rate": 1.9776260902540768e-05,
|
| 1911 |
+
"loss": 0.5373,
|
| 1912 |
+
"step": 2580
|
| 1913 |
+
},
|
| 1914 |
+
{
|
| 1915 |
+
"epoch": 0.331498784077819,
|
| 1916 |
+
"grad_norm": 0.938994288444519,
|
| 1917 |
+
"learning_rate": 1.9766780432309445e-05,
|
| 1918 |
+
"loss": 0.5179,
|
| 1919 |
+
"step": 2590
|
| 1920 |
+
},
|
| 1921 |
+
{
|
| 1922 |
+
"epoch": 0.33277870216306155,
|
| 1923 |
+
"grad_norm": 0.8827023506164551,
|
| 1924 |
+
"learning_rate": 1.975729996207812e-05,
|
| 1925 |
+
"loss": 0.5249,
|
| 1926 |
+
"step": 2600
|
| 1927 |
+
},
|
| 1928 |
+
{
|
| 1929 |
+
"epoch": 0.33405862024830413,
|
| 1930 |
+
"grad_norm": 0.8544501066207886,
|
| 1931 |
+
"learning_rate": 1.9747819491846797e-05,
|
| 1932 |
+
"loss": 0.5172,
|
| 1933 |
+
"step": 2610
|
| 1934 |
+
},
|
| 1935 |
+
{
|
| 1936 |
+
"epoch": 0.33533853833354665,
|
| 1937 |
+
"grad_norm": 0.7890387177467346,
|
| 1938 |
+
"learning_rate": 1.9738339021615475e-05,
|
| 1939 |
+
"loss": 0.5207,
|
| 1940 |
+
"step": 2620
|
| 1941 |
+
},
|
| 1942 |
+
{
|
| 1943 |
+
"epoch": 0.3366184564187892,
|
| 1944 |
+
"grad_norm": 0.7724694013595581,
|
| 1945 |
+
"learning_rate": 1.9728858551384153e-05,
|
| 1946 |
+
"loss": 0.5065,
|
| 1947 |
+
"step": 2630
|
| 1948 |
+
},
|
| 1949 |
+
{
|
| 1950 |
+
"epoch": 0.33789837450403176,
|
| 1951 |
+
"grad_norm": 0.7064672708511353,
|
| 1952 |
+
"learning_rate": 1.9719378081152827e-05,
|
| 1953 |
+
"loss": 0.4828,
|
| 1954 |
+
"step": 2640
|
| 1955 |
+
},
|
| 1956 |
+
{
|
| 1957 |
+
"epoch": 0.3391782925892743,
|
| 1958 |
+
"grad_norm": 0.9625025391578674,
|
| 1959 |
+
"learning_rate": 1.97098976109215e-05,
|
| 1960 |
+
"loss": 0.5108,
|
| 1961 |
+
"step": 2650
|
| 1962 |
+
},
|
| 1963 |
+
{
|
| 1964 |
+
"epoch": 0.3404582106745168,
|
| 1965 |
+
"grad_norm": 2.1604244709014893,
|
| 1966 |
+
"learning_rate": 1.970041714069018e-05,
|
| 1967 |
+
"loss": 0.5129,
|
| 1968 |
+
"step": 2660
|
| 1969 |
+
},
|
| 1970 |
+
{
|
| 1971 |
+
"epoch": 0.3417381287597594,
|
| 1972 |
+
"grad_norm": 0.8561355471611023,
|
| 1973 |
+
"learning_rate": 1.9690936670458856e-05,
|
| 1974 |
+
"loss": 0.521,
|
| 1975 |
+
"step": 2670
|
| 1976 |
+
},
|
| 1977 |
+
{
|
| 1978 |
+
"epoch": 0.3430180468450019,
|
| 1979 |
+
"grad_norm": 0.8122753500938416,
|
| 1980 |
+
"learning_rate": 1.9681456200227534e-05,
|
| 1981 |
+
"loss": 0.5039,
|
| 1982 |
+
"step": 2680
|
| 1983 |
+
},
|
| 1984 |
+
{
|
| 1985 |
+
"epoch": 0.34429796493024445,
|
| 1986 |
+
"grad_norm": 0.848838210105896,
|
| 1987 |
+
"learning_rate": 1.967197572999621e-05,
|
| 1988 |
+
"loss": 0.5559,
|
| 1989 |
+
"step": 2690
|
| 1990 |
+
},
|
| 1991 |
+
{
|
| 1992 |
+
"epoch": 0.34557788301548703,
|
| 1993 |
+
"grad_norm": 0.8424274325370789,
|
| 1994 |
+
"learning_rate": 1.9662495259764886e-05,
|
| 1995 |
+
"loss": 0.5299,
|
| 1996 |
+
"step": 2700
|
| 1997 |
+
},
|
| 1998 |
+
{
|
| 1999 |
+
"epoch": 0.34685780110072956,
|
| 2000 |
+
"grad_norm": 1.0475687980651855,
|
| 2001 |
+
"learning_rate": 1.9653014789533564e-05,
|
| 2002 |
+
"loss": 0.5003,
|
| 2003 |
+
"step": 2710
|
| 2004 |
+
},
|
| 2005 |
+
{
|
| 2006 |
+
"epoch": 0.3481377191859721,
|
| 2007 |
+
"grad_norm": 0.9720590710639954,
|
| 2008 |
+
"learning_rate": 1.964353431930224e-05,
|
| 2009 |
+
"loss": 0.4976,
|
| 2010 |
+
"step": 2720
|
| 2011 |
+
},
|
| 2012 |
+
{
|
| 2013 |
+
"epoch": 0.34941763727121466,
|
| 2014 |
+
"grad_norm": 1.1562892198562622,
|
| 2015 |
+
"learning_rate": 1.9634053849070916e-05,
|
| 2016 |
+
"loss": 0.4728,
|
| 2017 |
+
"step": 2730
|
| 2018 |
+
},
|
| 2019 |
+
{
|
| 2020 |
+
"epoch": 0.3506975553564572,
|
| 2021 |
+
"grad_norm": 1.2779773473739624,
|
| 2022 |
+
"learning_rate": 1.9624573378839593e-05,
|
| 2023 |
+
"loss": 0.4933,
|
| 2024 |
+
"step": 2740
|
| 2025 |
+
},
|
| 2026 |
+
{
|
| 2027 |
+
"epoch": 0.3519774734416997,
|
| 2028 |
+
"grad_norm": 1.047945499420166,
|
| 2029 |
+
"learning_rate": 1.9615092908608268e-05,
|
| 2030 |
+
"loss": 0.518,
|
| 2031 |
+
"step": 2750
|
| 2032 |
+
},
|
| 2033 |
+
{
|
| 2034 |
+
"epoch": 0.3532573915269423,
|
| 2035 |
+
"grad_norm": 0.7196894884109497,
|
| 2036 |
+
"learning_rate": 1.9605612438376945e-05,
|
| 2037 |
+
"loss": 0.5006,
|
| 2038 |
+
"step": 2760
|
| 2039 |
+
},
|
| 2040 |
+
{
|
| 2041 |
+
"epoch": 0.3545373096121848,
|
| 2042 |
+
"grad_norm": 0.9644359350204468,
|
| 2043 |
+
"learning_rate": 1.9596131968145623e-05,
|
| 2044 |
+
"loss": 0.5083,
|
| 2045 |
+
"step": 2770
|
| 2046 |
+
},
|
| 2047 |
+
{
|
| 2048 |
+
"epoch": 0.35581722769742735,
|
| 2049 |
+
"grad_norm": 0.9720232486724854,
|
| 2050 |
+
"learning_rate": 1.9586651497914297e-05,
|
| 2051 |
+
"loss": 0.5194,
|
| 2052 |
+
"step": 2780
|
| 2053 |
+
},
|
| 2054 |
+
{
|
| 2055 |
+
"epoch": 0.35709714578266993,
|
| 2056 |
+
"grad_norm": 0.8277772665023804,
|
| 2057 |
+
"learning_rate": 1.9577171027682975e-05,
|
| 2058 |
+
"loss": 0.5121,
|
| 2059 |
+
"step": 2790
|
| 2060 |
+
},
|
| 2061 |
+
{
|
| 2062 |
+
"epoch": 0.35837706386791246,
|
| 2063 |
+
"grad_norm": 0.7811515927314758,
|
| 2064 |
+
"learning_rate": 1.956769055745165e-05,
|
| 2065 |
+
"loss": 0.5106,
|
| 2066 |
+
"step": 2800
|
| 2067 |
+
},
|
| 2068 |
+
{
|
| 2069 |
+
"epoch": 0.359656981953155,
|
| 2070 |
+
"grad_norm": 0.7438198328018188,
|
| 2071 |
+
"learning_rate": 1.9558210087220327e-05,
|
| 2072 |
+
"loss": 0.5183,
|
| 2073 |
+
"step": 2810
|
| 2074 |
+
},
|
| 2075 |
+
{
|
| 2076 |
+
"epoch": 0.36093690003839757,
|
| 2077 |
+
"grad_norm": 1.1424624919891357,
|
| 2078 |
+
"learning_rate": 1.9548729616989005e-05,
|
| 2079 |
+
"loss": 0.4973,
|
| 2080 |
+
"step": 2820
|
| 2081 |
+
},
|
| 2082 |
+
{
|
| 2083 |
+
"epoch": 0.3622168181236401,
|
| 2084 |
+
"grad_norm": 1.0706868171691895,
|
| 2085 |
+
"learning_rate": 1.9539249146757682e-05,
|
| 2086 |
+
"loss": 0.5063,
|
| 2087 |
+
"step": 2830
|
| 2088 |
+
},
|
| 2089 |
+
{
|
| 2090 |
+
"epoch": 0.3634967362088826,
|
