vihangd/smartyplats-7b-v2
1
0
Fork 0
Code Issues Pull Requests Actions Projects Releases Wiki Activity

初始化项目，由ModelHub XC社区提供模型

Browse Source 
Model: vihangd/smartyplats-7b-v2
Source: Original Platform
This commit is contained in:
ModelHub XC
2026-05-29 05:56:39 +08:00
commit b805d9b402
93 changed files with 92090 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

35
.gitattributes vendored Normal file
View File

@@ -0,0 +1,35 @@
*.7z filter=lfs diff=lfs merge=lfs -text
*.arrow filter=lfs diff=lfs merge=lfs -text
*.bin filter=lfs diff=lfs merge=lfs -text
*.bz2 filter=lfs diff=lfs merge=lfs -text
*.ckpt filter=lfs diff=lfs merge=lfs -text
*.ftz filter=lfs diff=lfs merge=lfs -text
*.gz filter=lfs diff=lfs merge=lfs -text
*.h5 filter=lfs diff=lfs merge=lfs -text
*.joblib filter=lfs diff=lfs merge=lfs -text
*.lfs.* filter=lfs diff=lfs merge=lfs -text
*.mlmodel filter=lfs diff=lfs merge=lfs -text
*.model filter=lfs diff=lfs merge=lfs -text
*.msgpack filter=lfs diff=lfs merge=lfs -text
*.npy filter=lfs diff=lfs merge=lfs -text
*.npz filter=lfs diff=lfs merge=lfs -text
*.onnx filter=lfs diff=lfs merge=lfs -text
*.ot filter=lfs diff=lfs merge=lfs -text
*.parquet filter=lfs diff=lfs merge=lfs -text
*.pb filter=lfs diff=lfs merge=lfs -text
*.pickle filter=lfs diff=lfs merge=lfs -text
*.pkl filter=lfs diff=lfs merge=lfs -text
*.pt filter=lfs diff=lfs merge=lfs -text
*.pth filter=lfs diff=lfs merge=lfs -text
*.rar filter=lfs diff=lfs merge=lfs -text
*.safetensors filter=lfs diff=lfs merge=lfs -text
saved_model/**/* filter=lfs diff=lfs merge=lfs -text
*.tar.* filter=lfs diff=lfs merge=lfs -text
*.tar filter=lfs diff=lfs merge=lfs -text
*.tflite filter=lfs diff=lfs merge=lfs -text
*.tgz filter=lfs diff=lfs merge=lfs -text
*.wasm filter=lfs diff=lfs merge=lfs -text
*.xz filter=lfs diff=lfs merge=lfs -text
*.zip filter=lfs diff=lfs merge=lfs -text
*.zst filter=lfs diff=lfs merge=lfs -text
*tfevents* filter=lfs diff=lfs merge=lfs -text

11
README.md Normal file
View File

@@ -0,0 +1,11 @@
---
license: apache-2.0
---
<p><h1> SmartyPlats-7b-v2 </h1></p>
An experimental finetune of Mistrel 7b with QLoRA
<h2> Datasets </h2>
Trained on alpca style datasets
<p><h2> Prompt Template </h2></p>
Uses alpaca style prompt template

25
config.json Normal file
View File

@@ -0,0 +1,25 @@
{
"_name_or_path": "../alpaca-qlora/smartyplats-7b-v2",
"architectures": [
"MistralForCausalLM"
],
"bos_token_id": 1,
"eos_token_id": 2,
"hidden_act": "silu",
"hidden_size": 4096,
"initializer_range": 0.02,
"intermediate_size": 14336,
"max_position_embeddings": 32768,
"model_type": "mistral",
"num_attention_heads": 32,
"num_hidden_layers": 32,
"num_key_value_heads": 8,
"rms_norm_eps": 1e-05,
"rope_theta": 10000.0,
"sliding_window": 4096,
"tie_word_embeddings": false,
"torch_dtype": "float16",
"transformers_version": "4.35.0.dev0",
"use_cache": true,
"vocab_size": 32000
}

