Upload folder using huggingface_hub
Browse filesThis view is limited to 50 files because it contains too many changes.
See raw diff
- ablation_initialization/gla_2_normal_initialization/config.json +20 -0
- ablation_initialization/gla_2_normal_initialization/pytorch_model-00001-of-00003.bin +3 -0
- ablation_initialization/gla_2_normal_initialization/pytorch_model-00002-of-00003.bin +3 -0
- ablation_initialization/gla_2_normal_initialization/pytorch_model-00003-of-00003.bin +3 -0
- ablation_initialization/gla_2_normal_initialization/pytorch_model.bin.index.json +346 -0
- ablation_initialization/gla_4_normal_initialization/config.json +20 -0
- ablation_initialization/gla_4_normal_initialization/pytorch_model-00001-of-00003.bin +3 -0
- ablation_initialization/gla_4_normal_initialization/pytorch_model-00002-of-00003.bin +3 -0
- ablation_initialization/gla_4_normal_initialization/pytorch_model-00003-of-00003.bin +3 -0
- ablation_initialization/gla_4_normal_initialization/pytorch_model.bin.index.json +370 -0
- ablation_initialization/gqa_normal_initialization/config.json +18 -0
- ablation_initialization/gqa_normal_initialization/pytorch_model-00001-of-00003.bin +3 -0
- ablation_initialization/gqa_normal_initialization/pytorch_model-00002-of-00003.bin +3 -0
- ablation_initialization/gqa_normal_initialization/pytorch_model-00003-of-00003.bin +3 -0
- ablation_initialization/gqa_normal_initialization/pytorch_model.bin.index.json +226 -0
- ablation_initialization/gta_normal_initialization/config.json +18 -0
- ablation_initialization/gta_normal_initialization/pytorch_model-00001-of-00003.bin +3 -0
- ablation_initialization/gta_normal_initialization/pytorch_model-00002-of-00003.bin +3 -0
- ablation_initialization/gta_normal_initialization/pytorch_model-00003-of-00003.bin +3 -0
- ablation_initialization/gta_normal_initialization/pytorch_model.bin.index.json +226 -0
- ablation_initialization/mha_normal_initialization/config.json +17 -0
- ablation_initialization/mha_normal_initialization/pytorch_model-00001-of-00003.bin +3 -0
- ablation_initialization/mha_normal_initialization/pytorch_model-00002-of-00003.bin +3 -0
- ablation_initialization/mha_normal_initialization/pytorch_model-00003-of-00003.bin +3 -0
- ablation_initialization/mha_normal_initialization/pytorch_model.bin.index.json +226 -0
- ablation_initialization/mla_normal_initialization/config.json +20 -0
- ablation_initialization/mla_normal_initialization/pytorch_model-00001-of-00003.bin +3 -0
- ablation_initialization/mla_normal_initialization/pytorch_model-00002-of-00003.bin +3 -0
- ablation_initialization/mla_normal_initialization/pytorch_model-00003-of-00003.bin +3 -0
- ablation_initialization/mla_normal_initialization/pytorch_model.bin.index.json +298 -0
- ablation_initialization/mqa_normal_initialization/config.json +18 -0
- ablation_initialization/mqa_normal_initialization/pytorch_model-00001-of-00003.bin +3 -0
- ablation_initialization/mqa_normal_initialization/pytorch_model-00002-of-00003.bin +3 -0
- ablation_initialization/mqa_normal_initialization/pytorch_model-00003-of-00003.bin +3 -0
- ablation_initialization/mqa_normal_initialization/pytorch_model.bin.index.json +226 -0
- fineweb-edu100B/fineweb_train_000001.bin +3 -0
- fineweb-edu100B/fineweb_train_000002.bin +3 -0
- fineweb-edu100B/fineweb_train_000003.bin +3 -0
- fineweb-edu100B/fineweb_train_000004.bin +3 -0
- fineweb-edu100B/fineweb_train_000005.bin +3 -0
- fineweb-edu100B/fineweb_train_000006.bin +3 -0
- fineweb-edu100B/fineweb_train_000007.bin +3 -0
- fineweb-edu100B/fineweb_train_000008.bin +3 -0
- fineweb-edu100B/fineweb_train_000009.bin +3 -0
- fineweb-edu100B/fineweb_train_000010.bin +3 -0
- fineweb-edu100B/fineweb_train_000011.bin +3 -0
- fineweb-edu100B/fineweb_train_000012.bin +3 -0
- fineweb-edu100B/fineweb_train_000013.bin +3 -0
- fineweb-edu100B/fineweb_train_000014.bin +3 -0
- fineweb-edu100B/fineweb_train_000015.bin +3 -0
ablation_initialization/gla_2_normal_initialization/config.json
ADDED
|
@@ -0,0 +1,20 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"GPT"
|
| 4 |
+
],
|
| 5 |
+
"bias": false,
|
| 6 |
+
"block_size": 2048,
|
| 7 |
+
"dropout": 0.0,
|
| 8 |
+
"intermediate_size": 10048,
|
| 9 |
+
"kv_lora_rank": 512,
|
| 10 |
+
"model_type": "gpt2",
|
| 11 |
+
"n_embd": 3072,
|
| 12 |
+
"n_head": 24,
|
| 13 |
+
"n_layer": 24,
|
| 14 |
+
"q_lora_rank": 1024,
|
| 15 |
+
"qk_nope_head_dim": 128,
|
| 16 |
+
"qk_rope_head_dim": 64,
|
| 17 |
+
"torch_dtype": "float32",
|
| 18 |
+
"transformers_version": "4.51.3",
|
| 19 |
+
"vocab_size": 50304
|
| 20 |
+
}
|
ablation_initialization/gla_2_normal_initialization/pytorch_model-00001-of-00003.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b8a93335e8272f9ee33b256069debb2dadecb1e7e0b97f62bfbf46e0f37dc0db
|
| 3 |
+
size 4901375657
|
ablation_initialization/gla_2_normal_initialization/pytorch_model-00002-of-00003.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:25698f343b2d75b8968bf728fa48e9821835b638ab7bbd8958b28ce75f9686cc
|
| 3 |
+
size 4983225876
|
ablation_initialization/gla_2_normal_initialization/pytorch_model-00003-of-00003.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7654eee3799fd08013053d2c2113434d9030c5e3360b9f3c73d8f0232cfe5c58
|
| 3 |
+
size 1606039992
|
ablation_initialization/gla_2_normal_initialization/pytorch_model.bin.index.json
ADDED
|
@@ -0,0 +1,346 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"metadata": {
|
| 3 |
+
"total_size": 11490521088
|
| 4 |
+
},
|
| 5 |
+
"weight_map": {
|
| 6 |
+
"layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 7 |
+
"lm_head.weight": "pytorch_model-00001-of-00003.bin",
|
| 8 |
+
"transformer.h.0.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 9 |
+
"transformer.h.0.attn.kv_a_layernorm_1.weight": "pytorch_model-00001-of-00003.bin",
|
| 10 |
+
"transformer.h.0.attn.kv_a_layernorm_2.weight": "pytorch_model-00001-of-00003.bin",
|
| 11 |
+
"transformer.h.0.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00001-of-00003.bin",
|
| 12 |
+
"transformer.h.0.attn.kv_b_proj_1.weight": "pytorch_model-00001-of-00003.bin",
|
| 13 |
+
"transformer.h.0.attn.kv_b_proj_2.weight": "pytorch_model-00001-of-00003.bin",
|
| 14 |
+
"transformer.h.0.attn.q_a_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 15 |
+
"transformer.h.0.attn.q_a_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 16 |
+
"transformer.h.0.attn.q_b_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 17 |
+
"transformer.h.0.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 18 |
+
"transformer.h.0.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 19 |
+
"transformer.h.0.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 20 |
+
"transformer.h.0.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 21 |
+
"transformer.h.0.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 22 |
+
"transformer.h.1.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 23 |
+
"transformer.h.1.attn.kv_a_layernorm_1.weight": "pytorch_model-00001-of-00003.bin",
|
| 24 |
+
"transformer.h.1.attn.kv_a_layernorm_2.weight": "pytorch_model-00001-of-00003.bin",
|
| 25 |
+
"transformer.h.1.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00001-of-00003.bin",
|
| 26 |
+
"transformer.h.1.attn.kv_b_proj_1.weight": "pytorch_model-00001-of-00003.bin",
|
| 27 |
+
"transformer.h.1.attn.kv_b_proj_2.weight": "pytorch_model-00001-of-00003.bin",
|
| 28 |
+
"transformer.h.1.attn.q_a_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 29 |
+
"transformer.h.1.attn.q_a_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 30 |
+
"transformer.h.1.attn.q_b_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 31 |
+
"transformer.h.1.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 32 |
+
"transformer.h.1.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 33 |
+
"transformer.h.1.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 34 |
+
"transformer.h.1.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 35 |
+
"transformer.h.1.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 36 |
+
"transformer.h.10.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 37 |
+
"transformer.h.10.attn.kv_a_layernorm_1.weight": "pytorch_model-00002-of-00003.bin",
|
| 38 |
+
"transformer.h.10.attn.kv_a_layernorm_2.weight": "pytorch_model-00002-of-00003.bin",
|
| 39 |
+
"transformer.h.10.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00002-of-00003.bin",
|
| 40 |
+
"transformer.h.10.attn.kv_b_proj_1.weight": "pytorch_model-00002-of-00003.bin",
|
| 41 |
+
"transformer.h.10.attn.kv_b_proj_2.weight": "pytorch_model-00002-of-00003.bin",
|
| 42 |
+
"transformer.h.10.attn.q_a_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 43 |
+
"transformer.h.10.attn.q_a_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 44 |
+
"transformer.h.10.attn.q_b_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 45 |
+
"transformer.h.10.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 46 |
+
"transformer.h.10.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 47 |
+
"transformer.h.10.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 48 |
+
"transformer.h.10.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 49 |
+
"transformer.h.10.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 50 |
+
"transformer.h.11.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 51 |
+
"transformer.h.11.attn.kv_a_layernorm_1.weight": "pytorch_model-00002-of-00003.bin",
|
| 52 |
+
"transformer.h.11.attn.kv_a_layernorm_2.weight": "pytorch_model-00002-of-00003.bin",
|
| 53 |
+
"transformer.h.11.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00002-of-00003.bin",
|
| 54 |
+
"transformer.h.11.attn.kv_b_proj_1.weight": "pytorch_model-00002-of-00003.bin",
|
| 55 |
+
"transformer.h.11.attn.kv_b_proj_2.weight": "pytorch_model-00002-of-00003.bin",
|
| 56 |
+
"transformer.h.11.attn.q_a_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 57 |
+
"transformer.h.11.attn.q_a_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 58 |
+
"transformer.h.11.attn.q_b_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 59 |
+
"transformer.h.11.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 60 |
+
"transformer.h.11.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 61 |
+
"transformer.h.11.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 62 |
+
"transformer.h.11.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 63 |
+
"transformer.h.11.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 64 |
+
"transformer.h.12.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 65 |
+
"transformer.h.12.attn.kv_a_layernorm_1.weight": "pytorch_model-00002-of-00003.bin",
|
| 66 |
+
"transformer.h.12.attn.kv_a_layernorm_2.weight": "pytorch_model-00002-of-00003.bin",
|
| 67 |
+
"transformer.h.12.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00002-of-00003.bin",
|
| 68 |
+
"transformer.h.12.attn.kv_b_proj_1.weight": "pytorch_model-00002-of-00003.bin",
|
| 69 |
+
"transformer.h.12.attn.kv_b_proj_2.weight": "pytorch_model-00002-of-00003.bin",
|
| 70 |
+
"transformer.h.12.attn.q_a_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 71 |
+
"transformer.h.12.attn.q_a_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 72 |
+
"transformer.h.12.attn.q_b_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 73 |
+
"transformer.h.12.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 74 |
+
"transformer.h.12.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 75 |
+
"transformer.h.12.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 76 |
+
"transformer.h.12.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 77 |
+
"transformer.h.12.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 78 |
+
"transformer.h.13.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 79 |
+
"transformer.h.13.attn.kv_a_layernorm_1.weight": "pytorch_model-00002-of-00003.bin",
|
| 80 |
+
"transformer.h.13.attn.kv_a_layernorm_2.weight": "pytorch_model-00002-of-00003.bin",
|
| 81 |
+
"transformer.h.13.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00002-of-00003.bin",
|
| 82 |
+
"transformer.h.13.attn.kv_b_proj_1.weight": "pytorch_model-00002-of-00003.bin",
|
| 83 |
+
"transformer.h.13.attn.kv_b_proj_2.weight": "pytorch_model-00002-of-00003.bin",
|
| 84 |
+
"transformer.h.13.attn.q_a_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 85 |
+
"transformer.h.13.attn.q_a_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 86 |
+
"transformer.h.13.attn.q_b_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 87 |
+
"transformer.h.13.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 88 |
+
"transformer.h.13.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 89 |
+
"transformer.h.13.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 90 |
+
"transformer.h.13.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 91 |
+
"transformer.h.13.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 92 |
+
"transformer.h.14.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 93 |
+
"transformer.h.14.attn.kv_a_layernorm_1.weight": "pytorch_model-00002-of-00003.bin",
|
| 94 |
+
"transformer.h.14.attn.kv_a_layernorm_2.weight": "pytorch_model-00002-of-00003.bin",
|
| 95 |
+
"transformer.h.14.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00002-of-00003.bin",
|
| 96 |
+
"transformer.h.14.attn.kv_b_proj_1.weight": "pytorch_model-00002-of-00003.bin",
|
| 97 |
+
"transformer.h.14.attn.kv_b_proj_2.weight": "pytorch_model-00002-of-00003.bin",
|
| 98 |
+
"transformer.h.14.attn.q_a_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 99 |
+
"transformer.h.14.attn.q_a_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 100 |
+
"transformer.h.14.attn.q_b_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 101 |
+
"transformer.h.14.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 102 |
+
"transformer.h.14.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 103 |
+
"transformer.h.14.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 104 |
+
"transformer.h.14.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 105 |
+
"transformer.h.14.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 106 |
+
"transformer.h.15.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 107 |
+
"transformer.h.15.attn.kv_a_layernorm_1.weight": "pytorch_model-00002-of-00003.bin",
|
| 108 |
+
"transformer.h.15.attn.kv_a_layernorm_2.weight": "pytorch_model-00002-of-00003.bin",
|
| 109 |
+
"transformer.h.15.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00002-of-00003.bin",
|
| 110 |
+
"transformer.h.15.attn.kv_b_proj_1.weight": "pytorch_model-00002-of-00003.bin",
|
| 111 |
+
"transformer.h.15.attn.kv_b_proj_2.weight": "pytorch_model-00002-of-00003.bin",
|
| 112 |
+
"transformer.h.15.attn.q_a_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 113 |
+
"transformer.h.15.attn.q_a_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 114 |
+
"transformer.h.15.attn.q_b_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 115 |
+
"transformer.h.15.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 116 |
+
"transformer.h.15.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 117 |
+
"transformer.h.15.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 118 |
+
"transformer.h.15.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 119 |
+
"transformer.h.15.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 120 |
+
"transformer.h.16.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 121 |
+
"transformer.h.16.attn.kv_a_layernorm_1.weight": "pytorch_model-00002-of-00003.bin",
|
| 122 |
+
"transformer.h.16.attn.kv_a_layernorm_2.weight": "pytorch_model-00002-of-00003.bin",
|
| 123 |
+
"transformer.h.16.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00002-of-00003.bin",
|
| 124 |
+
"transformer.h.16.attn.kv_b_proj_1.weight": "pytorch_model-00002-of-00003.bin",
|
| 125 |
+
"transformer.h.16.attn.kv_b_proj_2.weight": "pytorch_model-00002-of-00003.bin",
|
| 126 |
+
"transformer.h.16.attn.q_a_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 127 |
+
"transformer.h.16.attn.q_a_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 128 |
+
"transformer.h.16.attn.q_b_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 129 |
+
"transformer.h.16.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 130 |
+
"transformer.h.16.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 131 |
+
"transformer.h.16.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 132 |
+
"transformer.h.16.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 133 |
+
"transformer.h.16.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 134 |
+
"transformer.h.17.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 135 |
+
"transformer.h.17.attn.kv_a_layernorm_1.weight": "pytorch_model-00002-of-00003.bin",
|
| 136 |
+
"transformer.h.17.attn.kv_a_layernorm_2.weight": "pytorch_model-00002-of-00003.bin",
|
| 137 |
+
"transformer.h.17.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00002-of-00003.bin",
|
| 138 |
+
"transformer.h.17.attn.kv_b_proj_1.weight": "pytorch_model-00002-of-00003.bin",
|
| 139 |
+
"transformer.h.17.attn.kv_b_proj_2.weight": "pytorch_model-00002-of-00003.bin",
|
| 140 |
+
"transformer.h.17.attn.q_a_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 141 |
+
"transformer.h.17.attn.q_a_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 142 |
+
"transformer.h.17.attn.q_b_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 143 |
+
"transformer.h.17.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 144 |
+
"transformer.h.17.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 145 |
+
"transformer.h.17.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 146 |
+
"transformer.h.17.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 147 |
+
"transformer.h.17.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 148 |
+
"transformer.h.18.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 149 |
+
"transformer.h.18.attn.kv_a_layernorm_1.weight": "pytorch_model-00002-of-00003.bin",
|
| 150 |
+
"transformer.h.18.attn.kv_a_layernorm_2.weight": "pytorch_model-00002-of-00003.bin",
|
| 151 |
+
"transformer.h.18.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00002-of-00003.bin",
|
| 152 |
+
"transformer.h.18.attn.kv_b_proj_1.weight": "pytorch_model-00002-of-00003.bin",
|
| 153 |
+
"transformer.h.18.attn.kv_b_proj_2.weight": "pytorch_model-00002-of-00003.bin",
|
| 154 |
+
"transformer.h.18.attn.q_a_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 155 |
+
"transformer.h.18.attn.q_a_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 156 |
+
"transformer.h.18.attn.q_b_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 157 |
+
"transformer.h.18.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 158 |
+
"transformer.h.18.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 159 |
+
"transformer.h.18.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 160 |
+
"transformer.h.18.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 161 |
+
"transformer.h.18.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 162 |
+
"transformer.h.19.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 163 |
+
"transformer.h.19.attn.kv_a_layernorm_1.weight": "pytorch_model-00002-of-00003.bin",
|
| 164 |
+
"transformer.h.19.attn.kv_a_layernorm_2.weight": "pytorch_model-00002-of-00003.bin",
|
| 165 |
+
"transformer.h.19.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00002-of-00003.bin",
|
| 166 |
+
"transformer.h.19.attn.kv_b_proj_1.weight": "pytorch_model-00002-of-00003.bin",
|
| 167 |
+
"transformer.h.19.attn.kv_b_proj_2.weight": "pytorch_model-00002-of-00003.bin",
|
| 168 |
+
"transformer.h.19.attn.q_a_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 169 |
+
"transformer.h.19.attn.q_a_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 170 |
+
"transformer.h.19.attn.q_b_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 171 |
+
"transformer.h.19.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 172 |
+
"transformer.h.19.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 173 |
+
"transformer.h.19.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 174 |
+
"transformer.h.19.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 175 |
+
"transformer.h.19.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 176 |
+
"transformer.h.2.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 177 |
+
"transformer.h.2.attn.kv_a_layernorm_1.weight": "pytorch_model-00001-of-00003.bin",
|
| 178 |
+
"transformer.h.2.attn.kv_a_layernorm_2.weight": "pytorch_model-00001-of-00003.bin",
|
| 179 |
+
"transformer.h.2.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00001-of-00003.bin",
|
| 180 |
+
"transformer.h.2.attn.kv_b_proj_1.weight": "pytorch_model-00001-of-00003.bin",
|
| 181 |
+
"transformer.h.2.attn.kv_b_proj_2.weight": "pytorch_model-00001-of-00003.bin",
|
| 182 |
+
"transformer.h.2.attn.q_a_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 183 |
+
"transformer.h.2.attn.q_a_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 184 |
+
"transformer.h.2.attn.q_b_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 185 |
+
"transformer.h.2.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 186 |
+
"transformer.h.2.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 187 |
+
"transformer.h.2.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 188 |
+
"transformer.h.2.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 189 |
+
"transformer.h.2.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 190 |
+
"transformer.h.20.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 191 |
+
"transformer.h.20.attn.kv_a_layernorm_1.weight": "pytorch_model-00002-of-00003.bin",
|
| 192 |
+
"transformer.h.20.attn.kv_a_layernorm_2.weight": "pytorch_model-00002-of-00003.bin",
|
| 193 |
+
"transformer.h.20.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00002-of-00003.bin",
|
| 194 |
+
"transformer.h.20.attn.kv_b_proj_1.weight": "pytorch_model-00002-of-00003.bin",