| 2091 |
+
"grad_norm": 1.1213792562484741,
|
| 2092 |
+
"learning_rate": 1.9529768676526357e-05,
|
| 2093 |
+
"loss": 0.4991,
|
| 2094 |
+
"step": 2840
|
| 2095 |
+
},
|
| 2096 |
+
{
|
| 2097 |
+
"epoch": 0.3647766542941252,
|
| 2098 |
+
"grad_norm": 1.0018831491470337,
|
| 2099 |
+
"learning_rate": 1.9520288206295034e-05,
|
| 2100 |
+
"loss": 0.5004,
|
| 2101 |
+
"step": 2850
|
| 2102 |
+
},
|
| 2103 |
+
{
|
| 2104 |
+
"epoch": 0.36605657237936773,
|
| 2105 |
+
"grad_norm": 0.8369762897491455,
|
| 2106 |
+
"learning_rate": 1.9510807736063712e-05,
|
| 2107 |
+
"loss": 0.5341,
|
| 2108 |
+
"step": 2860
|
| 2109 |
+
},
|
| 2110 |
+
{
|
| 2111 |
+
"epoch": 0.36733649046461025,
|
| 2112 |
+
"grad_norm": 0.8711100220680237,
|
| 2113 |
+
"learning_rate": 1.9501327265832386e-05,
|
| 2114 |
+
"loss": 0.4788,
|
| 2115 |
+
"step": 2870
|
| 2116 |
+
},
|
| 2117 |
+
{
|
| 2118 |
+
"epoch": 0.36861640854985284,
|
| 2119 |
+
"grad_norm": 1.0853824615478516,
|
| 2120 |
+
"learning_rate": 1.9491846795601064e-05,
|
| 2121 |
+
"loss": 0.5419,
|
| 2122 |
+
"step": 2880
|
| 2123 |
+
},
|
| 2124 |
+
{
|
| 2125 |
+
"epoch": 0.36989632663509536,
|
| 2126 |
+
"grad_norm": 1.4816603660583496,
|
| 2127 |
+
"learning_rate": 1.9482366325369738e-05,
|
| 2128 |
+
"loss": 0.5252,
|
| 2129 |
+
"step": 2890
|
| 2130 |
+
},
|
| 2131 |
+
{
|
| 2132 |
+
"epoch": 0.3711762447203379,
|
| 2133 |
+
"grad_norm": 1.3375904560089111,
|
| 2134 |
+
"learning_rate": 1.9472885855138416e-05,
|
| 2135 |
+
"loss": 0.5127,
|
| 2136 |
+
"step": 2900
|
| 2137 |
+
},
|
| 2138 |
+
{
|
| 2139 |
+
"epoch": 0.3724561628055804,
|
| 2140 |
+
"grad_norm": 0.7361944913864136,
|
| 2141 |
+
"learning_rate": 1.9463405384907093e-05,
|
| 2142 |
+
"loss": 0.5278,
|
| 2143 |
+
"step": 2910
|
| 2144 |
+
},
|
| 2145 |
+
{
|
| 2146 |
+
"epoch": 0.373736080890823,
|
| 2147 |
+
"grad_norm": 1.1056461334228516,
|
| 2148 |
+
"learning_rate": 1.945392491467577e-05,
|
| 2149 |
+
"loss": 0.5072,
|
| 2150 |
+
"step": 2920
|
| 2151 |
+
},
|
| 2152 |
+
{
|
| 2153 |
+
"epoch": 0.3750159989760655,
|
| 2154 |
+
"grad_norm": 0.921783447265625,
|
| 2155 |
+
"learning_rate": 1.9444444444444445e-05,
|
| 2156 |
+
"loss": 0.5161,
|
| 2157 |
+
"step": 2930
|
| 2158 |
+
},
|
| 2159 |
+
{
|
| 2160 |
+
"epoch": 0.37629591706130805,
|
| 2161 |
+
"grad_norm": 1.15033757686615,
|
| 2162 |
+
"learning_rate": 1.9434963974213123e-05,
|
| 2163 |
+
"loss": 0.5329,
|
| 2164 |
+
"step": 2940
|
| 2165 |
+
},
|
| 2166 |
+
{
|
| 2167 |
+
"epoch": 0.37757583514655063,
|
| 2168 |
+
"grad_norm": 1.3983162641525269,
|
| 2169 |
+
"learning_rate": 1.94254835039818e-05,
|
| 2170 |
+
"loss": 0.5034,
|
| 2171 |
+
"step": 2950
|
| 2172 |
+
},
|
| 2173 |
+
{
|
| 2174 |
+
"epoch": 0.37885575323179316,
|
| 2175 |
+
"grad_norm": 1.0419155359268188,
|
| 2176 |
+
"learning_rate": 1.9416003033750475e-05,
|
| 2177 |
+
"loss": 0.5281,
|
| 2178 |
+
"step": 2960
|
| 2179 |
+
},
|
| 2180 |
+
{
|
| 2181 |
+
"epoch": 0.3801356713170357,
|
| 2182 |
+
"grad_norm": 2.478982448577881,
|
| 2183 |
+
"learning_rate": 1.9406522563519153e-05,
|
| 2184 |
+
"loss": 0.4843,
|
| 2185 |
+
"step": 2970
|
| 2186 |
+
},
|
| 2187 |
+
{
|
| 2188 |
+
"epoch": 0.38141558940227827,
|
| 2189 |
+
"grad_norm": 0.7638375163078308,
|
| 2190 |
+
"learning_rate": 1.9397042093287827e-05,
|
| 2191 |
+
"loss": 0.5214,
|
| 2192 |
+
"step": 2980
|
| 2193 |
+
},
|
| 2194 |
+
{
|
| 2195 |
+
"epoch": 0.3826955074875208,
|
| 2196 |
+
"grad_norm": 2.18105411529541,
|
| 2197 |
+
"learning_rate": 1.9387561623056505e-05,
|
| 2198 |
+
"loss": 0.5373,
|
| 2199 |
+
"step": 2990
|
| 2200 |
+
},
|
| 2201 |
+
{
|
| 2202 |
+
"epoch": 0.3839754255727633,
|
| 2203 |
+
"grad_norm": 1.0936481952667236,
|
| 2204 |
+
"learning_rate": 1.9378081152825182e-05,
|
| 2205 |
+
"loss": 0.5187,
|
| 2206 |
+
"step": 3000
|
| 2207 |
+
},
|
| 2208 |
+
{
|
| 2209 |
+
"epoch": 0.3852553436580059,
|
| 2210 |
+
"grad_norm": 1.0057384967803955,
|
| 2211 |
+
"learning_rate": 1.936860068259386e-05,
|
| 2212 |
+
"loss": 0.5179,
|
| 2213 |
+
"step": 3010
|
| 2214 |
+
},
|
| 2215 |
+
{
|
| 2216 |
+
"epoch": 0.3865352617432484,
|
| 2217 |
+
"grad_norm": 1.9525461196899414,
|
| 2218 |
+
"learning_rate": 1.9359120212362534e-05,
|
| 2219 |
+
"loss": 0.4964,
|
| 2220 |
+
"step": 3020
|
| 2221 |
+
},
|
| 2222 |
+
{
|
| 2223 |
+
"epoch": 0.38781517982849095,
|
| 2224 |
+
"grad_norm": 0.9308017492294312,
|
| 2225 |
+
"learning_rate": 1.9349639742131212e-05,
|
| 2226 |
+
"loss": 0.4751,
|
| 2227 |
+
"step": 3030
|
| 2228 |
+
},
|
| 2229 |
+
{
|
| 2230 |
+
"epoch": 0.38909509791373353,
|
| 2231 |
+
"grad_norm": 0.6847271919250488,
|
| 2232 |
+
"learning_rate": 1.9340159271899886e-05,
|
| 2233 |
+
"loss": 0.5049,
|
| 2234 |
+
"step": 3040
|
| 2235 |
+
},
|
| 2236 |
+
{
|
| 2237 |
+
"epoch": 0.39037501599897606,
|
| 2238 |
+
"grad_norm": 0.9826173186302185,
|
| 2239 |
+
"learning_rate": 1.9330678801668564e-05,
|
| 2240 |
+
"loss": 0.5089,
|
| 2241 |
+
"step": 3050
|
| 2242 |
+
},
|
| 2243 |
+
{
|
| 2244 |
+
"epoch": 0.3916549340842186,
|
| 2245 |
+
"grad_norm": 1.1526622772216797,
|
| 2246 |
+
"learning_rate": 1.932119833143724e-05,
|
| 2247 |
+
"loss": 0.5263,
|
| 2248 |
+
"step": 3060
|
| 2249 |
+
},
|
| 2250 |
+
{
|
| 2251 |
+
"epoch": 0.39293485216946117,
|
| 2252 |
+
"grad_norm": 1.3890351057052612,
|
| 2253 |
+
"learning_rate": 1.9311717861205916e-05,
|
| 2254 |
+
"loss": 0.5293,
|
| 2255 |
+
"step": 3070
|
| 2256 |
+
},
|
| 2257 |
+
{
|
| 2258 |
+
"epoch": 0.3942147702547037,
|
| 2259 |
+
"grad_norm": 1.8960402011871338,
|
| 2260 |
+
"learning_rate": 1.9302237390974593e-05,
|
| 2261 |
+
"loss": 0.4944,
|
| 2262 |
+
"step": 3080
|
| 2263 |
+
},
|
| 2264 |
+
{
|
| 2265 |
+
"epoch": 0.3954946883399462,
|
| 2266 |
+
"grad_norm": 3.4239537715911865,
|
| 2267 |
+
"learning_rate": 1.929275692074327e-05,
|
| 2268 |
+
"loss": 0.5281,
|
| 2269 |
+
"step": 3090
|
| 2270 |
+