6
generation_config.json Normal file
View File

@@ -0,0 +1,6 @@
{
"_from_model_config": true,
"bos_token_id": 1,
"eos_token_id": 2,
"transformers_version": "4.35.0.dev0"
}

3
model-00001-of-00042.safetensors Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:fb13b8aba76c4ceeac19a3cc3a345378e08f76208b9ccf3f109e3c1543317ca0
size 346030672

3
model-00002-of-00042.safetensors Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:f2641a51db86b3bbee125b38a8d86175bcfc11c0772728d35da63b578649b79a
size 394281776

3
model-00003-of-00042.safetensors Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:5e8307ba3bd0e7f0fd3137c76b8cddc26c4b0d640be48055112f4e638c4e5591
size 394281776

3
model-00004-of-00042.safetensors Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:019fa521981101b1430e3816eb11ccbf178fa829c98cd9fcf0e1709333294125
size 318767816

3
model-00005-of-00042.safetensors Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:418f365b48b50b148db99dd1b02093edcec95424a6d08930e16c6da003ca7b9a
size 318784416

3
model-00006-of-00042.safetensors Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:8ee0310717d482abd42628526c7f4d5586c83cdfc9d61f25851d61c4adcc3a16
size 318784416

3
model-00007-of-00042.safetensors Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:361e1cdb1ade646f48a65c16e695afd539138e7de83f7f771f81542df021309a
size 394281776

3
model-00008-of-00042.safetensors Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:1474c28dc6840bc519248f46380f041db68152e370ba80060711b2907ced2922
size 394281776

3
model-00009-of-00042.safetensors Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:f56b19a0cfd6cd13019e7ab081fd065fd30e7fa11c7cee962184d221199f63c6
size 318767816

3
model-00010-of-00042.safetensors Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:322f50e13909bba476012396405cf8225ccb8793d9bc46c7fe4c0fd57fc69818
size 318784416

3
model-00011-of-00042.safetensors Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:1b2b662271ebc5cdbc98a13b69ef7798197dd5a5570e03a56ec26acbd162900d
size 318784416

3
model-00012-of-00042.safetensors Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:2ac6c5614e55bd61057f61f5b327d428ee68b27f3c32e605cca477f0d24346be
size 394281776

3
model-00013-of-00042.safetensors Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:0bccac819f8becdb76f4c0e5390e421619433fd0ace2e6e1475f467de3f6f6ac
size 394281776

3
model-00014-of-00042.safetensors Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:90c9a08b9aa26ebbe5722aa34e19aae708f313b8a9298a9dd4a78765ac9b00e7
size 318767824

3
model-00015-of-00042.safetensors Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:8bd98d323075c0b82c79b5d832b6ffc143fd3ca25a572b3d54f57a1871f0ec82
size 318784424

3
model-00016-of-00042.safetensors Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:22aec46eece3adbca3baeeafd2d0846bf9312fa73058282e9ac670720c83e701
size 318784424

3
model-00017-of-00042.safetensors Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:38fb7b49151690af7896fa8c845009171b30929209c933301a0d9af69ddb72bf
size 394281776

3
model-00018-of-00042.safetensors Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:153041583f4182b77e480dc55ea017bf3882372a94ba181a23a24aefea2a448d
size 394281776

3
model-00019-of-00042.safetensors Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:f7c6a0694f0fc60318c0e616849714247dfac9d4950df8a45d51b897221bb2bf
size 318767824

3
model-00020-of-00042.safetensors Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:cbada724f5b20b8caa938b04efbe6b4af65f4c2048791a6fd9027ed6dc85cd5f
size 318784424

3
model-00021-of-00042.safetensors Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:1acdfe0fbf0317bb6a11da0174ba0db1ea80d5d50adeafe858052db4ff4b2938
size 318784424

3
model-00022-of-00042.safetensors Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:57c155b6fe5b6c287e96f8aae228086a4e127cdfef158d8d0de7d9afb7bc864a
size 394281776

3
model-00023-of-00042.safetensors Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:35ddb6dddc58327b6826309042f46534a0fcf977d0b96b2c8d9d1be5821496c7
size 394281776

