Upload 7 files
Browse files- best/meta.json +4 -0
- best/model.pt +3 -0
- best/model_config.json +14 -0
- best/optimizer.pt +3 -0
- best/train_config.json +42 -0
- tokenizer/tokenizer.json +0 -0
- tokenizer/tokenizer_meta.json +18 -0
best/meta.json
ADDED
|
@@ -0,0 +1,4 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"step": 13500,
|
| 3 |
+
"best_val_loss": 3.1774327754974365
|
| 4 |
+
}
|
best/model.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f97768e8a3703727804e4fcd96fefe60934b2b34372beab87c9dbcda37a160be
|
| 3 |
+
size 133490789
|
best/model_config.json
ADDED
|
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"vocab_size": 16000,
|
| 3 |
+
"n_layers": 8,
|
| 4 |
+
"n_heads": 8,
|
| 5 |
+
"n_kv_heads": 4,
|
| 6 |
+
"d_model": 512,
|
| 7 |
+
"d_ff": 1536,
|
| 8 |
+
"context_len": 1024,
|
| 9 |
+
"dropout": 0.0,
|
| 10 |
+
"bias": false,
|
| 11 |
+
"norm_eps": 1e-05,
|
| 12 |
+
"rope_theta": 10000.0,
|
| 13 |
+
"tie_weights": true
|
| 14 |
+
}
|
best/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2ae45bd62440e2b9b749ede646a4e50c11061678507d6283ae2f7027d7081ebc
|
| 3 |
+
size 266996363
|
best/train_config.json
ADDED
|
@@ -0,0 +1,42 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"corpus": "data/turkish_corpus.txt",
|
| 3 |
+
"val_corpus": "None",
|
| 4 |
+
"tokenizer_path": "tokenizer/",
|
| 5 |
+
"output_dir": "runs/tr45m_pretrain",
|
| 6 |
+
"cache_dir": ".cache",
|
| 7 |
+
"resume": "None",
|
| 8 |
+
"strategy": "ddp",
|
| 9 |
+
"ds_stage": "2",
|
| 10 |
+
"cpu_offload": "False",
|
| 11 |
+
"preset": "small",
|
| 12 |
+
"n_layers": "12",
|
| 13 |
+
"n_heads": "12",
|
| 14 |
+
"n_kv_heads": "4",
|
| 15 |
+
"d_model": "768",
|
| 16 |
+
"context_len": "None",
|
| 17 |
+
"dropout": "None",
|
| 18 |
+
"no_tie_weights": "False",
|
| 19 |
+
"compile": "False",
|
| 20 |
+
"batch_size": "32",
|
| 21 |
+
"grad_accum": "2",
|
| 22 |
+
"max_steps": "13700",
|
| 23 |
+
"warmup_steps": "400",
|
| 24 |
+
"lr": "0.0003",
|
| 25 |
+
"min_lr": "3e-05",
|
| 26 |
+
"weight_decay": "0.1",
|
| 27 |
+
"beta1": "0.9",
|
| 28 |
+
"beta2": "0.95",
|
| 29 |
+
"grad_clip": "1.0",
|
| 30 |
+
"mixed_precision": "bf16",
|
| 31 |
+
"streaming": "False",
|
| 32 |
+
"num_workers": "4",
|
| 33 |
+
"seed": "42",
|
| 34 |
+
"log_every": "50",
|
| 35 |
+
"eval_every": "500",
|
| 36 |
+
"eval_batches": "100",
|
| 37 |
+
"save_every": "1000",
|
| 38 |
+
"wandb": "False",
|
| 39 |
+
"wandb_project": "my-llm",
|
| 40 |
+
"sample_prompt": "T\u00fcrkiye'nin ba\u015fkenti",
|
| 41 |
+
"sample_tokens": "100"
|
| 42 |
+
}
|
tokenizer/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer/tokenizer_meta.json
ADDED
|
@@ -0,0 +1,18 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"vocab_size": 16000,
|
| 3 |
+
"special_tokens": {
|
| 4 |
+
"<|bos|>": 0,
|
| 5 |
+
"<|eos|>": 1,
|
| 6 |
+
"<|pad|>": 2,
|
| 7 |
+
"<|sep|>": 3,
|
| 8 |
+
"<|unk|>": 4,
|
| 9 |
+
"<|sys|>": 5,
|
| 10 |
+
"<|usr|>": 6,
|
| 11 |
+
"<|ast|>": 7
|
| 12 |
+
},
|
| 13 |
+
"bos_id": 0,
|
| 14 |
+
"eos_id": 1,
|
| 15 |
+
"pad_id": 2,
|
| 16 |
+
"sep_id": 3,
|
| 17 |
+
"unk_id": 4
|
| 18 |
+
}
|