},
|
| 2271 |
+
{
|
| 2272 |
+
"epoch": 0.3967746064251888,
|
| 2273 |
+
"grad_norm": 0.8984873294830322,
|
| 2274 |
+
"learning_rate": 1.928327645051195e-05,
|
| 2275 |
+
"loss": 0.5169,
|
| 2276 |
+
"step": 3100
|
| 2277 |
+
},
|
| 2278 |
+
{
|
| 2279 |
+
"epoch": 0.39805452451043133,
|
| 2280 |
+
"grad_norm": 1.3759665489196777,
|
| 2281 |
+
"learning_rate": 1.9273795980280623e-05,
|
| 2282 |
+
"loss": 0.5086,
|
| 2283 |
+
"step": 3110
|
| 2284 |
+
},
|
| 2285 |
+
{
|
| 2286 |
+
"epoch": 0.39933444259567386,
|
| 2287 |
+
"grad_norm": 1.0449821949005127,
|
| 2288 |
+
"learning_rate": 1.92643155100493e-05,
|
| 2289 |
+
"loss": 0.4921,
|
| 2290 |
+
"step": 3120
|
| 2291 |
+
},
|
| 2292 |
+
{
|
| 2293 |
+
"epoch": 0.40061436068091644,
|
| 2294 |
+
"grad_norm": 0.9045126438140869,
|
| 2295 |
+
"learning_rate": 1.9254835039817975e-05,
|
| 2296 |
+
"loss": 0.4888,
|
| 2297 |
+
"step": 3130
|
| 2298 |
+
},
|
| 2299 |
+
{
|
| 2300 |
+
"epoch": 0.40189427876615896,
|
| 2301 |
+
"grad_norm": 0.8578636050224304,
|
| 2302 |
+
"learning_rate": 1.9245354569586653e-05,
|
| 2303 |
+
"loss": 0.4956,
|
| 2304 |
+
"step": 3140
|
| 2305 |
+
},
|
| 2306 |
+
{
|
| 2307 |
+
"epoch": 0.4031741968514015,
|
| 2308 |
+
"grad_norm": 1.2654173374176025,
|
| 2309 |
+
"learning_rate": 1.923587409935533e-05,
|
| 2310 |
+
"loss": 0.4942,
|
| 2311 |
+
"step": 3150
|
| 2312 |
+
},
|
| 2313 |
+
{
|
| 2314 |
+
"epoch": 0.40445411493664407,
|
| 2315 |
+
"grad_norm": 1.6073803901672363,
|
| 2316 |
+
"learning_rate": 1.9226393629124008e-05,
|
| 2317 |
+
"loss": 0.5037,
|
| 2318 |
+
"step": 3160
|
| 2319 |
+
},
|
| 2320 |
+
{
|
| 2321 |
+
"epoch": 0.4057340330218866,
|
| 2322 |
+
"grad_norm": 0.9294481873512268,
|
| 2323 |
+
"learning_rate": 1.9216913158892682e-05,
|
| 2324 |
+
"loss": 0.4836,
|
| 2325 |
+
"step": 3170
|
| 2326 |
+
},
|
| 2327 |
+
{
|
| 2328 |
+
"epoch": 0.4070139511071291,
|
| 2329 |
+
"grad_norm": 1.1646478176116943,
|
| 2330 |
+
"learning_rate": 1.9207432688661357e-05,
|
| 2331 |
+
"loss": 0.5189,
|
| 2332 |
+
"step": 3180
|
| 2333 |
+
},
|
| 2334 |
+
{
|
| 2335 |
+
"epoch": 0.4082938691923717,
|
| 2336 |
+
"grad_norm": 0.8620973825454712,
|
| 2337 |
+
"learning_rate": 1.9197952218430034e-05,
|
| 2338 |
+
"loss": 0.5002,
|
| 2339 |
+
"step": 3190
|
| 2340 |
+
},
|
| 2341 |
+
{
|
| 2342 |
+
"epoch": 0.40957378727761423,
|
| 2343 |
+
"grad_norm": 0.7730439305305481,
|
| 2344 |
+
"learning_rate": 1.9188471748198712e-05,
|
| 2345 |
+
"loss": 0.4802,
|
| 2346 |
+
"step": 3200
|
| 2347 |
+
},
|
| 2348 |
+
{
|
| 2349 |
+
"epoch": 0.41085370536285676,
|
| 2350 |
+
"grad_norm": 0.706767201423645,
|
| 2351 |
+
"learning_rate": 1.917899127796739e-05,
|
| 2352 |
+
"loss": 0.535,
|
| 2353 |
+
"step": 3210
|
| 2354 |
+
},
|
| 2355 |
+
{
|
| 2356 |
+
"epoch": 0.41213362344809934,
|
| 2357 |
+
"grad_norm": 1.1379191875457764,
|
| 2358 |
+
"learning_rate": 1.9169510807736064e-05,
|
| 2359 |
+
"loss": 0.5186,
|
| 2360 |
+
"step": 3220
|
| 2361 |
+
},
|
| 2362 |
+
{
|
| 2363 |
+
"epoch": 0.41341354153334187,
|
| 2364 |
+
"grad_norm": 0.6369794011116028,
|
| 2365 |
+
"learning_rate": 1.916003033750474e-05,
|
| 2366 |
+
"loss": 0.4896,
|
| 2367 |
+
"step": 3230
|
| 2368 |
+
},
|
| 2369 |
+
{
|
| 2370 |
+
"epoch": 0.4146934596185844,
|
| 2371 |
+
"grad_norm": 3.0110867023468018,
|
| 2372 |
+
"learning_rate": 1.915054986727342e-05,
|
| 2373 |
+
"loss": 0.5275,
|
| 2374 |
+
"step": 3240
|
| 2375 |
+
},
|
| 2376 |
+
{
|
| 2377 |
+
"epoch": 0.415973377703827,
|
| 2378 |
+
"grad_norm": 1.3241641521453857,
|
| 2379 |
+
"learning_rate": 1.9141069397042097e-05,
|
| 2380 |
+
"loss": 0.5078,
|
| 2381 |
+
"step": 3250
|
| 2382 |
+
},
|
| 2383 |
+
{
|
| 2384 |
+
"epoch": 0.4172532957890695,
|
| 2385 |
+
"grad_norm": 0.942676305770874,
|
| 2386 |
+
"learning_rate": 1.913158892681077e-05,
|
| 2387 |
+
"loss": 0.4901,
|
| 2388 |
+
"step": 3260
|
| 2389 |
+
},
|
| 2390 |
+
{
|
| 2391 |
+
"epoch": 0.418533213874312,
|
| 2392 |
+
"grad_norm": 2.2048139572143555,
|
| 2393 |
+
"learning_rate": 1.9122108456579446e-05,
|
| 2394 |
+
"loss": 0.5272,
|
| 2395 |
+
"step": 3270
|
| 2396 |
+
},
|
| 2397 |
+
{
|
| 2398 |
+
"epoch": 0.4198131319595546,
|
| 2399 |
+
"grad_norm": 1.8954252004623413,
|
| 2400 |
+
"learning_rate": 1.9112627986348123e-05,
|
| 2401 |
+
"loss": 0.4853,
|
| 2402 |
+
"step": 3280
|
| 2403 |
+
},
|
| 2404 |
+
{
|
| 2405 |
+
"epoch": 0.42109305004479713,
|
| 2406 |
+
"grad_norm": 0.8221827745437622,
|
| 2407 |
+
"learning_rate": 1.91031475161168e-05,
|
| 2408 |
+
"loss": 0.4919,
|
| 2409 |
+
"step": 3290
|
| 2410 |
+
},
|
| 2411 |
+
{
|
| 2412 |
+
"epoch": 0.42237296813003966,
|
| 2413 |
+
"grad_norm": 0.742501437664032,
|
| 2414 |
+
"learning_rate": 1.909366704588548e-05,
|
| 2415 |
+
"loss": 0.5139,
|
| 2416 |
+
"step": 3300
|
| 2417 |
+
},
|
| 2418 |
+
{
|
| 2419 |
+
"epoch": 0.42365288621528224,
|
| 2420 |
+
"grad_norm": 1.9924613237380981,
|
| 2421 |
+
"learning_rate": 1.9084186575654153e-05,
|
| 2422 |
+
"loss": 0.4803,
|
| 2423 |
+
"step": 3310
|
| 2424 |
+
},
|
| 2425 |
+
{
|
| 2426 |
+
"epoch": 0.42493280430052477,
|
| 2427 |
+
"grad_norm": 0.9771905541419983,
|
| 2428 |
+
"learning_rate": 1.907470610542283e-05,
|
| 2429 |
+
"loss": 0.475,
|
| 2430 |
+
"step": 3320
|
| 2431 |
+
},
|
| 2432 |
+
{
|
| 2433 |
+
"epoch": 0.4262127223857673,
|
| 2434 |
+
"grad_norm": 1.5012576580047607,
|
| 2435 |
+
"learning_rate": 1.9065225635191508e-05,
|
| 2436 |
+
"loss": 0.4941,
|
| 2437 |
+
"step": 3330
|
| 2438 |
+
},
|
| 2439 |
+
{
|
| 2440 |
+
"epoch": 0.4274926404710099,
|
| 2441 |
+
"grad_norm": 0.6949377059936523,
|
| 2442 |
+
"learning_rate": 1.9055745164960186e-05,
|
| 2443 |
+
"loss": 0.512,
|
| 2444 |
+
"step": 3340
|
| 2445 |
+
},
|
| 2446 |
+
{
|
| 2447 |
+
"epoch": 0.4287725585562524,
|
| 2448 |
+
"grad_norm": 0.9960921406745911,
|
| 2449 |
+
"learning_rate": 1.904626469472886e-05,
|
| 2450 |