3
model-00024-of-00042.safetensors Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:e9f7ae713d21ba980873e42d2495a4a70e90c409c5e220451bb9c9ca511c1f31
size 318767824

3
model-00025-of-00042.safetensors Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:a06aaf93bb4fd236f64a01064a865b8f5bf0da4fd1b68931a9f50551b99206c6
size 318784424

3
model-00026-of-00042.safetensors Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:035f3a278942e5478a396fd4bf28bdd682332c92fed321274bb3138ca0092561
size 318784424

3
model-00027-of-00042.safetensors Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:ad9cf639a44d112ca37ef83545a9795fff7eb4bb57e4d86f9b9292303cf2ff12
size 394281776

3
model-00028-of-00042.safetensors Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:72995d4e88445687517cca787d547ab5a6b2051e32887f83d60a2393ef4d5697
size 394281776

3
model-00029-of-00042.safetensors Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:3294da4a269d0909df31dc23c0072e1366adef0f55af21f7f5087fff26c1f25f
size 318767824

3
model-00030-of-00042.safetensors Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:cc2b03bfcb35f711f92734166099af061c8648c36c882677d61aff891fcc8767
size 318784424

3
model-00031-of-00042.safetensors Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:5e5230d6cf205c5f99b23e6dadaee1805ac41b28047e98ab8faa43b848ff0cfa
size 318784424

3
model-00032-of-00042.safetensors Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:499ad2d27a0a0c3653f968e99a761fcef4d68a07e347fb9b2ea26c3089125927
size 394281776

3
model-00033-of-00042.safetensors Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:eb9e6ceac64d0970ac611f5dae1b7e0e1b63645bf9130f379b5d5c5cc3626055
size 394281776

3
model-00034-of-00042.safetensors Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:93d173f3aeaf651f119cdeacb60de0d2e7282b7f06a1f7422e8dd0db72381d59
size 318767824

3
model-00035-of-00042.safetensors Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:94f9a0bc7a135b70fe530d8deb76f46f6d58260528eb7bdb4d6f7e59b1bef67a
size 318784424

3
model-00036-of-00042.safetensors Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:eaf691f55e773ec98395a117dbea84dcd197e3d8c373eb01020eb1934ad85d4c
size 318784424

3
model-00037-of-00042.safetensors Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:cbe721771b9f19fac3a360b57045ddd266f7c5d866af9f7f632167845e2b1211
size 394281776

3
model-00038-of-00042.safetensors Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:73f67ef7fb08789ef4e19f1507e5efb0e19a36f59fbaeb2c77b1db6694b63bd5
size 394281776

3
model-00039-of-00042.safetensors Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:edc0808025a57b635844570c418fd17eff5761eab0423631cfe91e02af0bfc06
size 318767824

3
model-00040-of-00042.safetensors Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:1ca5f139adfa061d63c2dc0dd0fdc7c875321c809cdfafef0167a160bbffd70e
size 318784424

3
model-00041-of-00042.safetensors Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:6b19be890cb1f9dce093ea6b21b01eb58c83d64dab82bf0d014f560c456d24b6
size 234906168

3
model-00042-of-00042.safetensors Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:034b1ae1bfcb814533780bdc91f78dfee402c1d08e3ebe15123ae7c9080dc22f
size 262144128