|
| 195 |
+
"transformer.h.20.attn.kv_b_proj_2.weight": "pytorch_model-00002-of-00003.bin",
|
| 196 |
+
"transformer.h.20.attn.q_a_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 197 |
+
"transformer.h.20.attn.q_a_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 198 |
+
"transformer.h.20.attn.q_b_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 199 |
+
"transformer.h.20.input_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 200 |
+
"transformer.h.20.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 201 |
+
"transformer.h.20.mlp.c_fc2.weight": "pytorch_model-00003-of-00003.bin",
|
| 202 |
+
"transformer.h.20.mlp.c_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 203 |
+
"transformer.h.20.post_attention_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 204 |
+
"transformer.h.21.attn.c_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 205 |
+
"transformer.h.21.attn.kv_a_layernorm_1.weight": "pytorch_model-00003-of-00003.bin",
|
| 206 |
+
"transformer.h.21.attn.kv_a_layernorm_2.weight": "pytorch_model-00003-of-00003.bin",
|
| 207 |
+
"transformer.h.21.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00003-of-00003.bin",
|
| 208 |
+
"transformer.h.21.attn.kv_b_proj_1.weight": "pytorch_model-00003-of-00003.bin",
|
| 209 |
+
"transformer.h.21.attn.kv_b_proj_2.weight": "pytorch_model-00003-of-00003.bin",
|
| 210 |
+
"transformer.h.21.attn.q_a_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 211 |
+
"transformer.h.21.attn.q_a_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 212 |
+
"transformer.h.21.attn.q_b_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 213 |
+
"transformer.h.21.input_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 214 |
+
"transformer.h.21.mlp.c_fc1.weight": "pytorch_model-00003-of-00003.bin",
|
| 215 |
+
"transformer.h.21.mlp.c_fc2.weight": "pytorch_model-00003-of-00003.bin",
|
| 216 |
+
"transformer.h.21.mlp.c_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 217 |
+
"transformer.h.21.post_attention_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 218 |
+
"transformer.h.22.attn.c_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 219 |
+
"transformer.h.22.attn.kv_a_layernorm_1.weight": "pytorch_model-00003-of-00003.bin",
|
| 220 |
+
"transformer.h.22.attn.kv_a_layernorm_2.weight": "pytorch_model-00003-of-00003.bin",
|
| 221 |
+
"transformer.h.22.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00003-of-00003.bin",
|
| 222 |
+
"transformer.h.22.attn.kv_b_proj_1.weight": "pytorch_model-00003-of-00003.bin",
|
| 223 |
+
"transformer.h.22.attn.kv_b_proj_2.weight": "pytorch_model-00003-of-00003.bin",
|
| 224 |
+
"transformer.h.22.attn.q_a_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 225 |
+
"transformer.h.22.attn.q_a_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 226 |
+
"transformer.h.22.attn.q_b_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 227 |
+
"transformer.h.22.input_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 228 |
+
"transformer.h.22.mlp.c_fc1.weight": "pytorch_model-00003-of-00003.bin",
|
| 229 |
+
"transformer.h.22.mlp.c_fc2.weight": "pytorch_model-00003-of-00003.bin",
|
| 230 |
+
"transformer.h.22.mlp.c_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 231 |
+
"transformer.h.22.post_attention_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 232 |
+
"transformer.h.23.attn.c_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 233 |
+
"transformer.h.23.attn.kv_a_layernorm_1.weight": "pytorch_model-00003-of-00003.bin",
|
| 234 |
+
"transformer.h.23.attn.kv_a_layernorm_2.weight": "pytorch_model-00003-of-00003.bin",
|
| 235 |
+
"transformer.h.23.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00003-of-00003.bin",
|
| 236 |
+
"transformer.h.23.attn.kv_b_proj_1.weight": "pytorch_model-00003-of-00003.bin",
|
| 237 |
+
"transformer.h.23.attn.kv_b_proj_2.weight": "pytorch_model-00003-of-00003.bin",
|
| 238 |
+
"transformer.h.23.attn.q_a_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 239 |
+
"transformer.h.23.attn.q_a_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 240 |
+
"transformer.h.23.attn.q_b_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 241 |
+
"transformer.h.23.input_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 242 |
+
"transformer.h.23.mlp.c_fc1.weight": "pytorch_model-00003-of-00003.bin",
|
| 243 |
+
"transformer.h.23.mlp.c_fc2.weight": "pytorch_model-00003-of-00003.bin",
|
| 244 |
+
"transformer.h.23.mlp.c_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 245 |
+
"transformer.h.23.post_attention_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 246 |
+
"transformer.h.3.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 247 |
+
"transformer.h.3.attn.kv_a_layernorm_1.weight": "pytorch_model-00001-of-00003.bin",
|
| 248 |
+
"transformer.h.3.attn.kv_a_layernorm_2.weight": "pytorch_model-00001-of-00003.bin",
|
| 249 |
+
"transformer.h.3.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00001-of-00003.bin",
|
| 250 |
+
"transformer.h.3.attn.kv_b_proj_1.weight": "pytorch_model-00001-of-00003.bin",
|
| 251 |
+
"transformer.h.3.attn.kv_b_proj_2.weight": "pytorch_model-00001-of-00003.bin",
|
| 252 |
+
"transformer.h.3.attn.q_a_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 253 |
+
"transformer.h.3.attn.q_a_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 254 |
+
"transformer.h.3.attn.q_b_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 255 |
+
"transformer.h.3.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 256 |
+
"transformer.h.3.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 257 |
+
"transformer.h.3.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 258 |
+
"transformer.h.3.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 259 |
+
"transformer.h.3.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 260 |
+
"transformer.h.4.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 261 |
+
"transformer.h.4.attn.kv_a_layernorm_1.weight": "pytorch_model-00001-of-00003.bin",
|
| 262 |
+
"transformer.h.4.attn.kv_a_layernorm_2.weight": "pytorch_model-00001-of-00003.bin",
|
| 263 |
+
"transformer.h.4.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00001-of-00003.bin",
|
| 264 |
+
"transformer.h.4.attn.kv_b_proj_1.weight": "pytorch_model-00001-of-00003.bin",
|
| 265 |
+
"transformer.h.4.attn.kv_b_proj_2.weight": "pytorch_model-00001-of-00003.bin",
|
| 266 |
+
"transformer.h.4.attn.q_a_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 267 |
+
"transformer.h.4.attn.q_a_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 268 |
+
"transformer.h.4.attn.q_b_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 269 |
+
"transformer.h.4.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 270 |
+
"transformer.h.4.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 271 |
+
"transformer.h.4.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 272 |
+
"transformer.h.4.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 273 |
+
"transformer.h.4.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 274 |
+
"transformer.h.5.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 275 |
+
"transformer.h.5.attn.kv_a_layernorm_1.weight": "pytorch_model-00001-of-00003.bin",
|
| 276 |
+
"transformer.h.5.attn.kv_a_layernorm_2.weight": "pytorch_model-00001-of-00003.bin",
|
| 277 |
+
"transformer.h.5.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00001-of-00003.bin",
|
| 278 |
+
"transformer.h.5.attn.kv_b_proj_1.weight": "pytorch_model-00001-of-00003.bin",
|
| 279 |
+
"transformer.h.5.attn.kv_b_proj_2.weight": "pytorch_model-00001-of-00003.bin",
|
| 280 |
+
"transformer.h.5.attn.q_a_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 281 |
+
"transformer.h.5.attn.q_a_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 282 |
+
"transformer.h.5.attn.q_b_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 283 |
+
"transformer.h.5.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 284 |
+
"transformer.h.5.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 285 |
+
"transformer.h.5.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 286 |
+
"transformer.h.5.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 287 |
+
"transformer.h.5.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 288 |
+
"transformer.h.6.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 289 |
+
"transformer.h.6.attn.kv_a_layernorm_1.weight": "pytorch_model-00001-of-00003.bin",
|
| 290 |
+
"transformer.h.6.attn.kv_a_layernorm_2.weight": "pytorch_model-00001-of-00003.bin",
|
| 291 |
+
"transformer.h.6.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00001-of-00003.bin",
|
| 292 |
+
"transformer.h.6.attn.kv_b_proj_1.weight": "pytorch_model-00001-of-00003.bin",
|
| 293 |
+
"transformer.h.6.attn.kv_b_proj_2.weight": "pytorch_model-00001-of-00003.bin",
|
| 294 |
+
"transformer.h.6.attn.q_a_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 295 |
+
"transformer.h.6.attn.q_a_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 296 |
+
"transformer.h.6.attn.q_b_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 297 |
+
"transformer.h.6.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 298 |
+
"transformer.h.6.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 299 |
+
"transformer.h.6.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 300 |
+
"transformer.h.6.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 301 |
+
"transformer.h.6.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 302 |
+
"transformer.h.7.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 303 |
+
"transformer.h.7.attn.kv_a_layernorm_1.weight": "pytorch_model-00001-of-00003.bin",
|
| 304 |
+
"transformer.h.7.attn.kv_a_layernorm_2.weight": "pytorch_model-00001-of-00003.bin",
|
| 305 |
+
"transformer.h.7.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00001-of-00003.bin",
|
| 306 |
+
"transformer.h.7.attn.kv_b_proj_1.weight": "pytorch_model-00001-of-00003.bin",
|
| 307 |
+
"transformer.h.7.attn.kv_b_proj_2.weight": "pytorch_model-00001-of-00003.bin",
|
| 308 |
+
"transformer.h.7.attn.q_a_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 309 |
+
"transformer.h.7.attn.q_a_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 310 |
+
"transformer.h.7.attn.q_b_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 311 |
+
"transformer.h.7.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 312 |
+
"transformer.h.7.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 313 |
+
"transformer.h.7.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 314 |
+
"transformer.h.7.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 315 |
+
"transformer.h.7.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 316 |
+
"transformer.h.8.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 317 |
+
"transformer.h.8.attn.kv_a_layernorm_1.weight": "pytorch_model-00001-of-00003.bin",
|
| 318 |
+
"transformer.h.8.attn.kv_a_layernorm_2.weight": "pytorch_model-00001-of-00003.bin",
|
| 319 |
+
"transformer.h.8.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00001-of-00003.bin",
|
| 320 |
+
"transformer.h.8.attn.kv_b_proj_1.weight": "pytorch_model-00001-of-00003.bin",
|
| 321 |
+
"transformer.h.8.attn.kv_b_proj_2.weight": "pytorch_model-00001-of-00003.bin",
|
| 322 |
+
"transformer.h.8.attn.q_a_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 323 |
+
"transformer.h.8.attn.q_a_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 324 |
+
"transformer.h.8.attn.q_b_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 325 |
+
"transformer.h.8.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 326 |
+
"transformer.h.8.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 327 |
+
"transformer.h.8.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 328 |
+
"transformer.h.8.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 329 |
+
"transformer.h.8.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 330 |
+
"transformer.h.9.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 331 |
+
"transformer.h.9.attn.kv_a_layernorm_1.weight": "pytorch_model-00001-of-00003.bin",
|
| 332 |
+
"transformer.h.9.attn.kv_a_layernorm_2.weight": "pytorch_model-00001-of-00003.bin",
|
| 333 |
+
"transformer.h.9.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00001-of-00003.bin",
|
| 334 |
+
"transformer.h.9.attn.kv_b_proj_1.weight": "pytorch_model-00001-of-00003.bin",
|
| 335 |
+
"transformer.h.9.attn.kv_b_proj_2.weight": "pytorch_model-00001-of-00003.bin",
|
| 336 |
+
"transformer.h.9.attn.q_a_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 337 |
+
"transformer.h.9.attn.q_a_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 338 |
+
"transformer.h.9.attn.q_b_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 339 |
+
"transformer.h.9.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 340 |
+
"transformer.h.9.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 341 |
+
"transformer.h.9.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 342 |
+
"transformer.h.9.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 343 |
+
"transformer.h.9.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 344 |
+
"transformer.wte.weight": "pytorch_model-00001-of-00003.bin"
|
| 345 |
+
}
|
| 346 |
+
}
|
ablation_initialization/gla_4_normal_initialization/config.json
ADDED
|
@@ -0,0 +1,20 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"GPT"
|
| 4 |
+
],
|
| 5 |
+
"bias": false,
|
| 6 |
+
"block_size": 2048,
|
| 7 |
+
"dropout": 0.0,
|
| 8 |
+
"intermediate_size": 10136,
|
| 9 |
+
"kv_lora_rank": 512,
|
| 10 |
+
"model_type": "gpt2",
|
| 11 |
+
"n_embd": 3072,
|
| 12 |
+
"n_head": 24,
|
| 13 |
+
"n_layer": 24,
|
| 14 |
+
"q_lora_rank": 1024,
|
| 15 |
+
"qk_nope_head_dim": 128,
|
| 16 |
+
"qk_rope_head_dim": 64,
|
| 17 |
+
"torch_dtype": "float32",
|
| 18 |
+
"transformers_version": "4.51.3",
|
| 19 |
+
"vocab_size": 50304
|
| 20 |
+
}
|
ablation_initialization/gla_4_normal_initialization/pytorch_model-00001-of-00003.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1ac87a26a45d8727191a4551a611c46888b4c46017452605ab27121fe51c698f
|
| 3 |
+
size 4900199583
|
ablation_initialization/gla_4_normal_initialization/pytorch_model-00002-of-00003.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:dde1286ef44552ef30e2222ef95c1e4f34439b494ad1e008b1a4af56e8d58f24
|
| 3 |
+
size 4984311133
|
ablation_initialization/gla_4_normal_initialization/pytorch_model-00003-of-00003.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ec83a1678b567a2361a2c91a3d0e013c15d1a9ea3aaf435796470a5abffdc5a9
|
| 3 |
+
size 1608498606
|
ablation_initialization/gla_4_normal_initialization/pytorch_model.bin.index.json
ADDED
|
@@ -0,0 +1,370 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"metadata": {
|
| 3 |
+
"total_size": 11492880384
|
| 4 |
+
},
|
| 5 |
+
"weight_map": {
|
| 6 |
+
"layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 7 |
+
"lm_head.weight": "pytorch_model-00001-of-00003.bin",
|
| 8 |
+
"transformer.h.0.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 9 |
+
"transformer.h.0.attn.kv_a_layernorms.0.weight": "pytorch_model-00001-of-00003.bin",
|
| 10 |
+
"transformer.h.0.attn.kv_a_layernorms.1.weight": "pytorch_model-00001-of-00003.bin",
|
| 11 |
+
"transformer.h.0.attn.kv_a_layernorms.2.weight": "pytorch_model-00001-of-00003.bin",
|
| 12 |
+
"transformer.h.0.attn.kv_a_layernorms.3.weight": "pytorch_model-00001-of-00003.bin",
|
| 13 |
+
"transformer.h.0.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00001-of-00003.bin",
|
| 14 |
+
"transformer.h.0.attn.kv_b_proj": "pytorch_model-00001-of-00003.bin",
|
| 15 |
+
"transformer.h.0.attn.q_a_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 16 |
+
"transformer.h.0.attn.q_a_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 17 |
+
"transformer.h.0.attn.q_b_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 18 |
+
"transformer.h.0.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 19 |
+
"transformer.h.0.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 20 |
+
"transformer.h.0.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 21 |
+
"transformer.h.0.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 22 |
+
"transformer.h.0.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 23 |
+
"transformer.h.1.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 24 |
+
"transformer.h.1.attn.kv_a_layernorms.0.weight": "pytorch_model-00001-of-00003.bin",
|
| 25 |
+
"transformer.h.1.attn.kv_a_layernorms.1.weight": "pytorch_model-00001-of-00003.bin",
|
| 26 |
+
"transformer.h.1.attn.kv_a_layernorms.2.weight": "pytorch_model-00001-of-00003.bin",
|
| 27 |
+
"transformer.h.1.attn.kv_a_layernorms.3.weight": "pytorch_model-00001-of-00003.bin",
|
| 28 |
+
"transformer.h.1.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00001-of-00003.bin",
|
| 29 |
+
"transformer.h.1.attn.kv_b_proj": "pytorch_model-00001-of-00003.bin",
|
| 30 |
+
"transformer.h.1.attn.q_a_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 31 |
+
"transformer.h.1.attn.q_a_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 32 |
+
"transformer.h.1.attn.q_b_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 33 |
+
"transformer.h.1.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 34 |
+
"transformer.h.1.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 35 |
+
"transformer.h.1.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 36 |
+
"transformer.h.1.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 37 |
+
"transformer.h.1.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 38 |
+
"transformer.h.10.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 39 |
+
"transformer.h.10.attn.kv_a_layernorms.0.weight": "pytorch_model-00002-of-00003.bin",
|
| 40 |
+
"transformer.h.10.attn.kv_a_layernorms.1.weight": "pytorch_model-00002-of-00003.bin",
|
| 41 |
+
"transformer.h.10.attn.kv_a_layernorms.2.weight": "pytorch_model-00002-of-00003.bin",
|
| 42 |
+
"transformer.h.10.attn.kv_a_layernorms.3.weight": "pytorch_model-00002-of-00003.bin",
|
| 43 |
+
"transformer.h.10.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00002-of-00003.bin",
|
| 44 |
+
"transformer.h.10.attn.kv_b_proj": "pytorch_model-00002-of-00003.bin",
|
| 45 |
+
"transformer.h.10.attn.q_a_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 46 |
+
"transformer.h.10.attn.q_a_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 47 |
+
"transformer.h.10.attn.q_b_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 48 |
+
"transformer.h.10.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 49 |
+
"transformer.h.10.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 50 |
+
"transformer.h.10.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 51 |
+
"transformer.h.10.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 52 |
+
"transformer.h.10.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 53 |
+
"transformer.h.11.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 54 |
+
"transformer.h.11.attn.kv_a_layernorms.0.weight": "pytorch_model-00002-of-00003.bin",
|
| 55 |
+
"transformer.h.11.attn.kv_a_layernorms.1.weight": "pytorch_model-00002-of-00003.bin",
|
| 56 |
+
"transformer.h.11.attn.kv_a_layernorms.2.weight": "pytorch_model-00002-of-00003.bin",
|
| 57 |
+
"transformer.h.11.attn.kv_a_layernorms.3.weight": "pytorch_model-00002-of-00003.bin",
|
| 58 |
+
"transformer.h.11.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00002-of-00003.bin",
|
| 59 |
+
"transformer.h.11.attn.kv_b_proj": "pytorch_model-00002-of-00003.bin",
|
| 60 |
+
"transformer.h.11.attn.q_a_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 61 |
+
"transformer.h.11.attn.q_a_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 62 |
+
"transformer.h.11.attn.q_b_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 63 |
+
"transformer.h.11.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 64 |
+
"transformer.h.11.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 65 |
+
"transformer.h.11.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 66 |
+
"transformer.h.11.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 67 |
+
"transformer.h.11.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 68 |
+
"transformer.h.12.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 69 |
+
"transformer.h.12.attn.kv_a_layernorms.0.weight": "pytorch_model-00002-of-00003.bin",
|
| 70 |
+
"transformer.h.12.attn.kv_a_layernorms.1.weight": "pytorch_model-00002-of-00003.bin",
|
| 71 |
+
"transformer.h.12.attn.kv_a_layernorms.2.weight": "pytorch_model-00002-of-00003.bin",
|
| 72 |
+
"transformer.h.12.attn.kv_a_layernorms.3.weight": "pytorch_model-00002-of-00003.bin",
|
| 73 |
+
"transformer.h.12.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00002-of-00003.bin",
|
| 74 |
+
"transformer.h.12.attn.kv_b_proj": "pytorch_model-00002-of-00003.bin",
|
| 75 |
+
"transformer.h.12.attn.q_a_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 76 |
+
"transformer.h.12.attn.q_a_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 77 |
+
"transformer.h.12.attn.q_b_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 78 |
+
"transformer.h.12.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 79 |
+
"transformer.h.12.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 80 |
+
"transformer.h.12.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 81 |
+
"transformer.h.12.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 82 |
+
"transformer.h.12.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 83 |
+
"transformer.h.13.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 84 |
+
"transformer.h.13.attn.kv_a_layernorms.0.weight": "pytorch_model-00002-of-00003.bin",
|
| 85 |
+
"transformer.h.13.attn.kv_a_layernorms.1.weight": "pytorch_model-00002-of-00003.bin",
|
| 86 |
+
"transformer.h.13.attn.kv_a_layernorms.2.weight": "pytorch_model-00002-of-00003.bin",
|
| 87 |
+
"transformer.h.13.attn.kv_a_layernorms.3.weight": "pytorch_model-00002-of-00003.bin",