+
"loss": 0.4825,
|
| 2451 |
+
"step": 3350
|
| 2452 |
+
},
|
| 2453 |
+
{
|
| 2454 |
+
"epoch": 0.43005247664149493,
|
| 2455 |
+
"grad_norm": 1.0516977310180664,
|
| 2456 |
+
"learning_rate": 1.9036784224497538e-05,
|
| 2457 |
+
"loss": 0.5117,
|
| 2458 |
+
"step": 3360
|
| 2459 |
+
},
|
| 2460 |
+
{
|
| 2461 |
+
"epoch": 0.4313323947267375,
|
| 2462 |
+
"grad_norm": 0.9433191418647766,
|
| 2463 |
+
"learning_rate": 1.9027303754266212e-05,
|
| 2464 |
+
"loss": 0.484,
|
| 2465 |
+
"step": 3370
|
| 2466 |
+
},
|
| 2467 |
+
{
|
| 2468 |
+
"epoch": 0.43261231281198004,
|
| 2469 |
+
"grad_norm": 2.7105135917663574,
|
| 2470 |
+
"learning_rate": 1.901782328403489e-05,
|
| 2471 |
+
"loss": 0.4986,
|
| 2472 |
+
"step": 3380
|
| 2473 |
+
},
|
| 2474 |
+
{
|
| 2475 |
+
"epoch": 0.43389223089722256,
|
| 2476 |
+
"grad_norm": 0.8293517827987671,
|
| 2477 |
+
"learning_rate": 1.9008342813803567e-05,
|
| 2478 |
+
"loss": 0.4797,
|
| 2479 |
+
"step": 3390
|
| 2480 |
+
},
|
| 2481 |
+
{
|
| 2482 |
+
"epoch": 0.43517214898246515,
|
| 2483 |
+
"grad_norm": 1.3104497194290161,
|
| 2484 |
+
"learning_rate": 1.899886234357224e-05,
|
| 2485 |
+
"loss": 0.4971,
|
| 2486 |
+
"step": 3400
|
| 2487 |
+
},
|
| 2488 |
+
{
|
| 2489 |
+
"epoch": 0.43645206706770767,
|
| 2490 |
+
"grad_norm": 2.3978216648101807,
|
| 2491 |
+
"learning_rate": 1.898938187334092e-05,
|
| 2492 |
+
"loss": 0.5012,
|
| 2493 |
+
"step": 3410
|
| 2494 |
+
},
|
| 2495 |
+
{
|
| 2496 |
+
"epoch": 0.4377319851529502,
|
| 2497 |
+
"grad_norm": 0.838647186756134,
|
| 2498 |
+
"learning_rate": 1.8979901403109594e-05,
|
| 2499 |
+
"loss": 0.5042,
|
| 2500 |
+
"step": 3420
|
| 2501 |
+
},
|
| 2502 |
+
{
|
| 2503 |
+
"epoch": 0.4390119032381928,
|
| 2504 |
+
"grad_norm": 0.7617325782775879,
|
| 2505 |
+
"learning_rate": 1.897042093287827e-05,
|
| 2506 |
+
"loss": 0.5094,
|
| 2507 |
+
"step": 3430
|
| 2508 |
+
},
|
| 2509 |
+
{
|
| 2510 |
+
"epoch": 0.4402918213234353,
|
| 2511 |
+
"grad_norm": 0.8923324942588806,
|
| 2512 |
+
"learning_rate": 1.896094046264695e-05,
|
| 2513 |
+
"loss": 0.487,
|
| 2514 |
+
"step": 3440
|
| 2515 |
+
},
|
| 2516 |
+
{
|
| 2517 |
+
"epoch": 0.44157173940867783,
|
| 2518 |
+
"grad_norm": 1.7616841793060303,
|
| 2519 |
+
"learning_rate": 1.8951459992415627e-05,
|
| 2520 |
+
"loss": 0.4929,
|
| 2521 |
+
"step": 3450
|
| 2522 |
+
},
|
| 2523 |
+
{
|
| 2524 |
+
"epoch": 0.4428516574939204,
|
| 2525 |
+
"grad_norm": 0.7590444087982178,
|
| 2526 |
+
"learning_rate": 1.89419795221843e-05,
|
| 2527 |
+
"loss": 0.4844,
|
| 2528 |
+
"step": 3460
|
| 2529 |
+
},
|
| 2530 |
+
{
|
| 2531 |
+
"epoch": 0.44413157557916294,
|
| 2532 |
+
"grad_norm": 0.8714098930358887,
|
| 2533 |
+
"learning_rate": 1.893249905195298e-05,
|
| 2534 |
+
"loss": 0.4736,
|
| 2535 |
+
"step": 3470
|
| 2536 |
+
},
|
| 2537 |
+
{
|
| 2538 |
+
"epoch": 0.44541149366440547,
|
| 2539 |
+
"grad_norm": 1.3062705993652344,
|
| 2540 |
+
"learning_rate": 1.8923018581721656e-05,
|
| 2541 |
+
"loss": 0.501,
|
| 2542 |
+
"step": 3480
|
| 2543 |
+
},
|
| 2544 |
+
{
|
| 2545 |
+
"epoch": 0.44669141174964805,
|
| 2546 |
+
"grad_norm": 1.3274645805358887,
|
| 2547 |
+
"learning_rate": 1.891353811149033e-05,
|
| 2548 |
+
"loss": 0.506,
|
| 2549 |
+
"step": 3490
|
| 2550 |
+
},
|
| 2551 |
+
{
|
| 2552 |
+
"epoch": 0.4479713298348906,
|
| 2553 |
+
"grad_norm": 3.321751832962036,
|
| 2554 |
+
"learning_rate": 1.8904057641259008e-05,
|
| 2555 |
+
"loss": 0.5244,
|
| 2556 |
+
"step": 3500
|
| 2557 |
+
},
|
| 2558 |
+
{
|
| 2559 |
+
"epoch": 0.4492512479201331,
|
| 2560 |
+
"grad_norm": 0.8489649891853333,
|
| 2561 |
+
"learning_rate": 1.8894577171027683e-05,
|
| 2562 |
+
"loss": 0.5003,
|
| 2563 |
+
"step": 3510
|
| 2564 |
+
},
|
| 2565 |
+
{
|
| 2566 |
+
"epoch": 0.4505311660053757,
|
| 2567 |
+
"grad_norm": 1.0105081796646118,
|
| 2568 |
+
"learning_rate": 1.888509670079636e-05,
|
| 2569 |
+
"loss": 0.4748,
|
| 2570 |
+
"step": 3520
|
| 2571 |
+
},
|
| 2572 |
+
{
|
| 2573 |
+
"epoch": 0.4518110840906182,
|
| 2574 |
+
"grad_norm": 1.452003836631775,
|
| 2575 |
+
"learning_rate": 1.8875616230565038e-05,
|
| 2576 |
+
"loss": 0.4842,
|
| 2577 |
+
"step": 3530
|
| 2578 |
+
},
|
| 2579 |
+
{
|
| 2580 |
+
"epoch": 0.45309100217586074,
|
| 2581 |
+
"grad_norm": 0.8805971741676331,
|
| 2582 |
+
"learning_rate": 1.8866135760333716e-05,
|
| 2583 |
+
"loss": 0.4923,
|
| 2584 |
+
"step": 3540
|
| 2585 |
+
},
|
| 2586 |
+
{
|
| 2587 |
+
"epoch": 0.4543709202611033,
|
| 2588 |
+
"grad_norm": 2.563197135925293,
|
| 2589 |
+
"learning_rate": 1.885665529010239e-05,
|
| 2590 |
+
"loss": 0.5021,
|
| 2591 |
+
"step": 3550
|
| 2592 |
+
},
|
| 2593 |
+
{
|
| 2594 |
+
"epoch": 0.45565083834634584,
|
| 2595 |
+
"grad_norm": 0.7883086204528809,
|
| 2596 |
+
"learning_rate": 1.8847174819871067e-05,
|
| 2597 |
+
"loss": 0.4891,
|
| 2598 |
+
"step": 3560
|
| 2599 |
+
},
|
| 2600 |
+
{
|
| 2601 |
+
"epoch": 0.45693075643158837,
|
| 2602 |
+
"grad_norm": 1.5375005006790161,
|
| 2603 |
+
"learning_rate": 1.8837694349639745e-05,
|
| 2604 |
+
"loss": 0.486,
|
| 2605 |
+
"step": 3570
|
| 2606 |
+
},
|
| 2607 |
+
{
|
| 2608 |
+
"epoch": 0.4582106745168309,
|
| 2609 |
+
"grad_norm": 1.0281457901000977,
|
| 2610 |
+
"learning_rate": 1.8828213879408423e-05,
|
| 2611 |
+
"loss": 0.5057,
|
| 2612 |
+
"step": 3580
|
| 2613 |
+
},
|
| 2614 |
+
{
|
| 2615 |
+
"epoch": 0.4594905926020735,
|
| 2616 |
+
"grad_norm": 1.281043291091919,
|
| 2617 |
+
"learning_rate": 1.8818733409177097e-05,
|
| 2618 |
+
"loss": 0.4998,
|
| 2619 |
+
"step": 3590
|
| 2620 |
+
},
|
| 2621 |
+
{
|
| 2622 |
+
"epoch": 0.460770510687316,
|
| 2623 |
+
"grad_norm": 0.9816468954086304,
|
| 2624 |
+
"learning_rate": 1.880925293894577e-05,
|
| 2625 |
+
"loss": 0.4944,
|
| 2626 |
+
"step": 3600
|
| 2627 |
+
},
|
| 2628 |
+
{
|
| 2629 |
+
"epoch": 0.46205042877255853,
|
| 2630 |
+
"grad_norm": 0.9405584335327148,
|
| 2631 |