298
model.safetensors.index.json Normal file
View File

@@ -0,0 +1,298 @@
{
"metadata": {
"total_size": 14483464192
},
"weight_map": {
"lm_head.weight": "model-00042-of-00042.safetensors",
"model.embed_tokens.weight": "model-00001-of-00042.safetensors",
"model.layers.0.input_layernorm.weight": "model-00002-of-00042.safetensors",
"model.layers.0.mlp.down_proj.weight": "model-00002-of-00042.safetensors",
"model.layers.0.mlp.gate_proj.weight": "model-00002-of-00042.safetensors",
"model.layers.0.mlp.up_proj.weight": "model-00002-of-00042.safetensors",
"model.layers.0.post_attention_layernorm.weight": "model-00002-of-00042.safetensors",
"model.layers.0.self_attn.k_proj.weight": "model-00001-of-00042.safetensors",
"model.layers.0.self_attn.o_proj.weight": "model-00001-of-00042.safetensors",
"model.layers.0.self_attn.q_proj.weight": "model-00001-of-00042.safetensors",
"model.layers.0.self_attn.v_proj.weight": "model-00001-of-00042.safetensors",
"model.layers.1.input_layernorm.weight": "model-00003-of-00042.safetensors",
"model.layers.1.mlp.down_proj.weight": "model-00003-of-00042.safetensors",
"model.layers.1.mlp.gate_proj.weight": "model-00003-of-00042.safetensors",
"model.layers.1.mlp.up_proj.weight": "model-00003-of-00042.safetensors",
"model.layers.1.post_attention_layernorm.weight": "model-00003-of-00042.safetensors",
"model.layers.1.self_attn.k_proj.weight": "model-00002-of-00042.safetensors",
"model.layers.1.self_attn.o_proj.weight": "model-00003-of-00042.safetensors",
"model.layers.1.self_attn.q_proj.weight": "model-00002-of-00042.safetensors",
"model.layers.1.self_attn.v_proj.weight": "model-00003-of-00042.safetensors",
"model.layers.10.input_layernorm.weight": "model-00015-of-00042.safetensors",
"model.layers.10.mlp.down_proj.weight": "model-00015-of-00042.safetensors",
"model.layers.10.mlp.gate_proj.weight": "model-00014-of-00042.safetensors",
"model.layers.10.mlp.up_proj.weight": "model-00014-of-00042.safetensors",
"model.layers.10.post_attention_layernorm.weight": "model-00015-of-00042.safetensors",
"model.layers.10.self_attn.k_proj.weight": "model-00014-of-00042.safetensors",
"model.layers.10.self_attn.o_proj.weight": "model-00014-of-00042.safetensors",
"model.layers.10.self_attn.q_proj.weight": "model-00014-of-00042.safetensors",
"model.layers.10.self_attn.v_proj.weight": "model-00014-of-00042.safetensors",
"model.layers.11.input_layernorm.weight": "model-00016-of-00042.safetensors",
"model.layers.11.mlp.down_proj.weight": "model-00016-of-00042.safetensors",
"model.layers.11.mlp.gate_proj.weight": "model-00015-of-00042.safetensors",
"model.layers.11.mlp.up_proj.weight": "model-00016-of-00042.safetensors",
"model.layers.11.post_attention_layernorm.weight": "model-00016-of-00042.safetensors",
"model.layers.11.self_attn.k_proj.weight": "model-00015-of-00042.safetensors",
"model.layers.11.self_attn.o_proj.weight": "model-00015-of-00042.safetensors",
"model.layers.11.self_attn.q_proj.weight": "model-00015-of-00042.safetensors",
"model.layers.11.self_attn.v_proj.weight": "model-00015-of-00042.safetensors",
"model.layers.12.input_layernorm.weight": "model-00017-of-00042.safetensors",
"model.layers.12.mlp.down_proj.weight": "model-00017-of-00042.safetensors",
"model.layers.12.mlp.gate_proj.weight": "model-00017-of-00042.safetensors",
"model.layers.12.mlp.up_proj.weight": "model-00017-of-00042.safetensors",
"model.layers.12.post_attention_layernorm.weight": "model-00017-of-00042.safetensors",
"model.layers.12.self_attn.k_proj.weight": "model-00016-of-00042.safetensors",