|
| 88 |
+
"transformer.h.13.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00002-of-00003.bin",
|
| 89 |
+
"transformer.h.13.attn.kv_b_proj": "pytorch_model-00002-of-00003.bin",
|
| 90 |
+
"transformer.h.13.attn.q_a_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 91 |
+
"transformer.h.13.attn.q_a_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 92 |
+
"transformer.h.13.attn.q_b_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 93 |
+
"transformer.h.13.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 94 |
+
"transformer.h.13.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 95 |
+
"transformer.h.13.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 96 |
+
"transformer.h.13.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 97 |
+
"transformer.h.13.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 98 |
+
"transformer.h.14.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 99 |
+
"transformer.h.14.attn.kv_a_layernorms.0.weight": "pytorch_model-00002-of-00003.bin",
|
| 100 |
+
"transformer.h.14.attn.kv_a_layernorms.1.weight": "pytorch_model-00002-of-00003.bin",
|
| 101 |
+
"transformer.h.14.attn.kv_a_layernorms.2.weight": "pytorch_model-00002-of-00003.bin",
|
| 102 |
+
"transformer.h.14.attn.kv_a_layernorms.3.weight": "pytorch_model-00002-of-00003.bin",
|
| 103 |
+
"transformer.h.14.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00002-of-00003.bin",
|
| 104 |
+
"transformer.h.14.attn.kv_b_proj": "pytorch_model-00002-of-00003.bin",
|
| 105 |
+
"transformer.h.14.attn.q_a_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 106 |
+
"transformer.h.14.attn.q_a_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 107 |
+
"transformer.h.14.attn.q_b_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 108 |
+
"transformer.h.14.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 109 |
+
"transformer.h.14.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 110 |
+
"transformer.h.14.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 111 |
+
"transformer.h.14.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 112 |
+
"transformer.h.14.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 113 |
+
"transformer.h.15.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 114 |
+
"transformer.h.15.attn.kv_a_layernorms.0.weight": "pytorch_model-00002-of-00003.bin",
|
| 115 |
+
"transformer.h.15.attn.kv_a_layernorms.1.weight": "pytorch_model-00002-of-00003.bin",
|
| 116 |
+
"transformer.h.15.attn.kv_a_layernorms.2.weight": "pytorch_model-00002-of-00003.bin",
|
| 117 |
+
"transformer.h.15.attn.kv_a_layernorms.3.weight": "pytorch_model-00002-of-00003.bin",
|
| 118 |
+
"transformer.h.15.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00002-of-00003.bin",
|
| 119 |
+
"transformer.h.15.attn.kv_b_proj": "pytorch_model-00002-of-00003.bin",
|
| 120 |
+
"transformer.h.15.attn.q_a_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 121 |
+
"transformer.h.15.attn.q_a_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 122 |
+
"transformer.h.15.attn.q_b_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 123 |
+
"transformer.h.15.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 124 |
+
"transformer.h.15.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 125 |
+
"transformer.h.15.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 126 |
+
"transformer.h.15.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 127 |
+
"transformer.h.15.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 128 |
+
"transformer.h.16.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 129 |
+
"transformer.h.16.attn.kv_a_layernorms.0.weight": "pytorch_model-00002-of-00003.bin",
|
| 130 |
+
"transformer.h.16.attn.kv_a_layernorms.1.weight": "pytorch_model-00002-of-00003.bin",
|
| 131 |
+
"transformer.h.16.attn.kv_a_layernorms.2.weight": "pytorch_model-00002-of-00003.bin",
|
| 132 |
+
"transformer.h.16.attn.kv_a_layernorms.3.weight": "pytorch_model-00002-of-00003.bin",
|
| 133 |
+
"transformer.h.16.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00002-of-00003.bin",
|
| 134 |
+
"transformer.h.16.attn.kv_b_proj": "pytorch_model-00002-of-00003.bin",
|
| 135 |
+
"transformer.h.16.attn.q_a_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 136 |
+
"transformer.h.16.attn.q_a_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 137 |
+
"transformer.h.16.attn.q_b_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 138 |
+
"transformer.h.16.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 139 |
+
"transformer.h.16.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 140 |
+
"transformer.h.16.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 141 |
+
"transformer.h.16.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 142 |
+
"transformer.h.16.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 143 |
+
"transformer.h.17.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 144 |
+
"transformer.h.17.attn.kv_a_layernorms.0.weight": "pytorch_model-00002-of-00003.bin",
|
| 145 |
+
"transformer.h.17.attn.kv_a_layernorms.1.weight": "pytorch_model-00002-of-00003.bin",
|
| 146 |
+
"transformer.h.17.attn.kv_a_layernorms.2.weight": "pytorch_model-00002-of-00003.bin",
|
| 147 |
+
"transformer.h.17.attn.kv_a_layernorms.3.weight": "pytorch_model-00002-of-00003.bin",
|
| 148 |
+
"transformer.h.17.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00002-of-00003.bin",
|
| 149 |
+
"transformer.h.17.attn.kv_b_proj": "pytorch_model-00002-of-00003.bin",
|
| 150 |
+
"transformer.h.17.attn.q_a_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 151 |
+
"transformer.h.17.attn.q_a_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 152 |
+
"transformer.h.17.attn.q_b_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 153 |
+
"transformer.h.17.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 154 |
+
"transformer.h.17.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 155 |
+
"transformer.h.17.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 156 |
+
"transformer.h.17.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 157 |
+
"transformer.h.17.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 158 |
+
"transformer.h.18.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 159 |
+
"transformer.h.18.attn.kv_a_layernorms.0.weight": "pytorch_model-00002-of-00003.bin",
|
| 160 |
+
"transformer.h.18.attn.kv_a_layernorms.1.weight": "pytorch_model-00002-of-00003.bin",
|
| 161 |
+
"transformer.h.18.attn.kv_a_layernorms.2.weight": "pytorch_model-00002-of-00003.bin",
|
| 162 |
+
"transformer.h.18.attn.kv_a_layernorms.3.weight": "pytorch_model-00002-of-00003.bin",
|
| 163 |
+
"transformer.h.18.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00002-of-00003.bin",
|
| 164 |
+
"transformer.h.18.attn.kv_b_proj": "pytorch_model-00002-of-00003.bin",
|
| 165 |
+
"transformer.h.18.attn.q_a_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 166 |
+
"transformer.h.18.attn.q_a_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 167 |
+
"transformer.h.18.attn.q_b_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 168 |
+
"transformer.h.18.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 169 |
+
"transformer.h.18.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 170 |
+
"transformer.h.18.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 171 |
+
"transformer.h.18.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 172 |
+
"transformer.h.18.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 173 |
+
"transformer.h.19.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 174 |
+
"transformer.h.19.attn.kv_a_layernorms.0.weight": "pytorch_model-00002-of-00003.bin",
|
| 175 |
+
"transformer.h.19.attn.kv_a_layernorms.1.weight": "pytorch_model-00002-of-00003.bin",
|
| 176 |
+
"transformer.h.19.attn.kv_a_layernorms.2.weight": "pytorch_model-00002-of-00003.bin",
|
| 177 |
+
"transformer.h.19.attn.kv_a_layernorms.3.weight": "pytorch_model-00002-of-00003.bin",
|
| 178 |
+
"transformer.h.19.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00002-of-00003.bin",
|
| 179 |
+
"transformer.h.19.attn.kv_b_proj": "pytorch_model-00002-of-00003.bin",
|
| 180 |
+
"transformer.h.19.attn.q_a_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 181 |
+
"transformer.h.19.attn.q_a_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 182 |
+
"transformer.h.19.attn.q_b_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 183 |
+
"transformer.h.19.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 184 |
+
"transformer.h.19.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 185 |
+
"transformer.h.19.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 186 |
+
"transformer.h.19.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 187 |
+
"transformer.h.19.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 188 |
+
"transformer.h.2.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 189 |
+
"transformer.h.2.attn.kv_a_layernorms.0.weight": "pytorch_model-00001-of-00003.bin",
|
| 190 |
+
"transformer.h.2.attn.kv_a_layernorms.1.weight": "pytorch_model-00001-of-00003.bin",
|
| 191 |
+
"transformer.h.2.attn.kv_a_layernorms.2.weight": "pytorch_model-00001-of-00003.bin",
|
| 192 |
+
"transformer.h.2.attn.kv_a_layernorms.3.weight": "pytorch_model-00001-of-00003.bin",
|
| 193 |
+
"transformer.h.2.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00001-of-00003.bin",
|
| 194 |
+
"transformer.h.2.attn.kv_b_proj": "pytorch_model-00001-of-00003.bin",
|
| 195 |
+
"transformer.h.2.attn.q_a_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 196 |
+
"transformer.h.2.attn.q_a_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 197 |
+
"transformer.h.2.attn.q_b_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 198 |
+
"transformer.h.2.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 199 |
+
"transformer.h.2.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 200 |
+
"transformer.h.2.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 201 |
+
"transformer.h.2.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 202 |
+
"transformer.h.2.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 203 |
+
"transformer.h.20.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 204 |
+
"transformer.h.20.attn.kv_a_layernorms.0.weight": "pytorch_model-00002-of-00003.bin",
|
| 205 |
+
"transformer.h.20.attn.kv_a_layernorms.1.weight": "pytorch_model-00002-of-00003.bin",
|
| 206 |
+
"transformer.h.20.attn.kv_a_layernorms.2.weight": "pytorch_model-00002-of-00003.bin",
|
| 207 |
+
"transformer.h.20.attn.kv_a_layernorms.3.weight": "pytorch_model-00002-of-00003.bin",
|
| 208 |
+
"transformer.h.20.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00002-of-00003.bin",
|
| 209 |
+
"transformer.h.20.attn.kv_b_proj": "pytorch_model-00002-of-00003.bin",
|
| 210 |
+
"transformer.h.20.attn.q_a_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 211 |
+
"transformer.h.20.attn.q_a_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 212 |
+
"transformer.h.20.attn.q_b_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 213 |
+
"transformer.h.20.input_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 214 |
+
"transformer.h.20.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 215 |
+
"transformer.h.20.mlp.c_fc2.weight": "pytorch_model-00003-of-00003.bin",
|
| 216 |
+
"transformer.h.20.mlp.c_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 217 |
+
"transformer.h.20.post_attention_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 218 |
+
"transformer.h.21.attn.c_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 219 |
+
"transformer.h.21.attn.kv_a_layernorms.0.weight": "pytorch_model-00003-of-00003.bin",
|
| 220 |
+
"transformer.h.21.attn.kv_a_layernorms.1.weight": "pytorch_model-00003-of-00003.bin",
|
| 221 |
+
"transformer.h.21.attn.kv_a_layernorms.2.weight": "pytorch_model-00003-of-00003.bin",
|
| 222 |
+
"transformer.h.21.attn.kv_a_layernorms.3.weight": "pytorch_model-00003-of-00003.bin",
|
| 223 |
+
"transformer.h.21.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00003-of-00003.bin",
|
| 224 |
+
"transformer.h.21.attn.kv_b_proj": "pytorch_model-00003-of-00003.bin",
|
| 225 |
+
"transformer.h.21.attn.q_a_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 226 |
+
"transformer.h.21.attn.q_a_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 227 |
+
"transformer.h.21.attn.q_b_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 228 |
+
"transformer.h.21.input_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 229 |
+
"transformer.h.21.mlp.c_fc1.weight": "pytorch_model-00003-of-00003.bin",
|
| 230 |
+
"transformer.h.21.mlp.c_fc2.weight": "pytorch_model-00003-of-00003.bin",
|
| 231 |
+
"transformer.h.21.mlp.c_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 232 |
+
"transformer.h.21.post_attention_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 233 |
+
"transformer.h.22.attn.c_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 234 |
+
"transformer.h.22.attn.kv_a_layernorms.0.weight": "pytorch_model-00003-of-00003.bin",
|
| 235 |
+
"transformer.h.22.attn.kv_a_layernorms.1.weight": "pytorch_model-00003-of-00003.bin",
|
| 236 |
+
"transformer.h.22.attn.kv_a_layernorms.2.weight": "pytorch_model-00003-of-00003.bin",
|
| 237 |
+
"transformer.h.22.attn.kv_a_layernorms.3.weight": "pytorch_model-00003-of-00003.bin",
|
| 238 |
+
"transformer.h.22.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00003-of-00003.bin",
|
| 239 |
+
"transformer.h.22.attn.kv_b_proj": "pytorch_model-00003-of-00003.bin",
|
| 240 |
+
"transformer.h.22.attn.q_a_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 241 |
+
"transformer.h.22.attn.q_a_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 242 |
+
"transformer.h.22.attn.q_b_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 243 |
+
"transformer.h.22.input_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 244 |
+
"transformer.h.22.mlp.c_fc1.weight": "pytorch_model-00003-of-00003.bin",
|
| 245 |
+
"transformer.h.22.mlp.c_fc2.weight": "pytorch_model-00003-of-00003.bin",
|
| 246 |
+
"transformer.h.22.mlp.c_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 247 |
+
"transformer.h.22.post_attention_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 248 |
+
"transformer.h.23.attn.c_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 249 |
+
"transformer.h.23.attn.kv_a_layernorms.0.weight": "pytorch_model-00003-of-00003.bin",
|
| 250 |
+
"transformer.h.23.attn.kv_a_layernorms.1.weight": "pytorch_model-00003-of-00003.bin",
|
| 251 |
+
"transformer.h.23.attn.kv_a_layernorms.2.weight": "pytorch_model-00003-of-00003.bin",
|
| 252 |
+
"transformer.h.23.attn.kv_a_layernorms.3.weight": "pytorch_model-00003-of-00003.bin",
|
| 253 |
+
"transformer.h.23.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00003-of-00003.bin",
|
| 254 |
+
"transformer.h.23.attn.kv_b_proj": "pytorch_model-00003-of-00003.bin",
|
| 255 |
+
"transformer.h.23.attn.q_a_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 256 |
+
"transformer.h.23.attn.q_a_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 257 |
+
"transformer.h.23.attn.q_b_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 258 |
+
"transformer.h.23.input_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 259 |
+
"transformer.h.23.mlp.c_fc1.weight": "pytorch_model-00003-of-00003.bin",
|
| 260 |
+
"transformer.h.23.mlp.c_fc2.weight": "pytorch_model-00003-of-00003.bin",
|
| 261 |
+
"transformer.h.23.mlp.c_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 262 |
+
"transformer.h.23.post_attention_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 263 |
+
"transformer.h.3.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 264 |
+
"transformer.h.3.attn.kv_a_layernorms.0.weight": "pytorch_model-00001-of-00003.bin",
|
| 265 |
+
"transformer.h.3.attn.kv_a_layernorms.1.weight": "pytorch_model-00001-of-00003.bin",
|
| 266 |
+
"transformer.h.3.attn.kv_a_layernorms.2.weight": "pytorch_model-00001-of-00003.bin",
|
| 267 |
+
"transformer.h.3.attn.kv_a_layernorms.3.weight": "pytorch_model-00001-of-00003.bin",
|
| 268 |
+
"transformer.h.3.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00001-of-00003.bin",
|
| 269 |
+
"transformer.h.3.attn.kv_b_proj": "pytorch_model-00001-of-00003.bin",
|
| 270 |
+
"transformer.h.3.attn.q_a_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 271 |
+
"transformer.h.3.attn.q_a_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 272 |
+
"transformer.h.3.attn.q_b_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 273 |
+
"transformer.h.3.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 274 |
+
"transformer.h.3.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 275 |
+
"transformer.h.3.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 276 |
+
"transformer.h.3.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 277 |
+
"transformer.h.3.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 278 |
+
"transformer.h.4.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 279 |
+
"transformer.h.4.attn.kv_a_layernorms.0.weight": "pytorch_model-00001-of-00003.bin",
|
| 280 |
+
"transformer.h.4.attn.kv_a_layernorms.1.weight": "pytorch_model-00001-of-00003.bin",
|
| 281 |
+
"transformer.h.4.attn.kv_a_layernorms.2.weight": "pytorch_model-00001-of-00003.bin",
|
| 282 |
+
"transformer.h.4.attn.kv_a_layernorms.3.weight": "pytorch_model-00001-of-00003.bin",
|
| 283 |
+
"transformer.h.4.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00001-of-00003.bin",
|
| 284 |
+
"transformer.h.4.attn.kv_b_proj": "pytorch_model-00001-of-00003.bin",
|
| 285 |
+
"transformer.h.4.attn.q_a_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 286 |
+
"transformer.h.4.attn.q_a_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 287 |
+
"transformer.h.4.attn.q_b_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 288 |
+
"transformer.h.4.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 289 |
+
"transformer.h.4.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 290 |
+
"transformer.h.4.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 291 |
+
"transformer.h.4.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 292 |
+
"transformer.h.4.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 293 |
+
"transformer.h.5.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 294 |
+
"transformer.h.5.attn.kv_a_layernorms.0.weight": "pytorch_model-00001-of-00003.bin",
|
| 295 |
+
"transformer.h.5.attn.kv_a_layernorms.1.weight": "pytorch_model-00001-of-00003.bin",
|
| 296 |
+
"transformer.h.5.attn.kv_a_layernorms.2.weight": "pytorch_model-00001-of-00003.bin",
|
| 297 |
+
"transformer.h.5.attn.kv_a_layernorms.3.weight": "pytorch_model-00001-of-00003.bin",
|
| 298 |
+
"transformer.h.5.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00001-of-00003.bin",
|
| 299 |
+
"transformer.h.5.attn.kv_b_proj": "pytorch_model-00001-of-00003.bin",
|
| 300 |
+
"transformer.h.5.attn.q_a_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 301 |
+
"transformer.h.5.attn.q_a_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 302 |
+
"transformer.h.5.attn.q_b_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 303 |
+
"transformer.h.5.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 304 |
+
"transformer.h.5.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 305 |
+
"transformer.h.5.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 306 |
+
"transformer.h.5.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 307 |
+
"transformer.h.5.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 308 |
+
"transformer.h.6.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 309 |
+
"transformer.h.6.attn.kv_a_layernorms.0.weight": "pytorch_model-00001-of-00003.bin",
|
| 310 |
+
"transformer.h.6.attn.kv_a_layernorms.1.weight": "pytorch_model-00001-of-00003.bin",
|
| 311 |
+
"transformer.h.6.attn.kv_a_layernorms.2.weight": "pytorch_model-00001-of-00003.bin",
|
| 312 |
+
"transformer.h.6.attn.kv_a_layernorms.3.weight": "pytorch_model-00001-of-00003.bin",
|
| 313 |
+
"transformer.h.6.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00001-of-00003.bin",
|
| 314 |
+
"transformer.h.6.attn.kv_b_proj": "pytorch_model-00001-of-00003.bin",
|
| 315 |
+
"transformer.h.6.attn.q_a_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 316 |
+
"transformer.h.6.attn.q_a_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 317 |
+
"transformer.h.6.attn.q_b_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 318 |
+
"transformer.h.6.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 319 |
+
"transformer.h.6.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 320 |
+
"transformer.h.6.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 321 |
+
"transformer.h.6.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 322 |
+
"transformer.h.6.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 323 |
+
"transformer.h.7.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 324 |
+
"transformer.h.7.attn.kv_a_layernorms.0.weight": "pytorch_model-00001-of-00003.bin",
|
| 325 |
+
"transformer.h.7.attn.kv_a_layernorms.1.weight": "pytorch_model-00001-of-00003.bin",
|
| 326 |
+
"transformer.h.7.attn.kv_a_layernorms.2.weight": "pytorch_model-00001-of-00003.bin",
|
| 327 |
+
"transformer.h.7.attn.kv_a_layernorms.3.weight": "pytorch_model-00001-of-00003.bin",
|
| 328 |
+
"transformer.h.7.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00001-of-00003.bin",
|
| 329 |
+
"transformer.h.7.attn.kv_b_proj": "pytorch_model-00001-of-00003.bin",
|
| 330 |
+
"transformer.h.7.attn.q_a_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 331 |
+
"transformer.h.7.attn.q_a_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 332 |
+
"transformer.h.7.attn.q_b_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 333 |
+
"transformer.h.7.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 334 |
+