+
"learning_rate": 1.879977246871445e-05,
|
| 2632 |
+
"loss": 0.5036,
|
| 2633 |
+
"step": 3610
|
| 2634 |
+
},
|
| 2635 |
+
{
|
| 2636 |
+
"epoch": 0.4633303468578011,
|
| 2637 |
+
"grad_norm": 1.3431020975112915,
|
| 2638 |
+
"learning_rate": 1.8790291998483127e-05,
|
| 2639 |
+
"loss": 0.5004,
|
| 2640 |
+
"step": 3620
|
| 2641 |
+
},
|
| 2642 |
+
{
|
| 2643 |
+
"epoch": 0.46461026494304364,
|
| 2644 |
+
"grad_norm": 1.1362318992614746,
|
| 2645 |
+
"learning_rate": 1.8780811528251804e-05,
|
| 2646 |
+
"loss": 0.494,
|
| 2647 |
+
"step": 3630
|
| 2648 |
+
},
|
| 2649 |
+
{
|
| 2650 |
+
"epoch": 0.46589018302828616,
|
| 2651 |
+
"grad_norm": 1.2532789707183838,
|
| 2652 |
+
"learning_rate": 1.877133105802048e-05,
|
| 2653 |
+
"loss": 0.4748,
|
| 2654 |
+
"step": 3640
|
| 2655 |
+
},
|
| 2656 |
+
{
|
| 2657 |
+
"epoch": 0.46717010111352875,
|
| 2658 |
+
"grad_norm": 0.729250431060791,
|
| 2659 |
+
"learning_rate": 1.8761850587789156e-05,
|
| 2660 |
+
"loss": 0.5086,
|
| 2661 |
+
"step": 3650
|
| 2662 |
+
},
|
| 2663 |
+
{
|
| 2664 |
+
"epoch": 0.4684500191987713,
|
| 2665 |
+
"grad_norm": 0.6976643800735474,
|
| 2666 |
+
"learning_rate": 1.875237011755783e-05,
|
| 2667 |
+
"loss": 0.4781,
|
| 2668 |
+
"step": 3660
|
| 2669 |
+
},
|
| 2670 |
+
{
|
| 2671 |
+
"epoch": 0.4697299372840138,
|
| 2672 |
+
"grad_norm": 0.9878288507461548,
|
| 2673 |
+
"learning_rate": 1.8742889647326508e-05,
|
| 2674 |
+
"loss": 0.4809,
|
| 2675 |
+
"step": 3670
|
| 2676 |
+
},
|
| 2677 |
+
{
|
| 2678 |
+
"epoch": 0.4710098553692564,
|
| 2679 |
+
"grad_norm": 0.9907483458518982,
|
| 2680 |
+
"learning_rate": 1.8733409177095186e-05,
|
| 2681 |
+
"loss": 0.4914,
|
| 2682 |
+
"step": 3680
|
| 2683 |
+
},
|
| 2684 |
+
{
|
| 2685 |
+
"epoch": 0.4722897734544989,
|
| 2686 |
+
"grad_norm": 1.046175479888916,
|
| 2687 |
+
"learning_rate": 1.872392870686386e-05,
|
| 2688 |
+
"loss": 0.5048,
|
| 2689 |
+
"step": 3690
|
| 2690 |
+
},
|
| 2691 |
+
{
|
| 2692 |
+
"epoch": 0.47356969153974143,
|
| 2693 |
+
"grad_norm": 1.8168991804122925,
|
| 2694 |
+
"learning_rate": 1.8714448236632538e-05,
|
| 2695 |
+
"loss": 0.4528,
|
| 2696 |
+
"step": 3700
|
| 2697 |
+
},
|
| 2698 |
+
{
|
| 2699 |
+
"epoch": 0.474849609624984,
|
| 2700 |
+
"grad_norm": 2.9127612113952637,
|
| 2701 |
+
"learning_rate": 1.8704967766401216e-05,
|
| 2702 |
+
"loss": 0.5226,
|
| 2703 |
+
"step": 3710
|
| 2704 |
+
},
|
| 2705 |
+
{
|
| 2706 |
+
"epoch": 0.47612952771022654,
|
| 2707 |
+
"grad_norm": 1.4959512948989868,
|
| 2708 |
+
"learning_rate": 1.8695487296169893e-05,
|
| 2709 |
+
"loss": 0.4727,
|
| 2710 |
+
"step": 3720
|
| 2711 |
+
},
|
| 2712 |
+
{
|
| 2713 |
+
"epoch": 0.47740944579546907,
|
| 2714 |
+
"grad_norm": 0.6817557215690613,
|
| 2715 |
+
"learning_rate": 1.8686006825938568e-05,
|
| 2716 |
+
"loss": 0.468,
|
| 2717 |
+
"step": 3730
|
| 2718 |
+
},
|
| 2719 |
+
{
|
| 2720 |
+
"epoch": 0.47868936388071165,
|
| 2721 |
+
"grad_norm": 1.5274559259414673,
|
| 2722 |
+
"learning_rate": 1.8676526355707245e-05,
|
| 2723 |
+
"loss": 0.5158,
|
| 2724 |
+
"step": 3740
|
| 2725 |
+
},
|
| 2726 |
+
{
|
| 2727 |
+
"epoch": 0.4799692819659542,
|
| 2728 |
+
"grad_norm": 1.8979755640029907,
|
| 2729 |
+
"learning_rate": 1.866704588547592e-05,
|
| 2730 |
+
"loss": 0.4628,
|
| 2731 |
+
"step": 3750
|
| 2732 |
+
},
|
| 2733 |
+
{
|
| 2734 |
+
"epoch": 0.4812492000511967,
|
| 2735 |
+
"grad_norm": 0.7936620712280273,
|
| 2736 |
+
"learning_rate": 1.8657565415244597e-05,
|
| 2737 |
+
"loss": 0.5159,
|
| 2738 |
+
"step": 3760
|
| 2739 |
+
},
|
| 2740 |
+
{
|
| 2741 |
+
"epoch": 0.4825291181364393,
|
| 2742 |
+
"grad_norm": 1.2421464920043945,
|
| 2743 |
+
"learning_rate": 1.8648084945013275e-05,
|
| 2744 |
+
"loss": 0.495,
|
| 2745 |
+
"step": 3770
|
| 2746 |
+
},
|
| 2747 |
+
{
|
| 2748 |
+
"epoch": 0.4838090362216818,
|
| 2749 |
+
"grad_norm": 0.8110264539718628,
|
| 2750 |
+
"learning_rate": 1.8638604474781953e-05,
|
| 2751 |
+
"loss": 0.514,
|
| 2752 |
+
"step": 3780
|
| 2753 |
+
},
|
| 2754 |
+
{
|
| 2755 |
+
"epoch": 0.48508895430692434,
|
| 2756 |
+
"grad_norm": 1.168480634689331,
|
| 2757 |
+
"learning_rate": 1.8629124004550627e-05,
|
| 2758 |
+
"loss": 0.4842,
|
| 2759 |
+
"step": 3790
|
| 2760 |
+
},
|
| 2761 |
+
{
|
| 2762 |
+
"epoch": 0.4863688723921669,
|
| 2763 |
+
"grad_norm": 1.4188495874404907,
|
| 2764 |
+
"learning_rate": 1.86196435343193e-05,
|
| 2765 |
+
"loss": 0.4855,
|
| 2766 |
+
"step": 3800
|
| 2767 |
+
},
|
| 2768 |
+
{
|
| 2769 |
+
"epoch": 0.48764879047740944,
|
| 2770 |
+
"grad_norm": 1.9594807624816895,
|
| 2771 |
+
"learning_rate": 1.8610163064087982e-05,
|
| 2772 |
+
"loss": 0.486,
|
| 2773 |
+
"step": 3810
|
| 2774 |
+
},
|
| 2775 |
+
{
|
| 2776 |
+
"epoch": 0.48892870856265197,
|
| 2777 |
+
"grad_norm": 0.8144770264625549,
|
| 2778 |
+
"learning_rate": 1.8600682593856656e-05,
|
| 2779 |
+
"loss": 0.4939,
|
| 2780 |
+
"step": 3820
|
| 2781 |
+
},
|
| 2782 |
+
{
|
| 2783 |
+
"epoch": 0.49020862664789455,
|
| 2784 |
+
"grad_norm": 1.4995648860931396,
|
| 2785 |
+
"learning_rate": 1.8591202123625334e-05,
|
| 2786 |
+
"loss": 0.4912,
|
| 2787 |
+
"step": 3830
|
| 2788 |
+
},
|
| 2789 |
+
{
|
| 2790 |
+
"epoch": 0.4914885447331371,
|
| 2791 |
+
"grad_norm": 1.1382790803909302,
|
| 2792 |
+
"learning_rate": 1.858172165339401e-05,
|
| 2793 |
+
"loss": 0.4634,
|
| 2794 |
+
"step": 3840
|
| 2795 |
+
},
|
| 2796 |
+
{
|
| 2797 |
+
"epoch": 0.4927684628183796,
|
| 2798 |
+
"grad_norm": 0.8508588075637817,
|
| 2799 |
+
"learning_rate": 1.8572241183162686e-05,
|
| 2800 |
+
"loss": 0.5236,
|
| 2801 |
+
"step": 3850
|
| 2802 |
+
},
|
| 2803 |
+
{
|
| 2804 |
+
"epoch": 0.4940483809036222,
|
| 2805 |
+
"grad_norm": 1.2297135591506958,
|
| 2806 |
+
"learning_rate": 1.8562760712931364e-05,
|
| 2807 |
+
"loss": 0.4806,
|
| 2808 |
+
"step": 3860
|
| 2809 |
+
},
|
| 2810 |
+
{
|
| 2811 |
+
"epoch": 0.4953282989888647,