"model.layers.12.self_attn.o_proj.weight": "model-00016-of-00042.safetensors",
"model.layers.12.self_attn.q_proj.weight": "model-00016-of-00042.safetensors",
"model.layers.12.self_attn.v_proj.weight": "model-00016-of-00042.safetensors",
"model.layers.13.input_layernorm.weight": "model-00018-of-00042.safetensors",
"model.layers.13.mlp.down_proj.weight": "model-00018-of-00042.safetensors",
"model.layers.13.mlp.gate_proj.weight": "model-00018-of-00042.safetensors",
"model.layers.13.mlp.up_proj.weight": "model-00018-of-00042.safetensors",
"model.layers.13.post_attention_layernorm.weight": "model-00018-of-00042.safetensors",
"model.layers.13.self_attn.k_proj.weight": "model-00017-of-00042.safetensors",
"model.layers.13.self_attn.o_proj.weight": "model-00018-of-00042.safetensors",
"model.layers.13.self_attn.q_proj.weight": "model-00017-of-00042.safetensors",
"model.layers.13.self_attn.v_proj.weight": "model-00018-of-00042.safetensors",
"model.layers.14.input_layernorm.weight": "model-00020-of-00042.safetensors",
"model.layers.14.mlp.down_proj.weight": "model-00020-of-00042.safetensors",
"model.layers.14.mlp.gate_proj.weight": "model-00019-of-00042.safetensors",
"model.layers.14.mlp.up_proj.weight": "model-00019-of-00042.safetensors",
"model.layers.14.post_attention_layernorm.weight": "model-00020-of-00042.safetensors",
"model.layers.14.self_attn.k_proj.weight": "model-00019-of-00042.safetensors",
"model.layers.14.self_attn.o_proj.weight": "model-00019-of-00042.safetensors",
"model.layers.14.self_attn.q_proj.weight": "model-00019-of-00042.safetensors",
"model.layers.14.self_attn.v_proj.weight": "model-00019-of-00042.safetensors",
"model.layers.15.input_layernorm.weight": "model-00021-of-00042.safetensors",
"model.layers.15.mlp.down_proj.weight": "model-00021-of-00042.safetensors",
"model.layers.15.mlp.gate_proj.weight": "model-00020-of-00042.safetensors",
"model.layers.15.mlp.up_proj.weight": "model-00021-of-00042.safetensors",
"model.layers.15.post_attention_layernorm.weight": "model-00021-of-00042.safetensors",
"model.layers.15.self_attn.k_proj.weight": "model-00020-of-00042.safetensors",
"model.layers.15.self_attn.o_proj.weight": "model-00020-of-00042.safetensors",
"model.layers.15.self_attn.q_proj.weight": "model-00020-of-00042.safetensors",
"model.layers.15.self_attn.v_proj.weight": "model-00020-of-00042.safetensors",
"model.layers.16.input_layernorm.weight": "model-00022-of-00042.safetensors",
"model.layers.16.mlp.down_proj.weight": "model-00022-of-00042.safetensors",
"model.layers.16.mlp.gate_proj.weight": "model-00022-of-00042.safetensors",
"model.layers.16.mlp.up_proj.weight": "model-00022-of-00042.safetensors",
"model.layers.16.post_attention_layernorm.weight": "model-00022-of-00042.safetensors",
"model.layers.16.self_attn.k_proj.weight": "model-00021-of-00042.safetensors",
"model.layers.16.self_attn.o_proj.weight": "model-00021-of-00042.safetensors",
"model.layers.16.self_attn.q_proj.weight": "model-00021-of-00042.safetensors",
"model.layers.16.self_attn.v_proj.weight": "model-00021-of-00042.safetensors",
"model.layers.17.input_layernorm.weight": "model-00023-of-00042.safetensors",
"model.layers.17.mlp.down_proj.weight": "model-00023-of-00042.safetensors",
"model.layers.17.mlp.gate_proj.weight": "model-00023-of-00042.safetensors",
"model.layers.17.mlp.up_proj.weight": "model-00023-of-00042.safetensors",
"model.layers.17.post_attention_layernorm.weight": "model-00023-of-00042.safetensors",
"model.layers.17.self_attn.k_proj.weight": "model-00022-of-00042.safetensors",