"transformer.h.7.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 335 |
+
"transformer.h.7.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 336 |
+
"transformer.h.7.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 337 |
+
"transformer.h.7.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 338 |
+
"transformer.h.8.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 339 |
+
"transformer.h.8.attn.kv_a_layernorms.0.weight": "pytorch_model-00001-of-00003.bin",
|
| 340 |
+
"transformer.h.8.attn.kv_a_layernorms.1.weight": "pytorch_model-00001-of-00003.bin",
|
| 341 |
+
"transformer.h.8.attn.kv_a_layernorms.2.weight": "pytorch_model-00001-of-00003.bin",
|
| 342 |
+
"transformer.h.8.attn.kv_a_layernorms.3.weight": "pytorch_model-00001-of-00003.bin",
|
| 343 |
+
"transformer.h.8.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00001-of-00003.bin",
|
| 344 |
+
"transformer.h.8.attn.kv_b_proj": "pytorch_model-00001-of-00003.bin",
|
| 345 |
+
"transformer.h.8.attn.q_a_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 346 |
+
"transformer.h.8.attn.q_a_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 347 |
+
"transformer.h.8.attn.q_b_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 348 |
+
"transformer.h.8.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 349 |
+
"transformer.h.8.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 350 |
+
"transformer.h.8.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 351 |
+
"transformer.h.8.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 352 |
+
"transformer.h.8.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 353 |
+
"transformer.h.9.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 354 |
+
"transformer.h.9.attn.kv_a_layernorms.0.weight": "pytorch_model-00001-of-00003.bin",
|
| 355 |
+
"transformer.h.9.attn.kv_a_layernorms.1.weight": "pytorch_model-00001-of-00003.bin",
|
| 356 |
+
"transformer.h.9.attn.kv_a_layernorms.2.weight": "pytorch_model-00001-of-00003.bin",
|
| 357 |
+
"transformer.h.9.attn.kv_a_layernorms.3.weight": "pytorch_model-00001-of-00003.bin",
|
| 358 |
+
"transformer.h.9.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00001-of-00003.bin",
|
| 359 |
+
"transformer.h.9.attn.kv_b_proj": "pytorch_model-00001-of-00003.bin",
|
| 360 |
+
"transformer.h.9.attn.q_a_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 361 |
+
"transformer.h.9.attn.q_a_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 362 |
+
"transformer.h.9.attn.q_b_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 363 |
+
"transformer.h.9.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 364 |
+
"transformer.h.9.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 365 |
+
"transformer.h.9.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 366 |
+
"transformer.h.9.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 367 |
+
"transformer.h.9.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 368 |
+
"transformer.wte.weight": "pytorch_model-00001-of-00003.bin"
|
| 369 |
+
}
|
| 370 |
+
}
|
ablation_initialization/gqa_normal_initialization/config.json
ADDED
|
@@ -0,0 +1,18 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"GPT"
|
| 4 |
+
],
|
| 5 |
+
"bias": false,
|
| 6 |
+
"block_size": 2048,
|
| 7 |
+
"dropout": 0.0,
|
| 8 |
+
"group_size": 4,
|
| 9 |
+
"head_dim": 128,
|
| 10 |
+
"intermediate_size": 9728,
|
| 11 |
+
"model_type": "gpt2",
|
| 12 |
+
"n_embd": 3072,
|
| 13 |
+
"n_head": 24,
|
| 14 |
+
"n_layer": 24,
|
| 15 |
+
"torch_dtype": "float32",
|
| 16 |
+
"transformers_version": "4.51.3",
|
| 17 |
+
"vocab_size": 50304
|
| 18 |
+
}
|
ablation_initialization/gqa_normal_initialization/pytorch_model-00001-of-00003.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f6508dc234bb3143a79c31544117c06744186f810493bad5c10f78938ca3a6b7
|
| 3 |
+
size 4909160580
|
ablation_initialization/gqa_normal_initialization/pytorch_model-00002-of-00003.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a31f9cb877f1e74be4d349da17200d77b6d744b021bc864b499321c8be527d88
|
| 3 |
+
size 4983138488
|
ablation_initialization/gqa_normal_initialization/pytorch_model-00003-of-00003.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3a9d5f0ee9e6facd70cf2e60e10e3adbda1591973280594c2dac8f84586c98b8
|
| 3 |
+
size 1598151850
|
ablation_initialization/gqa_normal_initialization/pytorch_model.bin.index.json
ADDED
|
@@ -0,0 +1,226 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"metadata": {
|
| 3 |
+
"total_size": 11490373632
|
| 4 |
+
},
|
| 5 |
+
"weight_map": {
|
| 6 |
+
"layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 7 |
+
"lm_head.weight": "pytorch_model-00001-of-00003.bin",
|
| 8 |
+
"transformer.h.0.attn.c_k.weight": "pytorch_model-00001-of-00003.bin",
|
| 9 |
+
"transformer.h.0.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 10 |
+
"transformer.h.0.attn.c_q.weight": "pytorch_model-00001-of-00003.bin",
|
| 11 |
+
"transformer.h.0.attn.c_v.weight": "pytorch_model-00001-of-00003.bin",
|
| 12 |
+
"transformer.h.0.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 13 |
+
"transformer.h.0.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 14 |
+
"transformer.h.0.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 15 |
+
"transformer.h.0.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 16 |
+
"transformer.h.0.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 17 |
+
"transformer.h.1.attn.c_k.weight": "pytorch_model-00001-of-00003.bin",
|
| 18 |
+
"transformer.h.1.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 19 |
+
"transformer.h.1.attn.c_q.weight": "pytorch_model-00001-of-00003.bin",
|
| 20 |
+
"transformer.h.1.attn.c_v.weight": "pytorch_model-00001-of-00003.bin",
|
| 21 |
+
"transformer.h.1.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 22 |
+
"transformer.h.1.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 23 |
+
"transformer.h.1.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 24 |
+
"transformer.h.1.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 25 |
+
"transformer.h.1.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 26 |
+
"transformer.h.10.attn.c_k.weight": "pytorch_model-00002-of-00003.bin",
|
| 27 |
+
"transformer.h.10.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 28 |
+
"transformer.h.10.attn.c_q.weight": "pytorch_model-00002-of-00003.bin",
|
| 29 |
+
"transformer.h.10.attn.c_v.weight": "pytorch_model-00002-of-00003.bin",
|
| 30 |
+
"transformer.h.10.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 31 |
+
"transformer.h.10.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 32 |
+
"transformer.h.10.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 33 |
+
"transformer.h.10.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 34 |
+
"transformer.h.10.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 35 |
+
"transformer.h.11.attn.c_k.weight": "pytorch_model-00002-of-00003.bin",
|
| 36 |
+
"transformer.h.11.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 37 |
+
"transformer.h.11.attn.c_q.weight": "pytorch_model-00002-of-00003.bin",
|
| 38 |
+
"transformer.h.11.attn.c_v.weight": "pytorch_model-00002-of-00003.bin",
|
| 39 |
+
"transformer.h.11.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 40 |
+
"transformer.h.11.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 41 |
+
"transformer.h.11.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 42 |
+
"transformer.h.11.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 43 |
+
"transformer.h.11.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 44 |
+
"transformer.h.12.attn.c_k.weight": "pytorch_model-00002-of-00003.bin",
|
| 45 |
+
"transformer.h.12.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 46 |
+
"transformer.h.12.attn.c_q.weight": "pytorch_model-00002-of-00003.bin",
|
| 47 |
+
"transformer.h.12.attn.c_v.weight": "pytorch_model-00002-of-00003.bin",
|
| 48 |
+
"transformer.h.12.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 49 |
+
"transformer.h.12.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 50 |
+
"transformer.h.12.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 51 |
+
"transformer.h.12.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 52 |
+
"transformer.h.12.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 53 |
+
"transformer.h.13.attn.c_k.weight": "pytorch_model-00002-of-00003.bin",
|
| 54 |
+
"transformer.h.13.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 55 |
+
"transformer.h.13.attn.c_q.weight": "pytorch_model-00002-of-00003.bin",
|
| 56 |
+
"transformer.h.13.attn.c_v.weight": "pytorch_model-00002-of-00003.bin",
|
| 57 |
+
"transformer.h.13.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 58 |
+
"transformer.h.13.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 59 |
+
"transformer.h.13.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 60 |
+
"transformer.h.13.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 61 |
+
"transformer.h.13.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 62 |
+
"transformer.h.14.attn.c_k.weight": "pytorch_model-00002-of-00003.bin",
|
| 63 |
+
"transformer.h.14.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 64 |
+
"transformer.h.14.attn.c_q.weight": "pytorch_model-00002-of-00003.bin",
|
| 65 |
+
"transformer.h.14.attn.c_v.weight": "pytorch_model-00002-of-00003.bin",
|
| 66 |
+
"transformer.h.14.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 67 |
+
"transformer.h.14.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 68 |
+
"transformer.h.14.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 69 |
+
"transformer.h.14.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 70 |
+
"transformer.h.14.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 71 |
+
"transformer.h.15.attn.c_k.weight": "pytorch_model-00002-of-00003.bin",
|
| 72 |
+
"transformer.h.15.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 73 |
+
"transformer.h.15.attn.c_q.weight": "pytorch_model-00002-of-00003.bin",
|
| 74 |
+
"transformer.h.15.attn.c_v.weight": "pytorch_model-00002-of-00003.bin",
|
| 75 |
+
"transformer.h.15.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 76 |
+
"transformer.h.15.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 77 |
+
"transformer.h.15.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 78 |
+
"transformer.h.15.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 79 |
+
"transformer.h.15.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 80 |
+
"transformer.h.16.attn.c_k.weight": "pytorch_model-00002-of-00003.bin",
|
| 81 |
+
"transformer.h.16.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 82 |
+
"transformer.h.16.attn.c_q.weight": "pytorch_model-00002-of-00003.bin",
|
| 83 |
+
"transformer.h.16.attn.c_v.weight": "pytorch_model-00002-of-00003.bin",
|
| 84 |
+
"transformer.h.16.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 85 |
+
"transformer.h.16.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 86 |
+
"transformer.h.16.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 87 |
+
"transformer.h.16.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 88 |
+
"transformer.h.16.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 89 |
+
"transformer.h.17.attn.c_k.weight": "pytorch_model-00002-of-00003.bin",
|
| 90 |
+
"transformer.h.17.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 91 |
+
"transformer.h.17.attn.c_q.weight": "pytorch_model-00002-of-00003.bin",
|
| 92 |
+
"transformer.h.17.attn.c_v.weight": "pytorch_model-00002-of-00003.bin",
|
| 93 |
+
"transformer.h.17.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 94 |
+
"transformer.h.17.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 95 |
+
"transformer.h.17.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 96 |
+
"transformer.h.17.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 97 |
+
"transformer.h.17.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 98 |
+
"transformer.h.18.attn.c_k.weight": "pytorch_model-00002-of-00003.bin",
|
| 99 |
+
"transformer.h.18.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 100 |
+
"transformer.h.18.attn.c_q.weight": "pytorch_model-00002-of-00003.bin",
|
| 101 |
+
"transformer.h.18.attn.c_v.weight": "pytorch_model-00002-of-00003.bin",
|
| 102 |
+
"transformer.h.18.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 103 |
+
"transformer.h.18.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 104 |
+
"transformer.h.18.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 105 |
+
"transformer.h.18.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 106 |
+
"transformer.h.18.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 107 |
+
"transformer.h.19.attn.c_k.weight": "pytorch_model-00002-of-00003.bin",
|
| 108 |
+
"transformer.h.19.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 109 |
+
"transformer.h.19.attn.c_q.weight": "pytorch_model-00002-of-00003.bin",
|
| 110 |
+
"transformer.h.19.attn.c_v.weight": "pytorch_model-00002-of-00003.bin",
|
| 111 |
+
"transformer.h.19.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 112 |
+
"transformer.h.19.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 113 |
+
"transformer.h.19.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 114 |
+
"transformer.h.19.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 115 |
+
"transformer.h.19.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 116 |
+
"transformer.h.2.attn.c_k.weight": "pytorch_model-00001-of-00003.bin",
|
| 117 |
+
"transformer.h.2.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 118 |
+
"transformer.h.2.attn.c_q.weight": "pytorch_model-00001-of-00003.bin",
|
| 119 |
+
"transformer.h.2.attn.c_v.weight": "pytorch_model-00001-of-00003.bin",
|
| 120 |
+
"transformer.h.2.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 121 |
+
"transformer.h.2.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 122 |
+
"transformer.h.2.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 123 |
+
"transformer.h.2.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 124 |
+
"transformer.h.2.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 125 |
+
"transformer.h.20.attn.c_k.weight": "pytorch_model-00002-of-00003.bin",
|
| 126 |
+
"transformer.h.20.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 127 |
+
"transformer.h.20.attn.c_q.weight": "pytorch_model-00002-of-00003.bin",
|
| 128 |
+
"transformer.h.20.attn.c_v.weight": "pytorch_model-00002-of-00003.bin",
|
| 129 |
+
"transformer.h.20.input_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 130 |
+
"transformer.h.20.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 131 |
+
"transformer.h.20.mlp.c_fc2.weight": "pytorch_model-00003-of-00003.bin",
|
| 132 |
+
"transformer.h.20.mlp.c_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 133 |
+
"transformer.h.20.post_attention_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 134 |
+
"transformer.h.21.attn.c_k.weight": "pytorch_model-00003-of-00003.bin",
|
| 135 |
+
"transformer.h.21.attn.c_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 136 |
+
"transformer.h.21.attn.c_q.weight": "pytorch_model-00003-of-00003.bin",
|
| 137 |
+
"transformer.h.21.attn.c_v.weight": "pytorch_model-00003-of-00003.bin",
|
| 138 |
+
"transformer.h.21.input_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 139 |
+
"transformer.h.21.mlp.c_fc1.weight": "pytorch_model-00003-of-00003.bin",
|
| 140 |
+
"transformer.h.21.mlp.c_fc2.weight": "pytorch_model-00003-of-00003.bin",
|
| 141 |
+
"transformer.h.21.mlp.c_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 142 |
+
"transformer.h.21.post_attention_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 143 |
+
"transformer.h.22.attn.c_k.weight": "pytorch_model-00003-of-00003.bin",
|
| 144 |
+
"transformer.h.22.attn.c_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 145 |
+
"transformer.h.22.attn.c_q.weight": "pytorch_model-00003-of-00003.bin",
|
| 146 |
+
"transformer.h.22.attn.c_v.weight": "pytorch_model-00003-of-00003.bin",
|
| 147 |
+
"transformer.h.22.input_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 148 |
+
"transformer.h.22.mlp.c_fc1.weight": "pytorch_model-00003-of-00003.bin",
|
| 149 |
+
"transformer.h.22.mlp.c_fc2.weight": "pytorch_model-00003-of-00003.bin",
|
| 150 |
+
"transformer.h.22.mlp.c_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 151 |
+
"transformer.h.22.post_attention_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 152 |
+
"transformer.h.23.attn.c_k.weight": "pytorch_model-00003-of-00003.bin",
|
| 153 |
+
"transformer.h.23.attn.c_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 154 |
+
"transformer.h.23.attn.c_q.weight": "pytorch_model-00003-of-00003.bin",
|
| 155 |
+
"transformer.h.23.attn.c_v.weight": "pytorch_model-00003-of-00003.bin",
|
| 156 |
+
"transformer.h.23.input_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 157 |
+
"transformer.h.23.mlp.c_fc1.weight": "pytorch_model-00003-of-00003.bin",
|
| 158 |
+
"transformer.h.23.mlp.c_fc2.weight": "pytorch_model-00003-of-00003.bin",
|
| 159 |
+
"transformer.h.23.mlp.c_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 160 |
+
"transformer.h.23.post_attention_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 161 |
+
"transformer.h.3.attn.c_k.weight": "pytorch_model-00001-of-00003.bin",
|
| 162 |
+
"transformer.h.3.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 163 |
+
"transformer.h.3.attn.c_q.weight": "pytorch_model-00001-of-00003.bin",
|
| 164 |
+
"transformer.h.3.attn.c_v.weight": "pytorch_model-00001-of-00003.bin",
|
| 165 |
+
"transformer.h.3.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 166 |
+
"transformer.h.3.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 167 |
+
"transformer.h.3.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 168 |
+
"transformer.h.3.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 169 |
+
"transformer.h.3.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 170 |
+
"transformer.h.4.attn.c_k.weight": "pytorch_model-00001-of-00003.bin",
|
| 171 |
+
"transformer.h.4.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 172 |
+
"transformer.h.4.attn.c_q.weight": "pytorch_model-00001-of-00003.bin",
|
| 173 |
+
"transformer.h.4.attn.c_v.weight": "pytorch_model-00001-of-00003.bin",
|
| 174 |
+
"transformer.h.4.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 175 |
+
"transformer.h.4.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 176 |
+
"transformer.h.4.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 177 |
+
"transformer.h.4.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 178 |
+
"transformer.h.4.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 179 |
+
"transformer.h.5.attn.c_k.weight": "pytorch_model-00001-of-00003.bin",
|
| 180 |
+
"transformer.h.5.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 181 |
+
"transformer.h.5.attn.c_q.weight": "pytorch_model-00001-of-00003.bin",
|
| 182 |
+
"transformer.h.5.attn.c_v.weight": "pytorch_model-00001-of-00003.bin",
|
| 183 |
+
"transformer.h.5.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 184 |
+
"transformer.h.5.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 185 |
+
"transformer.h.5.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 186 |
+
"transformer.h.5.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 187 |
+
"transformer.h.5.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 188 |
+
"transformer.h.6.attn.c_k.weight": "pytorch_model-00001-of-00003.bin",
|
| 189 |
+
"transformer.h.6.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 190 |
+
"transformer.h.6.attn.c_q.weight": "pytorch_model-00001-of-00003.bin",
|
| 191 |
+
"transformer.h.6.attn.c_v.weight": "pytorch_model-00001-of-00003.bin",
|
| 192 |
+
"transformer.h.6.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 193 |
+
"transformer.h.6.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 194 |
+
"transformer.h.6.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 195 |
+
"transformer.h.6.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 196 |
+
"transformer.h.6.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 197 |
+
"transformer.h.7.attn.c_k.weight": "pytorch_model-00001-of-00003.bin",
|
| 198 |
+
"transformer.h.7.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 199 |
+
"transformer.h.7.attn.c_q.weight": "pytorch_model-00001-of-00003.bin",
|
| 200 |
+
"transformer.h.7.attn.c_v.weight": "pytorch_model-00001-of-00003.bin",
|
| 201 |
+
"transformer.h.7.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 202 |
+
"transformer.h.7.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 203 |
+
"transformer.h.7.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 204 |
+
"transformer.h.7.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 205 |
+
"transformer.h.7.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 206 |
+
"transformer.h.8.attn.c_k.weight": "pytorch_model-00001-of-00003.bin",
|
| 207 |
+
"transformer.h.8.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 208 |
+
"transformer.h.8.attn.c_q.weight": "pytorch_model-00001-of-00003.bin",
|
| 209 |
+
"transformer.h.8.attn.c_v.weight": "pytorch_model-00001-of-00003.bin",
|
| 210 |
+
"transformer.h.8.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 211 |
+
"transformer.h.8.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 212 |
+
"transformer.h.8.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 213 |
+
"transformer.h.8.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 214 |
+
"transformer.h.8.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 215 |
+
"transformer.h.9.attn.c_k.weight": "pytorch_model-00001-of-00003.bin",
|
| 216 |
+
"transformer.h.9.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 217 |
+
"transformer.h.9.attn.c_q.weight": "pytorch_model-00001-of-00003.bin",
|
| 218 |
+
"transformer.h.9.attn.c_v.weight": "pytorch_model-00001-of-00003.bin",
|
| 219 |
+
"transformer.h.9.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 220 |
+
"transformer.h.9.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 221 |