|
| 2812 |
+
"grad_norm": 0.863645076751709,
|
| 2813 |
+
"learning_rate": 1.855328024270004e-05,
|
| 2814 |
+
"loss": 0.4713,
|
| 2815 |
+
"step": 3870
|
| 2816 |
+
},
|
| 2817 |
+
{
|
| 2818 |
+
"epoch": 0.49660821707410724,
|
| 2819 |
+
"grad_norm": 1.6313464641571045,
|
| 2820 |
+
"learning_rate": 1.8543799772468716e-05,
|
| 2821 |
+
"loss": 0.4734,
|
| 2822 |
+
"step": 3880
|
| 2823 |
+
},
|
| 2824 |
+
{
|
| 2825 |
+
"epoch": 0.4978881351593498,
|
| 2826 |
+
"grad_norm": 3.317040205001831,
|
| 2827 |
+
"learning_rate": 1.8534319302237393e-05,
|
| 2828 |
+
"loss": 0.495,
|
| 2829 |
+
"step": 3890
|
| 2830 |
+
},
|
| 2831 |
+
{
|
| 2832 |
+
"epoch": 0.49916805324459235,
|
| 2833 |
+
"grad_norm": 0.8402191996574402,
|
| 2834 |
+
"learning_rate": 1.8524838832006068e-05,
|
| 2835 |
+
"loss": 0.4916,
|
| 2836 |
+
"step": 3900
|
| 2837 |
+
},
|
| 2838 |
+
{
|
| 2839 |
+
"epoch": 0.5004479713298349,
|
| 2840 |
+
"grad_norm": 2.1389431953430176,
|
| 2841 |
+
"learning_rate": 1.8515358361774745e-05,
|
| 2842 |
+
"loss": 0.4868,
|
| 2843 |
+
"step": 3910
|
| 2844 |
+
},
|
| 2845 |
+
{
|
| 2846 |
+
"epoch": 0.5017278894150774,
|
| 2847 |
+
"grad_norm": 1.6004793643951416,
|
| 2848 |
+
"learning_rate": 1.8505877891543423e-05,
|
| 2849 |
+
"loss": 0.5038,
|
| 2850 |
+
"step": 3920
|
| 2851 |
+
},
|
| 2852 |
+
{
|
| 2853 |
+
"epoch": 0.5030078075003199,
|
| 2854 |
+
"grad_norm": 1.4133943319320679,
|
| 2855 |
+
"learning_rate": 1.8496397421312097e-05,
|
| 2856 |
+
"loss": 0.5067,
|
| 2857 |
+
"step": 3930
|
| 2858 |
+
},
|
| 2859 |
+
{
|
| 2860 |
+
"epoch": 0.5042877255855626,
|
| 2861 |
+
"grad_norm": 2.345825433731079,
|
| 2862 |
+
"learning_rate": 1.8486916951080775e-05,
|
| 2863 |
+
"loss": 0.4888,
|
| 2864 |
+
"step": 3940
|
| 2865 |
+
},
|
| 2866 |
+
{
|
| 2867 |
+
"epoch": 0.5055676436708051,
|
| 2868 |
+
"grad_norm": 0.8149171471595764,
|
| 2869 |
+
"learning_rate": 1.8477436480849453e-05,
|
| 2870 |
+
"loss": 0.4847,
|
| 2871 |
+
"step": 3950
|
| 2872 |
+
},
|
| 2873 |
+
{
|
| 2874 |
+
"epoch": 0.5068475617560476,
|
| 2875 |
+
"grad_norm": 0.8076463937759399,
|
| 2876 |
+
"learning_rate": 1.846795601061813e-05,
|
| 2877 |
+
"loss": 0.4786,
|
| 2878 |
+
"step": 3960
|
| 2879 |
+
},
|
| 2880 |
+
{
|
| 2881 |
+
"epoch": 0.5081274798412901,
|
| 2882 |
+
"grad_norm": 0.7464845180511475,
|
| 2883 |
+
"learning_rate": 1.8458475540386805e-05,
|
| 2884 |
+
"loss": 0.5039,
|
| 2885 |
+
"step": 3970
|
| 2886 |
+
},
|
| 2887 |
+
{
|
| 2888 |
+
"epoch": 0.5094073979265327,
|
| 2889 |
+
"grad_norm": 0.9748256802558899,
|
| 2890 |
+
"learning_rate": 1.8448995070155482e-05,
|
| 2891 |
+
"loss": 0.4732,
|
| 2892 |
+
"step": 3980
|
| 2893 |
+
},
|
| 2894 |
+
{
|
| 2895 |
+
"epoch": 0.5106873160117752,
|
| 2896 |
+
"grad_norm": 1.2856113910675049,
|
| 2897 |
+
"learning_rate": 1.8439514599924156e-05,
|
| 2898 |
+
"loss": 0.4914,
|
| 2899 |
+
"step": 3990
|
| 2900 |
+
},
|
| 2901 |
+
{
|
| 2902 |
+
"epoch": 0.5119672340970178,
|
| 2903 |
+
"grad_norm": 0.9492047429084778,
|
| 2904 |
+
"learning_rate": 1.8430034129692834e-05,
|
| 2905 |
+
"loss": 0.5132,
|
| 2906 |
+
"step": 4000
|
| 2907 |
+
},
|
| 2908 |
+
{
|
| 2909 |
+
"epoch": 0.5119672340970178,
|
| 2910 |
+
"eval_accuracy": 0.79022,
|
| 2911 |
+
"eval_loss": 0.44629669189453125,
|
| 2912 |
+
"eval_macro_f1": 0.7899137546751205,
|
| 2913 |
+
"eval_precision": 0.7975042104934613,
|
| 2914 |
+
"eval_recall": 0.7936668066341269,
|
| 2915 |
+
"eval_runtime": 705.8506,
|
| 2916 |
+
"eval_samples_per_second": 141.673,
|
| 2917 |
+
"eval_steps_per_second": 1.108,
|
| 2918 |
+
"step": 4000
|
| 2919 |
+
}
|
| 2920 |
+
],
|
| 2921 |
+
"logging_steps": 10,
|
| 2922 |
+
"max_steps": 23439,
|
| 2923 |
+
"num_input_tokens_seen": 0,
|
| 2924 |
+
"num_train_epochs": 3,
|
| 2925 |
+
"save_steps": 200,
|
| 2926 |
+
"stateful_callbacks": {
|
| 2927 |
+
"EarlyStoppingCallback": {
|
| 2928 |
+
"args": {
|
| 2929 |
+
"early_stopping_patience": 3,
|
| 2930 |
+
"early_stopping_threshold": 0.0
|
| 2931 |
+
},
|
| 2932 |
+
"attributes": {
|
| 2933 |
+
"early_stopping_patience_counter": 0
|
| 2934 |
+
}
|
| 2935 |
+
},
|
| 2936 |
+
"TrainerControl": {
|
| 2937 |
+
"args": {
|
| 2938 |
+
"should_epoch_stop": false,
|
| 2939 |
+
"should_evaluate": false,
|
| 2940 |
+
"should_log": false,
|
| 2941 |
+
"should_save": true,
|
| 2942 |
+
"should_training_stop": false
|
| 2943 |
+
},
|
| 2944 |
+
"attributes": {}
|
| 2945 |
+
}
|
| 2946 |
+
},
|
| 2947 |
+
"total_flos": 6.735643017216e+16,
|
| 2948 |
+
"train_batch_size": 64,
|
| 2949 |
+
"trial_name": null,
|
| 2950 |
+
"trial_params": null
|
| 2951 |
+
}
|
codebert-base-cross-entropy/checkpoint-4000/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:dbdc2f4c9cd8ad671ca740ab491512a481c34d40fafc15bd45db45cd63388bf1
|
| 3 |
+
size 5841
|
codebert-base-cross-entropy/checkpoint-4000/vocab.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
codebert-base-cross-entropy/training.log
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
-
2026-04-15
|
| 2 |
-
2026-04-15
|
| 3 |
-
2026-04-15
|
| 4 |
-
2026-04-15
|
| 5 |
-
2026-04-15
|
| 6 |
RobertaForSequenceClassification(
|
| 7 |
(roberta): RobertaModel(
|
| 8 |
(embeddings): RobertaEmbeddings(
|
|
@@ -47,16 +47,16 @@ RobertaForSequenceClassification(
|
|
| 47 |
(out_proj): Linear(in_features=768, out_features=2, bias=True)
|
| 48 |
)
|
| 49 |
)
|
| 50 |
-
2026-04-15
|
| 51 |
-
2026-04-15
|
| 52 |
-
2026-04-15
|
| 53 |
-
2026-04-15
|
| 54 |
-
2026-04-15
|
| 55 |
-
2026-04-15
|
| 56 |
-
2026-04-15
|
| 57 |
-
2026-04-15
|
| 58 |
-
2026-04-15
|
| 59 |
-
2026-04-15
|
| 60 |
RobertaForSequenceClassification(
|
| 61 |
(roberta): RobertaModel(
|
| 62 |
(embeddings): RobertaEmbeddings(
|
|
@@ -101,16 +101,16 @@ RobertaForSequenceClassification(
|
|
| 101 |
(out_proj): Linear(in_features=768, out_features=2, bias=True)
|
| 102 |
)
|
| 103 |
)
|
| 104 |
-
2026-04-15
|
| 105 |
-
2026-04-15
|
| 106 |
-
2026-04-15
|
| 107 |
-
2026-04-15
|
| 108 |
-
2026-04-15
|
| 109 |
-
2026-04-15
|
| 110 |
-
2026-04-15
|
| 111 |
-
2026-04-15
|
| 112 |
-
2026-04-15
|
| 113 |
-
2026-04-15
|
| 114 |
RobertaForSequenceClassification(
|
| 115 |
(roberta): RobertaModel(
|
| 116 |
(embeddings): RobertaEmbeddings(
|
|
@@ -155,16 +155,16 @@ RobertaForSequenceClassification(
|
|
| 155 |
(out_proj): Linear(in_features=768, out_features=2, bias=True)
|
| 156 |
)
|
| 157 |
)
|
| 158 |
-
2026-04-15
|
| 159 |
-
2026-04-15
|
| 160 |
-
2026-04-15
|
| 161 |
-
2026-04-15
|
| 162 |
-
2026-04-15
|
| 163 |
-
2026-04-15
|
| 164 |
-
2026-04-15
|
| 165 |
-
2026-04-15
|
| 166 |
-
2026-04-15
|
| 167 |
-
2026-04-15
|
| 168 |
RobertaForSequenceClassification(
|
| 169 |
(roberta): RobertaModel(
|
| 170 |
(embeddings): RobertaEmbeddings(
|
|
@@ -209,16 +209,16 @@ RobertaForSequenceClassification(
|
|
| 209 |
(out_proj): Linear(in_features=768, out_features=2, bias=True)
|
| 210 |
)
|
| 211 |
)
|
| 212 |
-
2026-04-15
|
| 213 |
-
2026-04-15
|
| 214 |
-
2026-04-15
|
| 215 |
-
2026-04-15
|
| 216 |
-
2026-04-15
|
| 217 |
-
2026-04-15
|
| 218 |
-
2026-04-15
|
| 219 |
-
2026-04-15
|
| 220 |
-
2026-04-15
|
| 221 |
-
2026-04-15
|
| 222 |
RobertaForSequenceClassification(
|
| 223 |
(roberta): RobertaModel(
|
| 224 |
(embeddings): RobertaEmbeddings(
|
|
@@ -263,16 +263,16 @@ RobertaForSequenceClassification(
|
|
| 263 |
(out_proj): Linear(in_features=768, out_features=2, bias=True)
|
| 264 |
)
|
| 265 |
)
|
| 266 |
-
2026-04-15
|
| 267 |
-
2026-04-15
|
| 268 |
-
2026-04-15
|
| 269 |
-
2026-04-15
|
| 270 |
-
2026-04-15
|
| 271 |
-
2026-04-15
|
| 272 |
-
2026-04-15
|
| 273 |
-
2026-04-15
|
| 274 |
-
2026-04-15
|
| 275 |
-
2026-04-15
|
| 276 |
RobertaForSequenceClassification(
|
| 277 |
(roberta): RobertaModel(
|
| 278 |
(embeddings): RobertaEmbeddings(
|
|
@@ -317,8 +317,62 @@ RobertaForSequenceClassification(
|
|
| 317 |
(out_proj): Linear(in_features=768, out_features=2, bias=True)
|
| 318 |
)
|
| 319 |
)
|
| 320 |
-
2026-04-15
|
| 321 |
-
2026-04-15
|
| 322 |
-
2026-04-15
|
| 323 |
-
2026-04-15
|
| 324 |
-
2026-04-15
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
2026-04-15 12:35:23,651 - INFO - train_pipeline - Logging to ./output_checkpoints/codebert-base-cross-entropy/training.log
|
| 2 |
+
2026-04-15 12:35:23,668 - INFO - train_pipeline - Loading model & tokenizer for 'microsoft/codebert-base'
|
| 3 |
+
2026-04-15 12:35:25,145 - INFO - train_pipeline - Model placed on cuda
|
| 4 |
+
2026-04-15 12:35:25,156 - INFO - train_pipeline - ===== Model Architecture =====
|
| 5 |
+
2026-04-15 12:35:25,169 - INFO - train_pipeline -
|
| 6 |
RobertaForSequenceClassification(
|
| 7 |
(roberta): RobertaModel(
|
| 8 |
(embeddings): RobertaEmbeddings(
|
|
|
|
| 47 |
(out_proj): Linear(in_features=768, out_features=2, bias=True)
|
| 48 |
)
|
| 49 |
)
|
| 50 |
+
2026-04-15 12:35:25,191 - INFO - train_pipeline - ===== Tokenizer Summary =====
|
| 51 |
+
2026-04-15 12:35:25,243 - INFO - train_pipeline - Vocab size: 50265 | Special tokens: ['<s>', '</s>', '<unk>', '<pad>', '<mask>']
|
| 52 |
+
2026-04-15 12:35:25,254 - INFO - train_pipeline - ===== End of Architecture Log =====
|
| 53 |
+
2026-04-15 12:35:25,265 - INFO - train_pipeline - Base model weights frozen – only classifier head will be trained.
|
| 54 |
+
2026-04-15 12:35:26,856 - INFO - train_pipeline - === Starting training ===
|
| 55 |
+
2026-04-15 12:35:23,651 - INFO - train_pipeline - Logging to ./output_checkpoints/codebert-base-cross-entropy/training.log
|
| 56 |
+
2026-04-15 12:35:23,668 - INFO - train_pipeline - Loading model & tokenizer for 'microsoft/codebert-base'
|
| 57 |
+
2026-04-15 12:35:25,145 - INFO - train_pipeline - Model placed on cuda
|
| 58 |
+
2026-04-15 12:35:25,156 - INFO - train_pipeline - ===== Model Architecture =====
|
| 59 |
+
2026-04-15 12:35:25,169 - INFO - train_pipeline -
|
| 60 |
RobertaForSequenceClassification(
|
| 61 |
(roberta): RobertaModel(
|
| 62 |
(embeddings): RobertaEmbeddings(
|
|
|
|
| 101 |
(out_proj): Linear(in_features=768, out_features=2, bias=True)
|
| 102 |
)
|
| 103 |
)
|
| 104 |
+
2026-04-15 12:35:25,191 - INFO - train_pipeline - ===== Tokenizer Summary =====
|
| 105 |
+
2026-04-15 12:35:25,243 - INFO - train_pipeline - Vocab size: 50265 | Special tokens: ['<s>', '</s>', '<unk>', '<pad>', '<mask>']
|
| 106 |
+
2026-04-15 12:35:25,254 - INFO - train_pipeline - ===== End of Architecture Log =====
|
| 107 |
+
2026-04-15 12:35:25,265 - INFO - train_pipeline - Base model weights frozen – only classifier head will be trained.
|
| 108 |
+
2026-04-15 12:35:26,856 - INFO - train_pipeline - === Starting training ===
|
| 109 |
+
2026-04-15 12:35:23,651 - INFO - train_pipeline - Logging to ./output_checkpoints/codebert-base-cross-entropy/training.log
|
| 110 |
+
2026-04-15 12:35:23,668 - INFO - train_pipeline - Loading model & tokenizer for 'microsoft/codebert-base'
|
| 111 |
+
2026-04-15 12:35:25,145 - INFO - train_pipeline - Model placed on cuda
|
| 112 |
+
2026-04-15 12:35:25,156 - INFO - train_pipeline - ===== Model Architecture =====
|
| 113 |
+
2026-04-15 12:35:25,169 - INFO - train_pipeline -
|
| 114 |
RobertaForSequenceClassification(
|
| 115 |
(roberta): RobertaModel(
|
| 116 |
(embeddings): RobertaEmbeddings(
|
|
|
|
| 155 |
(out_proj): Linear(in_features=768, out_features=2, bias=True)
|
| 156 |
)
|
| 157 |
)
|
| 158 |
+
2026-04-15 12:35:25,191 - INFO - train_pipeline - ===== Tokenizer Summary =====
|
| 159 |
+
2026-04-15 12:35:25,243 - INFO - train_pipeline - Vocab size: 50265 | Special tokens: ['<s>', '</s>', '<unk>', '<pad>', '<mask>']
|
| 160 |
+
2026-04-15 12:35:25,254 - INFO - train_pipeline - ===== End of Architecture Log =====
|
| 161 |
+
2026-04-15 12:35:25,265 - INFO - train_pipeline - Base model weights frozen – only classifier head will be trained.