+
"transformer.h.9.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 222 |
+
"transformer.h.9.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 223 |
+
"transformer.h.9.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 224 |
+
"transformer.wte.weight": "pytorch_model-00001-of-00003.bin"
|
| 225 |
+
}
|
| 226 |
+
}
|
ablation_initialization/gta_normal_initialization/config.json
ADDED
|
@@ -0,0 +1,18 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"GPT"
|
| 4 |
+
],
|
| 5 |
+
"bias": false,
|
| 6 |
+
"block_size": 2048,
|
| 7 |
+
"dropout": 0.0,
|
| 8 |
+
"group_size": 4,
|
| 9 |
+
"head_dim": 128,
|
| 10 |
+
"intermediate_size": 9960,
|
| 11 |
+
"model_type": "gpt2",
|
| 12 |
+
"n_embd": 3072,
|
| 13 |
+
"n_head": 24,
|
| 14 |
+
"n_layer": 24,
|
| 15 |
+
"torch_dtype": "float32",
|
| 16 |
+
"transformers_version": "4.51.3",
|
| 17 |
+
"vocab_size": 50304
|
| 18 |
+
}
|
ablation_initialization/gta_normal_initialization/pytorch_model-00001-of-00003.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:dc47d482ea37be50ece777f969b249bee42fad35f44b9bffa6ebee403b44c01a
|
| 3 |
+
size 4902475960
|
ablation_initialization/gta_normal_initialization/pytorch_model-00002-of-00003.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:be18d95991c6d9effa19ec0b5bcf8a08dfe06d33af2fbf336f44d74cc155269f
|
| 3 |
+
size 4982057208
|
ablation_initialization/gta_normal_initialization/pytorch_model-00003-of-00003.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cdd8f652e920ca5d3d5ea6a2722819d2853904ae0a261bbe83d22304bf1f48a1
|
| 3 |
+
size 1603558634
|
ablation_initialization/gta_normal_initialization/pytorch_model.bin.index.json
ADDED
|
@@ -0,0 +1,226 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"metadata": {
|
| 3 |
+
"total_size": 11488014336
|
| 4 |
+
},
|
| 5 |
+
"weight_map": {
|
| 6 |
+
"layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 7 |
+
"lm_head.weight": "pytorch_model-00001-of-00003.bin",
|
| 8 |
+
"transformer.h.0.attn.W_rope_k.weight": "pytorch_model-00001-of-00003.bin",
|
| 9 |
+
"transformer.h.0.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 10 |
+
"transformer.h.0.attn.c_q.weight": "pytorch_model-00001-of-00003.bin",
|
| 11 |
+
"transformer.h.0.attn.kv_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 12 |
+
"transformer.h.0.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 13 |
+
"transformer.h.0.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 14 |
+
"transformer.h.0.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 15 |
+
"transformer.h.0.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 16 |
+
"transformer.h.0.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 17 |
+
"transformer.h.1.attn.W_rope_k.weight": "pytorch_model-00001-of-00003.bin",
|
| 18 |
+
"transformer.h.1.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 19 |
+
"transformer.h.1.attn.c_q.weight": "pytorch_model-00001-of-00003.bin",
|
| 20 |
+
"transformer.h.1.attn.kv_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 21 |
+
"transformer.h.1.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 22 |
+
"transformer.h.1.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 23 |
+
"transformer.h.1.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 24 |
+
"transformer.h.1.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 25 |
+
"transformer.h.1.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 26 |
+
"transformer.h.10.attn.W_rope_k.weight": "pytorch_model-00002-of-00003.bin",
|
| 27 |
+
"transformer.h.10.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 28 |
+
"transformer.h.10.attn.c_q.weight": "pytorch_model-00002-of-00003.bin",
|
| 29 |
+
"transformer.h.10.attn.kv_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 30 |
+
"transformer.h.10.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 31 |
+
"transformer.h.10.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 32 |
+
"transformer.h.10.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 33 |
+
"transformer.h.10.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 34 |
+
"transformer.h.10.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 35 |
+
"transformer.h.11.attn.W_rope_k.weight": "pytorch_model-00002-of-00003.bin",
|
| 36 |
+
"transformer.h.11.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 37 |
+
"transformer.h.11.attn.c_q.weight": "pytorch_model-00002-of-00003.bin",
|
| 38 |
+
"transformer.h.11.attn.kv_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 39 |
+
"transformer.h.11.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 40 |
+
"transformer.h.11.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 41 |
+
"transformer.h.11.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 42 |
+
"transformer.h.11.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 43 |
+
"transformer.h.11.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 44 |
+
"transformer.h.12.attn.W_rope_k.weight": "pytorch_model-00002-of-00003.bin",
|
| 45 |
+
"transformer.h.12.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 46 |
+
"transformer.h.12.attn.c_q.weight": "pytorch_model-00002-of-00003.bin",
|
| 47 |
+
"transformer.h.12.attn.kv_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 48 |
+
"transformer.h.12.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 49 |
+
"transformer.h.12.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 50 |
+
"transformer.h.12.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 51 |
+
"transformer.h.12.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 52 |
+
"transformer.h.12.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 53 |
+
"transformer.h.13.attn.W_rope_k.weight": "pytorch_model-00002-of-00003.bin",
|
| 54 |
+
"transformer.h.13.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 55 |
+
"transformer.h.13.attn.c_q.weight": "pytorch_model-00002-of-00003.bin",
|
| 56 |
+
"transformer.h.13.attn.kv_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 57 |
+
"transformer.h.13.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 58 |
+
"transformer.h.13.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 59 |
+
"transformer.h.13.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 60 |
+
"transformer.h.13.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 61 |
+
"transformer.h.13.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 62 |
+
"transformer.h.14.attn.W_rope_k.weight": "pytorch_model-00002-of-00003.bin",
|
| 63 |
+
"transformer.h.14.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 64 |
+
"transformer.h.14.attn.c_q.weight": "pytorch_model-00002-of-00003.bin",
|
| 65 |
+
"transformer.h.14.attn.kv_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 66 |
+
"transformer.h.14.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 67 |
+
"transformer.h.14.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 68 |
+
"transformer.h.14.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 69 |
+
"transformer.h.14.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 70 |
+
"transformer.h.14.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 71 |
+
"transformer.h.15.attn.W_rope_k.weight": "pytorch_model-00002-of-00003.bin",
|
| 72 |
+
"transformer.h.15.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 73 |
+
"transformer.h.15.attn.c_q.weight": "pytorch_model-00002-of-00003.bin",
|
| 74 |
+
"transformer.h.15.attn.kv_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 75 |
+
"transformer.h.15.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 76 |
+
"transformer.h.15.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 77 |
+
"transformer.h.15.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 78 |
+
"transformer.h.15.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 79 |
+
"transformer.h.15.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 80 |
+
"transformer.h.16.attn.W_rope_k.weight": "pytorch_model-00002-of-00003.bin",
|
| 81 |
+
"transformer.h.16.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 82 |
+
"transformer.h.16.attn.c_q.weight": "pytorch_model-00002-of-00003.bin",
|
| 83 |
+
"transformer.h.16.attn.kv_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 84 |
+
"transformer.h.16.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 85 |
+
"transformer.h.16.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 86 |
+
"transformer.h.16.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 87 |
+
"transformer.h.16.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 88 |
+
"transformer.h.16.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 89 |
+
"transformer.h.17.attn.W_rope_k.weight": "pytorch_model-00002-of-00003.bin",
|
| 90 |
+
"transformer.h.17.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 91 |
+
"transformer.h.17.attn.c_q.weight": "pytorch_model-00002-of-00003.bin",
|
| 92 |
+
"transformer.h.17.attn.kv_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 93 |
+
"transformer.h.17.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 94 |
+
"transformer.h.17.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 95 |
+
"transformer.h.17.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 96 |
+
"transformer.h.17.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 97 |
+
"transformer.h.17.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 98 |
+
"transformer.h.18.attn.W_rope_k.weight": "pytorch_model-00002-of-00003.bin",
|
| 99 |
+
"transformer.h.18.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 100 |
+
"transformer.h.18.attn.c_q.weight": "pytorch_model-00002-of-00003.bin",
|
| 101 |
+
"transformer.h.18.attn.kv_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 102 |
+
"transformer.h.18.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 103 |
+
"transformer.h.18.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 104 |
+
"transformer.h.18.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 105 |
+
"transformer.h.18.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 106 |
+
"transformer.h.18.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 107 |
+
"transformer.h.19.attn.W_rope_k.weight": "pytorch_model-00002-of-00003.bin",
|
| 108 |
+
"transformer.h.19.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 109 |
+
"transformer.h.19.attn.c_q.weight": "pytorch_model-00002-of-00003.bin",
|
| 110 |
+
"transformer.h.19.attn.kv_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 111 |
+
"transformer.h.19.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 112 |
+
"transformer.h.19.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 113 |
+
"transformer.h.19.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 114 |
+
"transformer.h.19.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 115 |
+
"transformer.h.19.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 116 |
+
"transformer.h.2.attn.W_rope_k.weight": "pytorch_model-00001-of-00003.bin",
|
| 117 |
+
"transformer.h.2.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 118 |
+
"transformer.h.2.attn.c_q.weight": "pytorch_model-00001-of-00003.bin",
|
| 119 |
+
"transformer.h.2.attn.kv_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 120 |
+
"transformer.h.2.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 121 |
+
"transformer.h.2.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 122 |
+
"transformer.h.2.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 123 |
+
"transformer.h.2.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 124 |
+
"transformer.h.2.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 125 |
+
"transformer.h.20.attn.W_rope_k.weight": "pytorch_model-00002-of-00003.bin",
|
| 126 |
+
"transformer.h.20.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 127 |
+
"transformer.h.20.attn.c_q.weight": "pytorch_model-00002-of-00003.bin",
|
| 128 |
+
"transformer.h.20.attn.kv_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 129 |
+
"transformer.h.20.input_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 130 |
+
"transformer.h.20.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 131 |
+
"transformer.h.20.mlp.c_fc2.weight": "pytorch_model-00003-of-00003.bin",
|
| 132 |
+
"transformer.h.20.mlp.c_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 133 |
+
"transformer.h.20.post_attention_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 134 |
+
"transformer.h.21.attn.W_rope_k.weight": "pytorch_model-00003-of-00003.bin",
|
| 135 |
+
"transformer.h.21.attn.c_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 136 |
+
"transformer.h.21.attn.c_q.weight": "pytorch_model-00003-of-00003.bin",
|
| 137 |
+
"transformer.h.21.attn.kv_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 138 |
+
"transformer.h.21.input_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 139 |
+
"transformer.h.21.mlp.c_fc1.weight": "pytorch_model-00003-of-00003.bin",
|
| 140 |
+
"transformer.h.21.mlp.c_fc2.weight": "pytorch_model-00003-of-00003.bin",
|
| 141 |
+
"transformer.h.21.mlp.c_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 142 |
+
"transformer.h.21.post_attention_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 143 |
+
"transformer.h.22.attn.W_rope_k.weight": "pytorch_model-00003-of-00003.bin",
|
| 144 |
+
"transformer.h.22.attn.c_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 145 |
+
"transformer.h.22.attn.c_q.weight": "pytorch_model-00003-of-00003.bin",
|
| 146 |
+
"transformer.h.22.attn.kv_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 147 |
+
"transformer.h.22.input_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 148 |
+
"transformer.h.22.mlp.c_fc1.weight": "pytorch_model-00003-of-00003.bin",
|
| 149 |
+
"transformer.h.22.mlp.c_fc2.weight": "pytorch_model-00003-of-00003.bin",
|
| 150 |
+
"transformer.h.22.mlp.c_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 151 |
+
"transformer.h.22.post_attention_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 152 |
+
"transformer.h.23.attn.W_rope_k.weight": "pytorch_model-00003-of-00003.bin",
|
| 153 |
+
"transformer.h.23.attn.c_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 154 |
+
"transformer.h.23.attn.c_q.weight": "pytorch_model-00003-of-00003.bin",
|
| 155 |
+
"transformer.h.23.attn.kv_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 156 |
+
"transformer.h.23.input_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 157 |
+
"transformer.h.23.mlp.c_fc1.weight": "pytorch_model-00003-of-00003.bin",
|
| 158 |
+
"transformer.h.23.mlp.c_fc2.weight": "pytorch_model-00003-of-00003.bin",
|
| 159 |
+
"transformer.h.23.mlp.c_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 160 |
+
"transformer.h.23.post_attention_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 161 |
+
"transformer.h.3.attn.W_rope_k.weight": "pytorch_model-00001-of-00003.bin",
|
| 162 |
+
"transformer.h.3.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 163 |
+
"transformer.h.3.attn.c_q.weight": "pytorch_model-00001-of-00003.bin",
|
| 164 |
+
"transformer.h.3.attn.kv_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 165 |
+
"transformer.h.3.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 166 |
+
"transformer.h.3.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 167 |
+
"transformer.h.3.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 168 |
+
"transformer.h.3.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 169 |
+
"transformer.h.3.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 170 |
+
"transformer.h.4.attn.W_rope_k.weight": "pytorch_model-00001-of-00003.bin",
|
| 171 |
+
"transformer.h.4.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 172 |
+
"transformer.h.4.attn.c_q.weight": "pytorch_model-00001-of-00003.bin",
|
| 173 |
+
"transformer.h.4.attn.kv_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 174 |
+
"transformer.h.4.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 175 |
+
"transformer.h.4.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 176 |
+
"transformer.h.4.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 177 |
+
"transformer.h.4.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 178 |
+
"transformer.h.4.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 179 |
+
"transformer.h.5.attn.W_rope_k.weight": "pytorch_model-00001-of-00003.bin",
|
| 180 |
+
"transformer.h.5.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 181 |
+
"transformer.h.5.attn.c_q.weight": "pytorch_model-00001-of-00003.bin",
|
| 182 |
+
"transformer.h.5.attn.kv_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 183 |
+
"transformer.h.5.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 184 |
+
"transformer.h.5.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 185 |
+
"transformer.h.5.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 186 |
+
"transformer.h.5.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 187 |
+
"transformer.h.5.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 188 |
+
"transformer.h.6.attn.W_rope_k.weight": "pytorch_model-00001-of-00003.bin",
|
| 189 |
+
"transformer.h.6.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 190 |
+
"transformer.h.6.attn.c_q.weight": "pytorch_model-00001-of-00003.bin",
|
| 191 |
+
"transformer.h.6.attn.kv_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 192 |
+
"transformer.h.6.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 193 |
+
"transformer.h.6.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 194 |
+
"transformer.h.6.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 195 |
+
"transformer.h.6.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 196 |
+
"transformer.h.6.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 197 |
+
"transformer.h.7.attn.W_rope_k.weight": "pytorch_model-00001-of-00003.bin",
|
| 198 |
+
"transformer.h.7.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 199 |
+
"transformer.h.7.attn.c_q.weight": "pytorch_model-00001-of-00003.bin",
|
| 200 |
+
"transformer.h.7.attn.kv_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 201 |
+
"transformer.h.7.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 202 |
+
"transformer.h.7.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 203 |
+
"transformer.h.7.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 204 |
+
"transformer.h.7.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 205 |
+
"transformer.h.7.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 206 |
+
"transformer.h.8.attn.W_rope_k.weight": "pytorch_model-00001-of-00003.bin",
|
| 207 |
+
"transformer.h.8.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 208 |
+
"transformer.h.8.attn.c_q.weight": "pytorch_model-00001-of-00003.bin",
|
| 209 |
+
"transformer.h.8.attn.kv_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 210 |
+
"transformer.h.8.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 211 |
+
"transformer.h.8.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 212 |
+
"transformer.h.8.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 213 |
+
"transformer.h.8.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 214 |
+
"transformer.h.8.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 215 |
+
"transformer.h.9.attn.W_rope_k.weight": "pytorch_model-00001-of-00003.bin",
|
| 216 |
+
"transformer.h.9.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 217 |
+
"transformer.h.9.attn.c_q.weight": "pytorch_model-00001-of-00003.bin",
|
| 218 |
+
"transformer.h.9.attn.kv_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 219 |
+
"transformer.h.9.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 220 |
+
"transformer.h.9.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 221 |
+
"transformer.h.9.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 222 |
+
"transformer.h.9.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 223 |
+
"transformer.h.9.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 224 |
+
"transformer.wte.weight": "pytorch_model-00001-of-00003.bin"
|
| 225 |
+
}
|
| 226 |
+
}
|
ablation_initialization/mha_normal_initialization/config.json
ADDED
|
@@ -0,0 +1,17 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"GPT"
|
| 4 |
+
],
|
| 5 |
+
"bias": false,
|
| 6 |
+
"block_size": 2048,
|
| 7 |
+
"dropout": 0.0,
|
| 8 |
+
"head_dim": 128,
|
| 9 |
+
"intermediate_size": 8192,
|
| 10 |
+
"model_type": "gpt2",
|
| 11 |
+
"n_embd": 3072,
|
| 12 |
+
"n_head": 24,
|
| 13 |
+
"n_layer": 24,
|
| 14 |
+
"torch_dtype": "float32",
|
| 15 |
+
"transformers_version": "4.51.3",
|
| 16 |
+
"vocab_size": 50304
|
| 17 |
+
}
|
ablation_initialization/mha_normal_initialization/pytorch_model-00001-of-00003.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0837396ef88ae37faaf70575b45495fdecbcaabe9e8703b7ea561468fcf78f2d
|
| 3 |
+
size 4946909328
|
ablation_initialization/mha_normal_initialization/pytorch_model-00002-of-00003.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b1c20da24814e1ac3d0ab9bcb75ac1efa168ad61433c18b70e7605260f94b61d
|
| 3 |
+
size 4983138500
|
ablation_initialization/mha_normal_initialization/pytorch_model-00003-of-00003.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f76c00d3951406b7c1ea9dc60b90a5f70e30466e726694c08ca6f566df3d53f8
|
| 3 |
+
size 1560403114
|
ablation_initialization/mha_normal_initialization/pytorch_model.bin.index.json
ADDED
|
@@ -0,0 +1,226 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"metadata": {
|
| 3 |
+
"total_size": 11490373632
|
| 4 |
+
},
|
| 5 |
+
"weight_map": {
|
| 6 |
+
"layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 7 |
+
"lm_head.weight": "pytorch_model-00001-of-00003.bin",
|
| 8 |
+
"transformer.h.0.attn.c_k.weight": "pytorch_model-00001-of-00003.bin",
|
| 9 |
+
"transformer.h.0.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 10 |
+
"transformer.h.0.attn.c_q.weight": "pytorch_model-00001-of-00003.bin",
|
| 11 |
+
"transformer.h.0.attn.c_v.weight": "pytorch_model-00001-of-00003.bin",
|
| 12 |
+
"transformer.h.0.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 13 |
+
"transformer.h.0.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 14 |
+
"transformer.h.0.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 15 |
+
"transformer.h.0.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 16 |
+
"transformer.h.0.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 17 |
+
"transformer.h.1.attn.c_k.weight": "pytorch_model-00001-of-00003.bin",
|
| 18 |
+
"transformer.h.1.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 19 |
+
"transformer.h.1.attn.c_q.weight": "pytorch_model-00001-of-00003.bin",
|
| 20 |
+
"transformer.h.1.attn.c_v.weight": "pytorch_model-00001-of-00003.bin",
|
| 21 |
+
"transformer.h.1.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 22 |
+
"transformer.h.1.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 23 |
+
"transformer.h.1.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 24 |
+
"transformer.h.1.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 25 |
+
"transformer.h.1.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 26 |
+
"transformer.h.10.attn.c_k.weight": "pytorch_model-00002-of-00003.bin",
|
| 27 |
+
"transformer.h.10.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 28 |
+
"transformer.h.10.attn.c_q.weight": "pytorch_model-00002-of-00003.bin",
|
| 29 |
+
"transformer.h.10.attn.c_v.weight": "pytorch_model-00002-of-00003.bin",
|
| 30 |
+
"transformer.h.10.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 31 |
+
"transformer.h.10.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 32 |
+
"transformer.h.10.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 33 |
+
"transformer.h.10.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 34 |
+
"transformer.h.10.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 35 |
+
"transformer.h.11.attn.c_k.weight": "pytorch_model-00002-of-00003.bin",
|
| 36 |
+
"transformer.h.11.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 37 |
+
"transformer.h.11.attn.c_q.weight": "pytorch_model-00002-of-00003.bin",
|
| 38 |
+
"transformer.h.11.attn.c_v.weight": "pytorch_model-00002-of-00003.bin",
|
| 39 |
+
"transformer.h.11.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 40 |
+
"transformer.h.11.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 41 |
+
"transformer.h.11.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 42 |
+
"transformer.h.11.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 43 |
+
"transformer.h.11.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 44 |
+
"transformer.h.12.attn.c_k.weight": "pytorch_model-00002-of-00003.bin",
|
| 45 |
+
"transformer.h.12.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 46 |
+
"transformer.h.12.attn.c_q.weight": "pytorch_model-00002-of-00003.bin",
|
| 47 |
+
"transformer.h.12.attn.c_v.weight": "pytorch_model-00002-of-00003.bin",
|
| 48 |
+
"transformer.h.12.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 49 |
+
"transformer.h.12.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 50 |
+
"transformer.h.12.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 51 |
+
"transformer.h.12.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 52 |
+
"transformer.h.12.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 53 |
+
"transformer.h.13.attn.c_k.weight": "pytorch_model-00002-of-00003.bin",
|
| 54 |
+
"transformer.h.13.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 55 |
+
"transformer.h.13.attn.c_q.weight": "pytorch_model-00002-of-00003.bin",
|
| 56 |
+
"transformer.h.13.attn.c_v.weight": "pytorch_model-00002-of-00003.bin",
|
| 57 |
+
"transformer.h.13.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 58 |
+
"transformer.h.13.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 59 |
+
"transformer.h.13.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 60 |
+
"transformer.h.13.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 61 |
+
"transformer.h.13.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 62 |
+
"transformer.h.14.attn.c_k.weight": "pytorch_model-00002-of-00003.bin",
|
| 63 |
+
"transformer.h.14.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 64 |
+
"transformer.h.14.attn.c_q.weight": "pytorch_model-00002-of-00003.bin",
|
| 65 |
+
"transformer.h.14.attn.c_v.weight": "pytorch_model-00002-of-00003.bin",
|
| 66 |
+
"transformer.h.14.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 67 |
+
"transformer.h.14.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 68 |
+
"transformer.h.14.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 69 |
+
"transformer.h.14.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 70 |
+
"transformer.h.14.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 71 |
+
"transformer.h.15.attn.c_k.weight": "pytorch_model-00002-of-00003.bin",
|
| 72 |
+
"transformer.h.15.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 73 |
+
"transformer.h.15.attn.c_q.weight": "pytorch_model-00002-of-00003.bin",
|
| 74 |
+
"transformer.h.15.attn.c_v.weight": "pytorch_model-00002-of-00003.bin",
|
| 75 |
+
"transformer.h.15.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 76 |
+
"transformer.h.15.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 77 |
+
"transformer.h.15.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 78 |
+
"transformer.h.15.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 79 |
+
"transformer.h.15.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 80 |
+
"transformer.h.16.attn.c_k.weight": "pytorch_model-00002-of-00003.bin",
|
| 81 |
+
"transformer.h.16.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 82 |
+
"transformer.h.16.attn.c_q.weight": "pytorch_model-00002-of-00003.bin",
|
| 83 |
+
"transformer.h.16.attn.c_v.weight": "pytorch_model-00002-of-00003.bin",
|
| 84 |
+
"transformer.h.16.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 85 |
+
"transformer.h.16.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 86 |
+
"transformer.h.16.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 87 |
+
"transformer.h.16.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 88 |
+
"transformer.h.16.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 89 |
+
"transformer.h.17.attn.c_k.weight": "pytorch_model-00002-of-00003.bin",
|
| 90 |
+
"transformer.h.17.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 91 |
+
"transformer.h.17.attn.c_q.weight": "pytorch_model-00002-of-00003.bin",
|
| 92 |
+
"transformer.h.17.attn.c_v.weight": "pytorch_model-00002-of-00003.bin",
|
| 93 |
+
"transformer.h.17.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 94 |
+
"transformer.h.17.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 95 |
+
"transformer.h.17.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 96 |
+
"transformer.h.17.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 97 |
+
"transformer.h.17.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 98 |
+
"transformer.h.18.attn.c_k.weight": "pytorch_model-00002-of-00003.bin",
|
| 99 |
+
"transformer.h.18.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 100 |
+
"transformer.h.18.attn.c_q.weight": "pytorch_model-00002-of-00003.bin",
|
| 101 |
+
"transformer.h.18.attn.c_v.weight": "pytorch_model-00002-of-00003.bin",
|
| 102 |
+
"transformer.h.18.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 103 |
+
"transformer.h.18.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 104 |
+
"transformer.h.18.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 105 |
+
"transformer.h.18.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 106 |
+
"transformer.h.18.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 107 |
+
"transformer.h.19.attn.c_k.weight": "pytorch_model-00002-of-00003.bin",
|
| 108 |
+
"transformer.h.19.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 109 |
+
"transformer.h.19.attn.c_q.weight": "pytorch_model-00002-of-00003.bin",
|
| 110 |
+
"transformer.h.19.attn.c_v.weight": "pytorch_model-00002-of-00003.bin",
|
| 111 |
+
"transformer.h.19.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 112 |
+
"transformer.h.19.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 113 |
+
"transformer.h.19.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 114 |
+
"transformer.h.19.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 115 |
+
"transformer.h.19.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 116 |
+
"transformer.h.2.attn.c_k.weight": "pytorch_model-00001-of-00003.bin",
|
| 117 |
+
"transformer.h.2.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 118 |
+
"transformer.h.2.attn.c_q.weight": "pytorch_model-00001-of-00003.bin",
|
| 119 |
+
"transformer.h.2.attn.c_v.weight": "pytorch_model-00001-of-00003.bin",
|
| 120 |
+
"transformer.h.2.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 121 |
+
"transformer.h.2.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 122 |
+
"transformer.h.2.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 123 |
+
"transformer.h.2.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 124 |
+
"transformer.h.2.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 125 |
+
"transformer.h.20.attn.c_k.weight": "pytorch_model-00002-of-00003.bin",
|
| 126 |
+
"transformer.h.20.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 127 |
+
"transformer.h.20.attn.c_q.weight": "pytorch_model-00002-of-00003.bin",
|
| 128 |
+
"transformer.h.20.attn.c_v.weight": "pytorch_model-00002-of-00003.bin",
|
| 129 |
+
"transformer.h.20.input_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 130 |
+
"transformer.h.20.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 131 |
+
"transformer.h.20.mlp.c_fc2.weight": "pytorch_model-00003-of-00003.bin",
|
| 132 |
+
"transformer.h.20.mlp.c_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 133 |
+
"transformer.h.20.post_attention_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 134 |
+
"transformer.h.21.attn.c_k.weight": "pytorch_model-00003-of-00003.bin",
|
| 135 |
+
"transformer.h.21.attn.c_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 136 |
+
"transformer.h.21.attn.c_q.weight": "pytorch_model-00003-of-00003.bin",
|
| 137 |
+
"transformer.h.21.attn.c_v.weight": "pytorch_model-00003-of-00003.bin",
|
| 138 |
+
"transformer.h.21.input_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 139 |
+
"transformer.h.21.mlp.c_fc1.weight": "pytorch_model-00003-of-00003.bin",
|
| 140 |
+
"transformer.h.21.mlp.c_fc2.weight": "pytorch_model-00003-of-00003.bin",
|
| 141 |
+
"transformer.h.21.mlp.c_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 142 |
+
"transformer.h.21.post_attention_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 143 |
+
"transformer.h.22.attn.c_k.weight": "pytorch_model-00003-of-00003.bin",
|
| 144 |
+
"transformer.h.22.attn.c_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 145 |
+
"transformer.h.22.attn.c_q.weight": "pytorch_model-00003-of-00003.bin",
|
| 146 |
+
"transformer.h.22.attn.c_v.weight": "pytorch_model-00003-of-00003.bin",
|
| 147 |
+
"transformer.h.22.input_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 148 |
+
"transformer.h.22.mlp.c_fc1.weight": "pytorch_model-00003-of-00003.bin",
|
| 149 |
+
"transformer.h.22.mlp.c_fc2.weight": "pytorch_model-00003-of-00003.bin",
|
| 150 |
+
"transformer.h.22.mlp.c_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 151 |
+
"transformer.h.22.post_attention_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 152 |
+
"transformer.h.23.attn.c_k.weight": "pytorch_model-00003-of-00003.bin",
|
| 153 |
+
"transformer.h.23.attn.c_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 154 |
+
"transformer.h.23.attn.c_q.weight": "pytorch_model-00003-of-00003.bin",
|
| 155 |
+
"transformer.h.23.attn.c_v.weight": "pytorch_model-00003-of-00003.bin",
|
| 156 |
+
"transformer.h.23.input_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 157 |
+
"transformer.h.23.mlp.c_fc1.weight": "pytorch_model-00003-of-00003.bin",
|
| 158 |
+
"transformer.h.23.mlp.c_fc2.weight": "pytorch_model-00003-of-00003.bin",
|
| 159 |
+
"transformer.h.23.mlp.c_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 160 |
+
"transformer.h.23.post_attention_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 161 |
+
"transformer.h.3.attn.c_k.weight": "pytorch_model-00001-of-00003.bin",
|
| 162 |
+
"transformer.h.3.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 163 |
+
"transformer.h.3.attn.c_q.weight": "pytorch_model-00001-of-00003.bin",
|
| 164 |
+
"transformer.h.3.attn.c_v.weight": "pytorch_model-00001-of-00003.bin",
|
| 165 |
+
"transformer.h.3.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 166 |
+
"transformer.h.3.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 167 |
+
"transformer.h.3.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 168 |
+
"transformer.h.3.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 169 |
+
"transformer.h.3.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 170 |
+
"transformer.h.4.attn.c_k.weight": "pytorch_model-00001-of-00003.bin",
|
| 171 |
+
"transformer.h.4.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 172 |
+
"transformer.h.4.attn.c_q.weight": "pytorch_model-00001-of-00003.bin",
|
| 173 |
+
"transformer.h.4.attn.c_v.weight": "pytorch_model-00001-of-00003.bin",
|
| 174 |
+
"transformer.h.4.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 175 |
+
"transformer.h.4.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 176 |
+
"transformer.h.4.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 177 |
+
"transformer.h.4.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 178 |
+
"transformer.h.4.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 179 |
+
"transformer.h.5.attn.c_k.weight": "pytorch_model-00001-of-00003.bin",
|
| 180 |
+
"transformer.h.5.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 181 |
+
"transformer.h.5.attn.c_q.weight": "pytorch_model-00001-of-00003.bin",
|
| 182 |
+
"transformer.h.5.attn.c_v.weight": "pytorch_model-00001-of-00003.bin",
|
| 183 |
+
"transformer.h.5.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 184 |
+
"transformer.h.5.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 185 |
+
"transformer.h.5.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 186 |
+
"transformer.h.5.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 187 |
+
"transformer.h.5.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 188 |
+
"transformer.h.6.attn.c_k.weight": "pytorch_model-00001-of-00003.bin",
|
| 189 |
+
"transformer.h.6.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 190 |
+
"transformer.h.6.attn.c_q.weight": "pytorch_model-00001-of-00003.bin",
|
| 191 |
+
"transformer.h.6.attn.c_v.weight": "pytorch_model-00001-of-00003.bin",
|
| 192 |
+
"transformer.h.6.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 193 |
+
"transformer.h.6.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 194 |
+
"transformer.h.6.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 195 |
+
"transformer.h.6.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 196 |
+
"transformer.h.6.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 197 |
+
"transformer.h.7.attn.c_k.weight": "pytorch_model-00001-of-00003.bin",
|
| 198 |
+
"transformer.h.7.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 199 |
+
"transformer.h.7.attn.c_q.weight": "pytorch_model-00001-of-00003.bin",
|
| 200 |
+
"transformer.h.7.attn.c_v.weight": "pytorch_model-00001-of-00003.bin",
|
| 201 |
+
"transformer.h.7.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 202 |
+
"transformer.h.7.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 203 |
+
"transformer.h.7.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 204 |
+
"transformer.h.7.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 205 |
+
"transformer.h.7.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 206 |
+
"transformer.h.8.attn.c_k.weight": "pytorch_model-00001-of-00003.bin",
|
| 207 |
+
"transformer.h.8.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 208 |
+
"transformer.h.8.attn.c_q.weight": "pytorch_model-00001-of-00003.bin",
|
| 209 |
+
"transformer.h.8.attn.c_v.weight": "pytorch_model-00001-of-00003.bin",
|
| 210 |
+
"transformer.h.8.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 211 |
+
"transformer.h.8.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 212 |
+
"transformer.h.8.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 213 |
+
"transformer.h.8.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 214 |
+
"transformer.h.8.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 215 |
+
"transformer.h.9.attn.c_k.weight": "pytorch_model-00001-of-00003.bin",
|
| 216 |
+
"transformer.h.9.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 217 |
+
"transformer.h.9.attn.c_q.weight": "pytorch_model-00001-of-00003.bin",
|
| 218 |
+
"transformer.h.9.attn.c_v.weight": "pytorch_model-00001-of-00003.bin",
|
| 219 |
+
"transformer.h.9.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 220 |
+
"transformer.h.9.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 221 |
+
"transformer.h.9.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 222 |
+
"transformer.h.9.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 223 |
+
"transformer.h.9.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 224 |
+
"transformer.wte.weight": "pytorch_model-00001-of-00003.bin"
|
| 225 |
+
}
|
| 226 |
+
}
|
ablation_initialization/mla_normal_initialization/config.json
ADDED
|
@@ -0,0 +1,20 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"GPT"
|
| 4 |
+
],
|
| 5 |
+
"bias": false,
|
| 6 |
+
"block_size": 2048,
|
| 7 |
+
"dropout": 0.0,
|
| 8 |
+
"intermediate_size": 9448,
|
| 9 |
+
"kv_lora_rank": 512,
|
| 10 |
+
"model_type": "gpt2",
|
| 11 |
+
"n_embd": 3072,
|
| 12 |
+
"n_head": 24,
|
| 13 |
+
"n_layer": 24,
|
| 14 |
+
"q_lora_rank": 1536,
|
| 15 |
+
"qk_nope_head_dim": 128,
|
| 16 |
+
"qk_rope_head_dim": 64,
|
| 17 |
+
"torch_dtype": "float32",
|
| 18 |
+
"transformers_version": "4.51.3",
|
| 19 |
+
"vocab_size": 50304
|
| 20 |
+
}
|
ablation_initialization/mla_normal_initialization/pytorch_model-00001-of-00003.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:fb0613dc1786ef189bc019452ebac316eca5efe4ec4f434041ee17bbacf75189
|
| 3 |
+
size 4915151597
|
ablation_initialization/mla_normal_initialization/pytorch_model-00002-of-00003.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2e5b48d3c29305d458b880ef3d47abee7f97283fdb34de5cd37853d465cb4e86
|
| 3 |
+
size 4982159218
|
ablation_initialization/mla_normal_initialization/pytorch_model-00003-of-00003.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:34cc93bea34d7b96bc37aec029d16ac04a9e0d0948a9eab6e5a730e1252984bf
|
| 3 |
+
size 1591003468
|
ablation_initialization/mla_normal_initialization/pytorch_model.bin.index.json
ADDED
|
@@ -0,0 +1,298 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"metadata": {
|
| 3 |
+
"total_size": 11488210944
|
| 4 |
+
},
|
| 5 |
+
"weight_map": {
|
| 6 |
+
"layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 7 |
+
"lm_head.weight": "pytorch_model-00001-of-00003.bin",
|
| 8 |
+
"transformer.h.0.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 9 |
+
"transformer.h.0.attn.kv_a_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 10 |
+
"transformer.h.0.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00001-of-00003.bin",
|
| 11 |
+
"transformer.h.0.attn.kv_b_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 12 |
+
"transformer.h.0.attn.q_a_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 13 |
+
"transformer.h.0.attn.q_a_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 14 |
+
"transformer.h.0.attn.q_b_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 15 |
+
"transformer.h.0.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 16 |
+
"transformer.h.0.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 17 |
+
"transformer.h.0.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 18 |
+
"transformer.h.0.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 19 |
+
"transformer.h.0.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 20 |
+
"transformer.h.1.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 21 |
+
"transformer.h.1.attn.kv_a_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 22 |
+
"transformer.h.1.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00001-of-00003.bin",
|
| 23 |
+
"transformer.h.1.attn.kv_b_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 24 |
+
"transformer.h.1.attn.q_a_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 25 |
+
"transformer.h.1.attn.q_a_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 26 |
+
"transformer.h.1.attn.q_b_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 27 |
+
"transformer.h.1.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 28 |
+
"transformer.h.1.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 29 |
+
"transformer.h.1.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 30 |
+
"transformer.h.1.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 31 |
+
"transformer.h.1.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 32 |
+
"transformer.h.10.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 33 |
+
"transformer.h.10.attn.kv_a_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 34 |
+
"transformer.h.10.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00002-of-00003.bin",
|
| 35 |
+
"transformer.h.10.attn.kv_b_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 36 |
+
"transformer.h.10.attn.q_a_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 37 |
+
"transformer.h.10.attn.q_a_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 38 |
+
"transformer.h.10.attn.q_b_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 39 |
+
"transformer.h.10.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 40 |
+
"transformer.h.10.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 41 |
+
"transformer.h.10.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 42 |
+
"transformer.h.10.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 43 |
+
"transformer.h.10.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 44 |