|
| 162 |
+
2026-04-15 12:35:26,856 - INFO - train_pipeline - === Starting training ===
|
| 163 |
+
2026-04-15 12:35:23,651 - INFO - train_pipeline - Logging to ./output_checkpoints/codebert-base-cross-entropy/training.log
|
| 164 |
+
2026-04-15 12:35:23,668 - INFO - train_pipeline - Loading model & tokenizer for 'microsoft/codebert-base'
|
| 165 |
+
2026-04-15 12:35:25,145 - INFO - train_pipeline - Model placed on cuda
|
| 166 |
+
2026-04-15 12:35:25,156 - INFO - train_pipeline - ===== Model Architecture =====
|
| 167 |
+
2026-04-15 12:35:25,169 - INFO - train_pipeline -
|
| 168 |
RobertaForSequenceClassification(
|
| 169 |
(roberta): RobertaModel(
|
| 170 |
(embeddings): RobertaEmbeddings(
|
|
|
|
| 209 |
(out_proj): Linear(in_features=768, out_features=2, bias=True)
|
| 210 |
)
|
| 211 |
)
|
| 212 |
+
2026-04-15 12:35:25,191 - INFO - train_pipeline - ===== Tokenizer Summary =====
|
| 213 |
+
2026-04-15 12:35:25,243 - INFO - train_pipeline - Vocab size: 50265 | Special tokens: ['<s>', '</s>', '<unk>', '<pad>', '<mask>']
|
| 214 |
+
2026-04-15 12:35:25,254 - INFO - train_pipeline - ===== End of Architecture Log =====
|
| 215 |
+
2026-04-15 12:35:25,265 - INFO - train_pipeline - Base model weights frozen – only classifier head will be trained.
|
| 216 |
+
2026-04-15 12:35:26,856 - INFO - train_pipeline - === Starting training ===
|
| 217 |
+
2026-04-15 12:35:23,651 - INFO - train_pipeline - Logging to ./output_checkpoints/codebert-base-cross-entropy/training.log
|
| 218 |
+
2026-04-15 12:35:23,668 - INFO - train_pipeline - Loading model & tokenizer for 'microsoft/codebert-base'
|
| 219 |
+
2026-04-15 12:35:25,145 - INFO - train_pipeline - Model placed on cuda
|
| 220 |
+
2026-04-15 12:35:25,156 - INFO - train_pipeline - ===== Model Architecture =====
|
| 221 |
+
2026-04-15 12:35:25,169 - INFO - train_pipeline -
|
| 222 |
RobertaForSequenceClassification(
|
| 223 |
(roberta): RobertaModel(
|
| 224 |
(embeddings): RobertaEmbeddings(
|
|
|
|
| 263 |
(out_proj): Linear(in_features=768, out_features=2, bias=True)
|
| 264 |
)
|
| 265 |
)
|
| 266 |
+
2026-04-15 12:35:25,191 - INFO - train_pipeline - ===== Tokenizer Summary =====
|
| 267 |
+
2026-04-15 12:35:25,243 - INFO - train_pipeline - Vocab size: 50265 | Special tokens: ['<s>', '</s>', '<unk>', '<pad>', '<mask>']
|
| 268 |
+
2026-04-15 12:35:25,254 - INFO - train_pipeline - ===== End of Architecture Log =====
|
| 269 |
+
2026-04-15 12:35:25,265 - INFO - train_pipeline - Base model weights frozen – only classifier head will be trained.
|
| 270 |
+
2026-04-15 12:35:26,856 - INFO - train_pipeline - === Starting training ===
|
| 271 |
+
2026-04-15 12:35:23,651 - INFO - train_pipeline - Logging to ./output_checkpoints/codebert-base-cross-entropy/training.log
|
| 272 |
+
2026-04-15 12:35:23,668 - INFO - train_pipeline - Loading model & tokenizer for 'microsoft/codebert-base'
|
| 273 |
+
2026-04-15 12:35:25,145 - INFO - train_pipeline - Model placed on cuda
|
| 274 |
+
2026-04-15 12:35:25,156 - INFO - train_pipeline - ===== Model Architecture =====
|
| 275 |
+
2026-04-15 12:35:25,169 - INFO - train_pipeline -
|
| 276 |
RobertaForSequenceClassification(
|
| 277 |
(roberta): RobertaModel(
|
| 278 |
(embeddings): RobertaEmbeddings(
|
|
|
|
| 317 |
(out_proj): Linear(in_features=768, out_features=2, bias=True)
|
| 318 |
)
|
| 319 |
)
|
| 320 |
+
2026-04-15 12:35:25,191 - INFO - train_pipeline - ===== Tokenizer Summary =====
|
| 321 |
+
2026-04-15 12:35:25,243 - INFO - train_pipeline - Vocab size: 50265 | Special tokens: ['<s>', '</s>', '<unk>', '<pad>', '<mask>']
|
| 322 |
+
2026-04-15 12:35:25,254 - INFO - train_pipeline - ===== End of Architecture Log =====
|
| 323 |
+
2026-04-15 12:35:25,265 - INFO - train_pipeline - Base model weights frozen – only classifier head will be trained.
|
| 324 |
+
2026-04-15 12:35:26,856 - INFO - train_pipeline - === Starting training ===
|
| 325 |
+
2026-04-15 12:35:23,651 - INFO - train_pipeline - Logging to ./output_checkpoints/codebert-base-cross-entropy/training.log
|
| 326 |
+
2026-04-15 12:35:23,668 - INFO - train_pipeline - Loading model & tokenizer for 'microsoft/codebert-base'
|
| 327 |
+
2026-04-15 12:35:25,145 - INFO - train_pipeline - Model placed on cuda
|
| 328 |
+
2026-04-15 12:35:25,156 - INFO - train_pipeline - ===== Model Architecture =====
|
| 329 |
+
2026-04-15 12:35:25,169 - INFO - train_pipeline -
|
| 330 |
+
RobertaForSequenceClassification(
|
| 331 |
+
(roberta): RobertaModel(
|
| 332 |
+
(embeddings): RobertaEmbeddings(
|
| 333 |
+
(word_embeddings): Embedding(50265, 768, padding_idx=1)
|
| 334 |
+
(position_embeddings): Embedding(514, 768, padding_idx=1)
|
| 335 |
+
(token_type_embeddings): Embedding(1, 768)
|
| 336 |
+
(LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
|
| 337 |
+
(dropout): Dropout(p=0.1, inplace=False)
|
| 338 |
+
)
|
| 339 |
+
(encoder): RobertaEncoder(
|
| 340 |
+
(layer): ModuleList(
|
| 341 |
+
(0-11): 12 x RobertaLayer(
|
| 342 |
+
(attention): RobertaAttention(
|
| 343 |
+
(self): RobertaSdpaSelfAttention(
|
| 344 |
+
(query): Linear(in_features=768, out_features=768, bias=True)
|
| 345 |
+
(key): Linear(in_features=768, out_features=768, bias=True)
|
| 346 |
+
(value): Linear(in_features=768, out_features=768, bias=True)
|
| 347 |
+
(dropout): Dropout(p=0.1, inplace=False)
|
| 348 |
+
)
|
| 349 |
+
(output): RobertaSelfOutput(
|
| 350 |
+
(dense): Linear(in_features=768, out_features=768, bias=True)
|
| 351 |
+
(LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
|
| 352 |
+
(dropout): Dropout(p=0.1, inplace=False)
|
| 353 |
+
)
|
| 354 |
+
)
|
| 355 |
+
(intermediate): RobertaIntermediate(
|
| 356 |
+
(dense): Linear(in_features=768, out_features=3072, bias=True)
|
| 357 |
+
(intermediate_act_fn): GELUActivation()
|
| 358 |
+
)
|
| 359 |
+
(output): RobertaOutput(
|
| 360 |
+
(dense): Linear(in_features=3072, out_features=768, bias=True)
|
| 361 |
+
(LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
|
| 362 |
+
(dropout): Dropout(p=0.1, inplace=False)
|
| 363 |
+
)
|
| 364 |
+
)
|
| 365 |
+
)
|
| 366 |
+
)
|
| 367 |
+
)
|
| 368 |
+
(classifier): RobertaClassificationHead(
|
| 369 |
+
(dense): Linear(in_features=768, out_features=768, bias=True)
|
| 370 |
+
(dropout): Dropout(p=0.1, inplace=False)
|
| 371 |
+
(out_proj): Linear(in_features=768, out_features=2, bias=True)
|
| 372 |
+
)
|
| 373 |
+
)
|
| 374 |
+
2026-04-15 12:35:25,191 - INFO - train_pipeline - ===== Tokenizer Summary =====
|
| 375 |
+
2026-04-15 12:35:25,243 - INFO - train_pipeline - Vocab size: 50265 | Special tokens: ['<s>', '</s>', '<unk>', '<pad>', '<mask>']
|
| 376 |
+
2026-04-15 12:35:25,254 - INFO - train_pipeline - ===== End of Architecture Log =====
|
| 377 |
+
2026-04-15 12:35:25,265 - INFO - train_pipeline - Base model weights frozen – only classifier head will be trained.
|
| 378 |
+
2026-04-15 12:35:26,856 - INFO - train_pipeline - === Starting training ===
|