+
"transformer.h.11.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 45 |
+
"transformer.h.11.attn.kv_a_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 46 |
+
"transformer.h.11.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00002-of-00003.bin",
|
| 47 |
+
"transformer.h.11.attn.kv_b_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 48 |
+
"transformer.h.11.attn.q_a_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 49 |
+
"transformer.h.11.attn.q_a_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 50 |
+
"transformer.h.11.attn.q_b_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 51 |
+
"transformer.h.11.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 52 |
+
"transformer.h.11.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 53 |
+
"transformer.h.11.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 54 |
+
"transformer.h.11.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 55 |
+
"transformer.h.11.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 56 |
+
"transformer.h.12.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 57 |
+
"transformer.h.12.attn.kv_a_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 58 |
+
"transformer.h.12.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00002-of-00003.bin",
|
| 59 |
+
"transformer.h.12.attn.kv_b_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 60 |
+
"transformer.h.12.attn.q_a_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 61 |
+
"transformer.h.12.attn.q_a_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 62 |
+
"transformer.h.12.attn.q_b_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 63 |
+
"transformer.h.12.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 64 |
+
"transformer.h.12.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 65 |
+
"transformer.h.12.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 66 |
+
"transformer.h.12.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 67 |
+
"transformer.h.12.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 68 |
+
"transformer.h.13.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 69 |
+
"transformer.h.13.attn.kv_a_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 70 |
+
"transformer.h.13.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00002-of-00003.bin",
|
| 71 |
+
"transformer.h.13.attn.kv_b_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 72 |
+
"transformer.h.13.attn.q_a_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 73 |
+
"transformer.h.13.attn.q_a_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 74 |
+
"transformer.h.13.attn.q_b_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 75 |
+
"transformer.h.13.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 76 |
+
"transformer.h.13.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 77 |
+
"transformer.h.13.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 78 |
+
"transformer.h.13.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 79 |
+
"transformer.h.13.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 80 |
+
"transformer.h.14.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 81 |
+
"transformer.h.14.attn.kv_a_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 82 |
+
"transformer.h.14.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00002-of-00003.bin",
|
| 83 |
+
"transformer.h.14.attn.kv_b_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 84 |
+
"transformer.h.14.attn.q_a_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 85 |
+
"transformer.h.14.attn.q_a_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 86 |
+
"transformer.h.14.attn.q_b_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 87 |
+
"transformer.h.14.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 88 |
+
"transformer.h.14.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 89 |
+
"transformer.h.14.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 90 |
+
"transformer.h.14.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 91 |
+
"transformer.h.14.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 92 |
+
"transformer.h.15.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 93 |
+
"transformer.h.15.attn.kv_a_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 94 |
+
"transformer.h.15.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00002-of-00003.bin",
|
| 95 |
+
"transformer.h.15.attn.kv_b_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 96 |
+
"transformer.h.15.attn.q_a_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 97 |
+
"transformer.h.15.attn.q_a_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 98 |
+
"transformer.h.15.attn.q_b_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 99 |
+
"transformer.h.15.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 100 |
+
"transformer.h.15.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 101 |
+
"transformer.h.15.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 102 |
+
"transformer.h.15.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 103 |
+
"transformer.h.15.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 104 |
+
"transformer.h.16.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 105 |
+
"transformer.h.16.attn.kv_a_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 106 |
+
"transformer.h.16.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00002-of-00003.bin",
|
| 107 |
+
"transformer.h.16.attn.kv_b_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 108 |
+
"transformer.h.16.attn.q_a_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 109 |
+
"transformer.h.16.attn.q_a_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 110 |
+
"transformer.h.16.attn.q_b_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 111 |
+
"transformer.h.16.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 112 |
+
"transformer.h.16.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 113 |
+
"transformer.h.16.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 114 |
+
"transformer.h.16.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 115 |
+
"transformer.h.16.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 116 |
+
"transformer.h.17.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 117 |
+
"transformer.h.17.attn.kv_a_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 118 |
+
"transformer.h.17.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00002-of-00003.bin",
|
| 119 |
+
"transformer.h.17.attn.kv_b_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 120 |
+
"transformer.h.17.attn.q_a_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 121 |
+
"transformer.h.17.attn.q_a_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 122 |
+
"transformer.h.17.attn.q_b_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 123 |
+
"transformer.h.17.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 124 |
+
"transformer.h.17.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 125 |
+
"transformer.h.17.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 126 |
+
"transformer.h.17.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 127 |
+
"transformer.h.17.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 128 |
+
"transformer.h.18.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 129 |
+
"transformer.h.18.attn.kv_a_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 130 |
+
"transformer.h.18.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00002-of-00003.bin",
|
| 131 |
+
"transformer.h.18.attn.kv_b_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 132 |
+
"transformer.h.18.attn.q_a_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 133 |
+
"transformer.h.18.attn.q_a_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 134 |
+
"transformer.h.18.attn.q_b_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 135 |
+
"transformer.h.18.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 136 |
+
"transformer.h.18.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 137 |
+
"transformer.h.18.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 138 |
+
"transformer.h.18.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 139 |
+
"transformer.h.18.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 140 |
+
"transformer.h.19.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 141 |
+
"transformer.h.19.attn.kv_a_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 142 |
+
"transformer.h.19.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00002-of-00003.bin",
|
| 143 |
+
"transformer.h.19.attn.kv_b_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 144 |
+
"transformer.h.19.attn.q_a_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 145 |
+
"transformer.h.19.attn.q_a_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 146 |
+
"transformer.h.19.attn.q_b_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 147 |
+
"transformer.h.19.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 148 |
+
"transformer.h.19.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 149 |
+
"transformer.h.19.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 150 |
+
"transformer.h.19.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 151 |
+
"transformer.h.19.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 152 |
+
"transformer.h.2.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 153 |
+
"transformer.h.2.attn.kv_a_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 154 |
+
"transformer.h.2.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00001-of-00003.bin",
|
| 155 |
+
"transformer.h.2.attn.kv_b_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 156 |
+
"transformer.h.2.attn.q_a_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 157 |
+
"transformer.h.2.attn.q_a_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 158 |
+
"transformer.h.2.attn.q_b_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 159 |
+
"transformer.h.2.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 160 |
+
"transformer.h.2.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 161 |
+
"transformer.h.2.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 162 |
+
"transformer.h.2.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 163 |
+
"transformer.h.2.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 164 |
+
"transformer.h.20.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 165 |
+
"transformer.h.20.attn.kv_a_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 166 |
+
"transformer.h.20.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00002-of-00003.bin",
|
| 167 |
+
"transformer.h.20.attn.kv_b_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 168 |
+
"transformer.h.20.attn.q_a_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 169 |
+
"transformer.h.20.attn.q_a_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 170 |
+
"transformer.h.20.attn.q_b_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 171 |
+
"transformer.h.20.input_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 172 |
+
"transformer.h.20.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 173 |
+
"transformer.h.20.mlp.c_fc2.weight": "pytorch_model-00003-of-00003.bin",
|
| 174 |
+
"transformer.h.20.mlp.c_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 175 |
+
"transformer.h.20.post_attention_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 176 |
+
"transformer.h.21.attn.c_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 177 |
+
"transformer.h.21.attn.kv_a_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 178 |
+
"transformer.h.21.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00003-of-00003.bin",
|
| 179 |
+
"transformer.h.21.attn.kv_b_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 180 |
+
"transformer.h.21.attn.q_a_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 181 |
+
"transformer.h.21.attn.q_a_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 182 |
+
"transformer.h.21.attn.q_b_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 183 |
+
"transformer.h.21.input_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 184 |
+
"transformer.h.21.mlp.c_fc1.weight": "pytorch_model-00003-of-00003.bin",
|
| 185 |
+
"transformer.h.21.mlp.c_fc2.weight": "pytorch_model-00003-of-00003.bin",
|
| 186 |
+
"transformer.h.21.mlp.c_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 187 |
+
"transformer.h.21.post_attention_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 188 |
+
"transformer.h.22.attn.c_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 189 |
+
"transformer.h.22.attn.kv_a_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 190 |
+
"transformer.h.22.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00003-of-00003.bin",
|
| 191 |
+
"transformer.h.22.attn.kv_b_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 192 |
+
"transformer.h.22.attn.q_a_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 193 |
+
"transformer.h.22.attn.q_a_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 194 |
+
"transformer.h.22.attn.q_b_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 195 |
+
"transformer.h.22.input_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 196 |
+
"transformer.h.22.mlp.c_fc1.weight": "pytorch_model-00003-of-00003.bin",
|
| 197 |
+
"transformer.h.22.mlp.c_fc2.weight": "pytorch_model-00003-of-00003.bin",
|
| 198 |
+
"transformer.h.22.mlp.c_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 199 |
+
"transformer.h.22.post_attention_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 200 |
+
"transformer.h.23.attn.c_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 201 |
+
"transformer.h.23.attn.kv_a_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 202 |
+
"transformer.h.23.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00003-of-00003.bin",
|
| 203 |
+
"transformer.h.23.attn.kv_b_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 204 |
+
"transformer.h.23.attn.q_a_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 205 |
+
"transformer.h.23.attn.q_a_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 206 |
+
"transformer.h.23.attn.q_b_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 207 |
+
"transformer.h.23.input_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 208 |
+
"transformer.h.23.mlp.c_fc1.weight": "pytorch_model-00003-of-00003.bin",
|
| 209 |
+
"transformer.h.23.mlp.c_fc2.weight": "pytorch_model-00003-of-00003.bin",
|
| 210 |
+
"transformer.h.23.mlp.c_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 211 |
+
"transformer.h.23.post_attention_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 212 |
+
"transformer.h.3.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 213 |
+
"transformer.h.3.attn.kv_a_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 214 |
+
"transformer.h.3.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00001-of-00003.bin",
|
| 215 |
+
"transformer.h.3.attn.kv_b_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 216 |
+
"transformer.h.3.attn.q_a_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 217 |
+
"transformer.h.3.attn.q_a_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 218 |
+
"transformer.h.3.attn.q_b_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 219 |
+
"transformer.h.3.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 220 |
+
"transformer.h.3.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 221 |
+
"transformer.h.3.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 222 |
+
"transformer.h.3.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 223 |
+
"transformer.h.3.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 224 |
+
"transformer.h.4.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 225 |
+
"transformer.h.4.attn.kv_a_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 226 |
+
"transformer.h.4.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00001-of-00003.bin",
|
| 227 |
+
"transformer.h.4.attn.kv_b_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 228 |
+
"transformer.h.4.attn.q_a_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 229 |
+
"transformer.h.4.attn.q_a_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 230 |
+
"transformer.h.4.attn.q_b_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 231 |
+
"transformer.h.4.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 232 |
+
"transformer.h.4.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 233 |
+
"transformer.h.4.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 234 |
+
"transformer.h.4.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 235 |
+
"transformer.h.4.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 236 |
+
"transformer.h.5.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 237 |
+
"transformer.h.5.attn.kv_a_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 238 |
+
"transformer.h.5.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00001-of-00003.bin",
|
| 239 |
+
"transformer.h.5.attn.kv_b_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 240 |
+
"transformer.h.5.attn.q_a_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 241 |
+
"transformer.h.5.attn.q_a_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 242 |
+
"transformer.h.5.attn.q_b_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 243 |
+
"transformer.h.5.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 244 |
+
"transformer.h.5.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 245 |
+
"transformer.h.5.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 246 |
+
"transformer.h.5.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 247 |
+
"transformer.h.5.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 248 |
+
"transformer.h.6.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 249 |
+
"transformer.h.6.attn.kv_a_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 250 |
+
"transformer.h.6.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00001-of-00003.bin",
|
| 251 |
+
"transformer.h.6.attn.kv_b_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 252 |
+
"transformer.h.6.attn.q_a_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 253 |
+
"transformer.h.6.attn.q_a_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 254 |
+
"transformer.h.6.attn.q_b_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 255 |
+
"transformer.h.6.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 256 |
+
"transformer.h.6.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 257 |
+
"transformer.h.6.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 258 |
+
"transformer.h.6.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 259 |
+
"transformer.h.6.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 260 |
+
"transformer.h.7.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 261 |
+
"transformer.h.7.attn.kv_a_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 262 |
+
"transformer.h.7.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00001-of-00003.bin",
|
| 263 |
+
"transformer.h.7.attn.kv_b_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 264 |
+
"transformer.h.7.attn.q_a_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 265 |
+
"transformer.h.7.attn.q_a_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 266 |
+
"transformer.h.7.attn.q_b_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 267 |
+
"transformer.h.7.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 268 |
+
"transformer.h.7.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 269 |
+
"transformer.h.7.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 270 |
+
"transformer.h.7.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 271 |
+
"transformer.h.7.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 272 |
+
"transformer.h.8.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 273 |
+
"transformer.h.8.attn.kv_a_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 274 |
+
"transformer.h.8.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00001-of-00003.bin",
|
| 275 |
+
"transformer.h.8.attn.kv_b_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 276 |
+
"transformer.h.8.attn.q_a_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 277 |
+
"transformer.h.8.attn.q_a_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 278 |
+
"transformer.h.8.attn.q_b_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 279 |
+
"transformer.h.8.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 280 |
+
"transformer.h.8.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 281 |
+
"transformer.h.8.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 282 |
+
"transformer.h.8.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 283 |
+
"transformer.h.8.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 284 |
+
"transformer.h.9.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 285 |
+
"transformer.h.9.attn.kv_a_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 286 |
+
"transformer.h.9.attn.kv_a_proj_with_mqa.weight": "pytorch_model-00001-of-00003.bin",
|
| 287 |
+
"transformer.h.9.attn.kv_b_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 288 |
+
"transformer.h.9.attn.q_a_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 289 |
+
"transformer.h.9.attn.q_a_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 290 |
+
"transformer.h.9.attn.q_b_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 291 |
+
"transformer.h.9.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 292 |
+
"transformer.h.9.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 293 |
+
"transformer.h.9.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 294 |
+
"transformer.h.9.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 295 |
+
"transformer.h.9.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 296 |
+
"transformer.wte.weight": "pytorch_model-00001-of-00003.bin"
|
| 297 |
+
}
|
| 298 |
+
}
|
ablation_initialization/mqa_normal_initialization/config.json
ADDED
|
@@ -0,0 +1,18 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"GPT"
|
| 4 |
+
],
|
| 5 |
+
"bias": false,
|
| 6 |
+
"block_size": 2048,
|
| 7 |
+
"dropout": 0.0,
|
| 8 |
+
"group_size": 24,
|
| 9 |
+
"head_dim": 128,
|
| 10 |
+
"intermediate_size": 10152,
|
| 11 |
+
"model_type": "gpt2",
|
| 12 |
+
"n_embd": 3072,
|
| 13 |
+
"n_head": 24,
|
| 14 |
+
"n_layer": 24,
|
| 15 |
+
"torch_dtype": "float32",
|
| 16 |
+
"transformers_version": "4.51.3",
|
| 17 |
+
"vocab_size": 50304
|
| 18 |
+
}
|
ablation_initialization/mqa_normal_initialization/pytorch_model-00001-of-00003.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cb18b76bd91302de57a2f999c01af42505291918b129d6c9bdf3656dd9daa6af
|
| 3 |
+
size 4897757304
|
ablation_initialization/mqa_normal_initialization/pytorch_model-00002-of-00003.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b4794827dee4855f9cf8016141b8f86729242aac6782c51e195c7367004997a3
|
| 3 |
+
size 4982057080
|
ablation_initialization/mqa_normal_initialization/pytorch_model-00003-of-00003.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:51e2bf833d07acbe1e3578da5f1add364661a2119bf389d4665e131737df5a0d
|
| 3 |
+
size 1608277162
|
ablation_initialization/mqa_normal_initialization/pytorch_model.bin.index.json
ADDED
|
@@ -0,0 +1,226 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"metadata": {
|
| 3 |
+
"total_size": 11488014336
|
| 4 |
+
},
|
| 5 |
+
"weight_map": {
|
| 6 |
+
"layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 7 |
+
"lm_head.weight": "pytorch_model-00001-of-00003.bin",
|
| 8 |
+
"transformer.h.0.attn.c_k.weight": "pytorch_model-00001-of-00003.bin",
|
| 9 |
+
"transformer.h.0.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 10 |
+
"transformer.h.0.attn.c_q.weight": "pytorch_model-00001-of-00003.bin",
|
| 11 |
+
"transformer.h.0.attn.c_v.weight": "pytorch_model-00001-of-00003.bin",
|
| 12 |
+
"transformer.h.0.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 13 |
+
"transformer.h.0.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 14 |
+
"transformer.h.0.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 15 |
+
"transformer.h.0.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 16 |
+
"transformer.h.0.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 17 |
+
"transformer.h.1.attn.c_k.weight": "pytorch_model-00001-of-00003.bin",
|
| 18 |
+
"transformer.h.1.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 19 |
+
"transformer.h.1.attn.c_q.weight": "pytorch_model-00001-of-00003.bin",
|
| 20 |
+
"transformer.h.1.attn.c_v.weight": "pytorch_model-00001-of-00003.bin",
|
| 21 |
+
"transformer.h.1.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 22 |
+
"transformer.h.1.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 23 |
+
"transformer.h.1.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 24 |
+
"transformer.h.1.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 25 |
+
"transformer.h.1.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 26 |
+
"transformer.h.10.attn.c_k.weight": "pytorch_model-00002-of-00003.bin",
|
| 27 |
+
"transformer.h.10.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 28 |
+
"transformer.h.10.attn.c_q.weight": "pytorch_model-00002-of-00003.bin",
|
| 29 |
+
"transformer.h.10.attn.c_v.weight": "pytorch_model-00002-of-00003.bin",
|
| 30 |
+
"transformer.h.10.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 31 |
+
"transformer.h.10.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 32 |
+
"transformer.h.10.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 33 |
+
"transformer.h.10.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 34 |
+
"transformer.h.10.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 35 |
+
"transformer.h.11.attn.c_k.weight": "pytorch_model-00002-of-00003.bin",
|
| 36 |
+
"transformer.h.11.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 37 |
+
"transformer.h.11.attn.c_q.weight": "pytorch_model-00002-of-00003.bin",
|
| 38 |
+
"transformer.h.11.attn.c_v.weight": "pytorch_model-00002-of-00003.bin",
|
| 39 |
+
"transformer.h.11.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 40 |
+
"transformer.h.11.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 41 |
+
"transformer.h.11.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 42 |
+
"transformer.h.11.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 43 |
+
"transformer.h.11.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 44 |
+
"transformer.h.12.attn.c_k.weight": "pytorch_model-00002-of-00003.bin",
|
| 45 |
+
"transformer.h.12.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 46 |
+
"transformer.h.12.attn.c_q.weight": "pytorch_model-00002-of-00003.bin",
|
| 47 |
+
"transformer.h.12.attn.c_v.weight": "pytorch_model-00002-of-00003.bin",
|
| 48 |
+
"transformer.h.12.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 49 |
+
"transformer.h.12.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 50 |
+
"transformer.h.12.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 51 |
+
"transformer.h.12.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 52 |
+
"transformer.h.12.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 53 |
+
"transformer.h.13.attn.c_k.weight": "pytorch_model-00002-of-00003.bin",
|
| 54 |
+
"transformer.h.13.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 55 |
+
"transformer.h.13.attn.c_q.weight": "pytorch_model-00002-of-00003.bin",
|
| 56 |
+
"transformer.h.13.attn.c_v.weight": "pytorch_model-00002-of-00003.bin",
|
| 57 |
+
"transformer.h.13.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 58 |
+
"transformer.h.13.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 59 |
+
"transformer.h.13.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 60 |
+
"transformer.h.13.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 61 |
+
"transformer.h.13.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 62 |
+
"transformer.h.14.attn.c_k.weight": "pytorch_model-00002-of-00003.bin",
|
| 63 |
+
"transformer.h.14.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 64 |
+
"transformer.h.14.attn.c_q.weight": "pytorch_model-00002-of-00003.bin",
|
| 65 |
+
"transformer.h.14.attn.c_v.weight": "pytorch_model-00002-of-00003.bin",
|
| 66 |
+
"transformer.h.14.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 67 |
+
"transformer.h.14.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 68 |
+
"transformer.h.14.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 69 |
+
"transformer.h.14.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 70 |
+
"transformer.h.14.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 71 |
+
"transformer.h.15.attn.c_k.weight": "pytorch_model-00002-of-00003.bin",
|
| 72 |
+
"transformer.h.15.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 73 |
+
"transformer.h.15.attn.c_q.weight": "pytorch_model-00002-of-00003.bin",
|
| 74 |
+
"transformer.h.15.attn.c_v.weight": "pytorch_model-00002-of-00003.bin",
|
| 75 |
+
"transformer.h.15.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 76 |
+
"transformer.h.15.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 77 |
+
"transformer.h.15.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 78 |
+
"transformer.h.15.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 79 |
+
"transformer.h.15.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 80 |
+
"transformer.h.16.attn.c_k.weight": "pytorch_model-00002-of-00003.bin",
|
| 81 |
+
"transformer.h.16.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 82 |
+
"transformer.h.16.attn.c_q.weight": "pytorch_model-00002-of-00003.bin",
|
| 83 |
+
"transformer.h.16.attn.c_v.weight": "pytorch_model-00002-of-00003.bin",
|
| 84 |
+
"transformer.h.16.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 85 |
+
"transformer.h.16.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 86 |
+
"transformer.h.16.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 87 |
+
"transformer.h.16.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 88 |
+
"transformer.h.16.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 89 |
+
"transformer.h.17.attn.c_k.weight": "pytorch_model-00002-of-00003.bin",
|
| 90 |
+
"transformer.h.17.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 91 |
+
"transformer.h.17.attn.c_q.weight": "pytorch_model-00002-of-00003.bin",
|
| 92 |
+
"transformer.h.17.attn.c_v.weight": "pytorch_model-00002-of-00003.bin",
|
| 93 |
+
"transformer.h.17.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 94 |
+
"transformer.h.17.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 95 |
+
"transformer.h.17.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 96 |
+
"transformer.h.17.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 97 |
+
"transformer.h.17.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 98 |
+
"transformer.h.18.attn.c_k.weight": "pytorch_model-00002-of-00003.bin",
|
| 99 |
+
"transformer.h.18.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 100 |
+
"transformer.h.18.attn.c_q.weight": "pytorch_model-00002-of-00003.bin",
|
| 101 |
+
"transformer.h.18.attn.c_v.weight": "pytorch_model-00002-of-00003.bin",
|
| 102 |
+
"transformer.h.18.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 103 |
+
"transformer.h.18.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 104 |
+
"transformer.h.18.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 105 |
+
"transformer.h.18.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 106 |
+
"transformer.h.18.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 107 |
+
"transformer.h.19.attn.c_k.weight": "pytorch_model-00002-of-00003.bin",
|
| 108 |
+
"transformer.h.19.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 109 |
+
"transformer.h.19.attn.c_q.weight": "pytorch_model-00002-of-00003.bin",
|
| 110 |
+
"transformer.h.19.attn.c_v.weight": "pytorch_model-00002-of-00003.bin",
|
| 111 |
+
"transformer.h.19.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 112 |
+
"transformer.h.19.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 113 |
+
"transformer.h.19.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 114 |
+
"transformer.h.19.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 115 |
+
"transformer.h.19.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 116 |
+
"transformer.h.2.attn.c_k.weight": "pytorch_model-00001-of-00003.bin",
|
| 117 |
+
"transformer.h.2.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 118 |
+
"transformer.h.2.attn.c_q.weight": "pytorch_model-00001-of-00003.bin",
|
| 119 |
+
"transformer.h.2.attn.c_v.weight": "pytorch_model-00001-of-00003.bin",
|
| 120 |
+
"transformer.h.2.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 121 |
+
"transformer.h.2.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 122 |
+
"transformer.h.2.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 123 |
+
"transformer.h.2.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 124 |
+
"transformer.h.2.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 125 |
+
"transformer.h.20.attn.c_k.weight": "pytorch_model-00002-of-00003.bin",
|
| 126 |
+
"transformer.h.20.attn.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 127 |
+
"transformer.h.20.attn.c_q.weight": "pytorch_model-00002-of-00003.bin",
|
| 128 |
+
"transformer.h.20.attn.c_v.weight": "pytorch_model-00002-of-00003.bin",
|
| 129 |
+
"transformer.h.20.input_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 130 |
+
"transformer.h.20.mlp.c_fc1.weight": "pytorch_model-00002-of-00003.bin",
|
| 131 |
+
"transformer.h.20.mlp.c_fc2.weight": "pytorch_model-00003-of-00003.bin",
|
| 132 |
+
"transformer.h.20.mlp.c_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 133 |
+
"transformer.h.20.post_attention_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 134 |
+
"transformer.h.21.attn.c_k.weight": "pytorch_model-00003-of-00003.bin",
|
| 135 |
+
"transformer.h.21.attn.c_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 136 |
+
"transformer.h.21.attn.c_q.weight": "pytorch_model-00003-of-00003.bin",
|
| 137 |
+
"transformer.h.21.attn.c_v.weight": "pytorch_model-00003-of-00003.bin",
|
| 138 |
+
"transformer.h.21.input_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 139 |
+
"transformer.h.21.mlp.c_fc1.weight": "pytorch_model-00003-of-00003.bin",
|
| 140 |
+
"transformer.h.21.mlp.c_fc2.weight": "pytorch_model-00003-of-00003.bin",
|
| 141 |
+
"transformer.h.21.mlp.c_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 142 |
+
"transformer.h.21.post_attention_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 143 |
+
"transformer.h.22.attn.c_k.weight": "pytorch_model-00003-of-00003.bin",
|
| 144 |
+
"transformer.h.22.attn.c_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 145 |
+
"transformer.h.22.attn.c_q.weight": "pytorch_model-00003-of-00003.bin",
|
| 146 |
+
"transformer.h.22.attn.c_v.weight": "pytorch_model-00003-of-00003.bin",
|
| 147 |
+
"transformer.h.22.input_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 148 |
+
"transformer.h.22.mlp.c_fc1.weight": "pytorch_model-00003-of-00003.bin",
|
| 149 |
+
"transformer.h.22.mlp.c_fc2.weight": "pytorch_model-00003-of-00003.bin",
|
| 150 |
+
"transformer.h.22.mlp.c_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 151 |
+
"transformer.h.22.post_attention_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 152 |
+
"transformer.h.23.attn.c_k.weight": "pytorch_model-00003-of-00003.bin",
|
| 153 |
+
"transformer.h.23.attn.c_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 154 |
+
"transformer.h.23.attn.c_q.weight": "pytorch_model-00003-of-00003.bin",
|
| 155 |
+
"transformer.h.23.attn.c_v.weight": "pytorch_model-00003-of-00003.bin",
|
| 156 |
+
"transformer.h.23.input_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 157 |
+
"transformer.h.23.mlp.c_fc1.weight": "pytorch_model-00003-of-00003.bin",
|
| 158 |
+
"transformer.h.23.mlp.c_fc2.weight": "pytorch_model-00003-of-00003.bin",
|
| 159 |
+
"transformer.h.23.mlp.c_proj.weight": "pytorch_model-00003-of-00003.bin",
|
| 160 |
+
"transformer.h.23.post_attention_layernorm.weight": "pytorch_model-00003-of-00003.bin",
|
| 161 |
+
"transformer.h.3.attn.c_k.weight": "pytorch_model-00001-of-00003.bin",
|
| 162 |
+
"transformer.h.3.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 163 |
+
"transformer.h.3.attn.c_q.weight": "pytorch_model-00001-of-00003.bin",
|
| 164 |
+
"transformer.h.3.attn.c_v.weight": "pytorch_model-00001-of-00003.bin",
|
| 165 |
+
"transformer.h.3.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 166 |
+
"transformer.h.3.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 167 |
+
"transformer.h.3.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 168 |
+
"transformer.h.3.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 169 |
+
"transformer.h.3.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 170 |
+
"transformer.h.4.attn.c_k.weight": "pytorch_model-00001-of-00003.bin",
|
| 171 |
+
"transformer.h.4.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 172 |
+
"transformer.h.4.attn.c_q.weight": "pytorch_model-00001-of-00003.bin",
|
| 173 |
+
"transformer.h.4.attn.c_v.weight": "pytorch_model-00001-of-00003.bin",
|
| 174 |
+
"transformer.h.4.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 175 |
+
"transformer.h.4.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 176 |
+
"transformer.h.4.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 177 |
+
"transformer.h.4.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 178 |
+
"transformer.h.4.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 179 |
+
"transformer.h.5.attn.c_k.weight": "pytorch_model-00001-of-00003.bin",
|
| 180 |
+
"transformer.h.5.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 181 |
+
"transformer.h.5.attn.c_q.weight": "pytorch_model-00001-of-00003.bin",
|
| 182 |
+
"transformer.h.5.attn.c_v.weight": "pytorch_model-00001-of-00003.bin",
|
| 183 |
+
"transformer.h.5.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 184 |
+
"transformer.h.5.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 185 |
+
"transformer.h.5.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 186 |
+
"transformer.h.5.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 187 |
+
"transformer.h.5.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 188 |
+
"transformer.h.6.attn.c_k.weight": "pytorch_model-00001-of-00003.bin",
|
| 189 |
+
"transformer.h.6.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 190 |
+
"transformer.h.6.attn.c_q.weight": "pytorch_model-00001-of-00003.bin",
|
| 191 |
+
"transformer.h.6.attn.c_v.weight": "pytorch_model-00001-of-00003.bin",
|
| 192 |
+
"transformer.h.6.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 193 |
+
"transformer.h.6.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 194 |
+
"transformer.h.6.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 195 |
+
"transformer.h.6.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 196 |
+
"transformer.h.6.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 197 |
+
"transformer.h.7.attn.c_k.weight": "pytorch_model-00001-of-00003.bin",
|
| 198 |
+
"transformer.h.7.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 199 |
+
"transformer.h.7.attn.c_q.weight": "pytorch_model-00001-of-00003.bin",
|
| 200 |
+
"transformer.h.7.attn.c_v.weight": "pytorch_model-00001-of-00003.bin",
|
| 201 |
+
"transformer.h.7.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 202 |
+
"transformer.h.7.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 203 |
+
"transformer.h.7.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 204 |
+
"transformer.h.7.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 205 |
+
"transformer.h.7.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 206 |
+
"transformer.h.8.attn.c_k.weight": "pytorch_model-00001-of-00003.bin",
|
| 207 |
+
"transformer.h.8.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 208 |
+
"transformer.h.8.attn.c_q.weight": "pytorch_model-00001-of-00003.bin",
|
| 209 |
+
"transformer.h.8.attn.c_v.weight": "pytorch_model-00001-of-00003.bin",
|
| 210 |
+
"transformer.h.8.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 211 |
+
"transformer.h.8.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 212 |
+
"transformer.h.8.mlp.c_fc2.weight": "pytorch_model-00001-of-00003.bin",
|
| 213 |
+
"transformer.h.8.mlp.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 214 |
+
"transformer.h.8.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
|
| 215 |
+
"transformer.h.9.attn.c_k.weight": "pytorch_model-00001-of-00003.bin",
|
| 216 |
+
"transformer.h.9.attn.c_proj.weight": "pytorch_model-00001-of-00003.bin",
|
| 217 |
+
"transformer.h.9.attn.c_q.weight": "pytorch_model-00001-of-00003.bin",
|
| 218 |
+
"transformer.h.9.attn.c_v.weight": "pytorch_model-00001-of-00003.bin",
|
| 219 |
+
"transformer.h.9.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 220 |
+
"transformer.h.9.mlp.c_fc1.weight": "pytorch_model-00001-of-00003.bin",
|
| 221 |
+
"transformer.h.9.mlp.c_fc2.weight": "pytorch_model-00002-of-00003.bin",
|
| 222 |
+
"transformer.h.9.mlp.c_proj.weight": "pytorch_model-00002-of-00003.bin",
|
| 223 |
+
"transformer.h.9.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
|
| 224 |
+
"transformer.wte.weight": "pytorch_model-00001-of-00003.bin"
|
| 225 |
+
}
|
| 226 |
+
}
|
fineweb-edu100B/fineweb_train_000001.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6bb7ce7bcac8e11463433767ec3402311c7527c3d8d766e7d65ef86dc4546bb2
|
| 3 |
+
size 200001024
|
fineweb-edu100B/fineweb_train_000002.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1143f8ca46f1700c5a50550bd3846fba733bb7f325b25930c32123f6b1c4b07b
|
| 3 |
+
size 200001024
|
fineweb-edu100B/fineweb_train_000003.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2a8d025e7816e14eb9f58e13481445511ac8590ad564f6dc8938fd8c10e5a572
|
| 3 |
+
size 200001024
|
fineweb-edu100B/fineweb_train_000004.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c068d8f775f3d5c161703869bb7fcbf9e3039ee0c1be875def40b2f78f0cddc3
|
| 3 |
+
size 200001024
|
fineweb-edu100B/fineweb_train_000005.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b0bb385b80d8029e8e2c7181d60616883aff3f5004b349bedff38c195ac43c45
|
| 3 |
+
size 200001024
|
fineweb-edu100B/fineweb_train_000006.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:73786405a21c30173cce692db9d776b4f2aa4233a2df66281433e79c79d3a9b8
|
| 3 |
+
size 200001024
|
fineweb-edu100B/fineweb_train_000007.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6b858eaa2f71963a4762c7362be426e6938a9463e8539d64982b9d83d90b5c9d
|
| 3 |
+
size 200001024
|
fineweb-edu100B/fineweb_train_000008.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8a459610f9b4229d787f39748070b0ff126c402e56e27bf0666b8924a786de1b
|
| 3 |
+
size 200001024
|
fineweb-edu100B/fineweb_train_000009.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9b3b462df8835f5b42b2844521788a33e7fdadbf1d160f84dd78fefa349041b7
|
| 3 |
+
size 200001024
|
fineweb-edu100B/fineweb_train_000010.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e67647ba7bdb6cc5d43e3cc8f1bbc4bd06b7728c5a59464b7f3846a2af22a8a8
|
| 3 |
+
size 200001024
|
fineweb-edu100B/fineweb_train_000011.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ab9532f842667f09030df577ec8f6c49ea805221e3951123e9ec237f0ec958f1
|
| 3 |
+
size 200001024
|
fineweb-edu100B/fineweb_train_000012.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:811702e10a25df10f70c117fd8ea0222494481db24c3c3fc82f461b2dde1449e
|
| 3 |
+
size 200001024
|
fineweb-edu100B/fineweb_train_000013.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1886f8783ff7ca653e6e6f55f7c0f38738bd9312639dfd900f08d27c96f8b522
|
| 3 |
+
size 200001024
|
fineweb-edu100B/fineweb_train_000014.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7cbf2ca0ca46465651d123df8e6b7927c0f048efc141109f341ecade37c63804
|
| 3 |
+
size 200001024
|
fineweb-edu100B/fineweb_train_000015.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b92cda9a41098de3d8d7cbccf8701d75651c273c78439fe65119451a171e143c
|
| 3 |
+
size 200001024
|