"model.layers.17.self_attn.o_proj.weight": "model-00023-of-00042.safetensors",
"model.layers.17.self_attn.q_proj.weight": "model-00022-of-00042.safetensors",
"model.layers.17.self_attn.v_proj.weight": "model-00023-of-00042.safetensors",
"model.layers.18.input_layernorm.weight": "model-00025-of-00042.safetensors",
"model.layers.18.mlp.down_proj.weight": "model-00025-of-00042.safetensors",
"model.layers.18.mlp.gate_proj.weight": "model-00024-of-00042.safetensors",
"model.layers.18.mlp.up_proj.weight": "model-00024-of-00042.safetensors",
"model.layers.18.post_attention_layernorm.weight": "model-00025-of-00042.safetensors",
"model.layers.18.self_attn.k_proj.weight": "model-00024-of-00042.safetensors",
"model.layers.18.self_attn.o_proj.weight": "model-00024-of-00042.safetensors",
"model.layers.18.self_attn.q_proj.weight": "model-00024-of-00042.safetensors",
"model.layers.18.self_attn.v_proj.weight": "model-00024-of-00042.safetensors",
"model.layers.19.input_layernorm.weight": "model-00026-of-00042.safetensors",
"model.layers.19.mlp.down_proj.weight": "model-00026-of-00042.safetensors",
"model.layers.19.mlp.gate_proj.weight": "model-00025-of-00042.safetensors",
"model.layers.19.mlp.up_proj.weight": "model-00026-of-00042.safetensors",
"model.layers.19.post_attention_layernorm.weight": "model-00026-of-00042.safetensors",
"model.layers.19.self_attn.k_proj.weight": "model-00025-of-00042.safetensors",
"model.layers.19.self_attn.o_proj.weight": "model-00025-of-00042.safetensors",
"model.layers.19.self_attn.q_proj.weight": "model-00025-of-00042.safetensors",
"model.layers.19.self_attn.v_proj.weight": "model-00025-of-00042.safetensors",
"model.layers.2.input_layernorm.weight": "model-00005-of-00042.safetensors",
"model.layers.2.mlp.down_proj.weight": "model-00005-of-00042.safetensors",
"model.layers.2.mlp.gate_proj.weight": "model-00004-of-00042.safetensors",
"model.layers.2.mlp.up_proj.weight": "model-00004-of-00042.safetensors",
"model.layers.2.post_attention_layernorm.weight": "model-00005-of-00042.safetensors",
"model.layers.2.self_attn.k_proj.weight": "model-00004-of-00042.safetensors",
"model.layers.2.self_attn.o_proj.weight": "model-00004-of-00042.safetensors",
"model.layers.2.self_attn.q_proj.weight": "model-00004-of-00042.safetensors",
"model.layers.2.self_attn.v_proj.weight": "model-00004-of-00042.safetensors",
"model.layers.20.input_layernorm.weight": "model-00027-of-00042.safetensors",
"model.layers.20.mlp.down_proj.weight": "model-00027-of-00042.safetensors",
"model.layers.20.mlp.gate_proj.weight": "model-00027-of-00042.safetensors",
"model.layers.20.mlp.up_proj.weight": "model-00027-of-00042.safetensors",
"model.layers.20.post_attention_layernorm.weight": "model-00027-of-00042.safetensors",
"model.layers.20.self_attn.k_proj.weight": "model-00026-of-00042.safetensors",
"model.layers.20.self_attn.o_proj.weight": "model-00026-of-00042.safetensors",
"model.layers.20.self_attn.q_proj.weight": "model-00026-of-00042.safetensors",
"model.layers.20.self_attn.v_proj.weight": "model-00026-of-00042.safetensors",
"model.layers.21.input_layernorm.weight": "model-00028-of-00042.safetensors",
"model.layers.21.mlp.down_proj.weight": "model-00028-of-00042.safetensors",
"model.layers.21.mlp.gate_proj.weight": "model-00028-of-00042.safetensors",
"model.layers.21.mlp.up_proj.weight": "model-00028-of-00042.safetensors",
"model.layers.21.post_attention_layernorm.weight": "model-00028-of-00042.safetensors",
"model.layers.21.self_attn.k_proj.weight": "model-00027-of-00042.safetensors",
"model.layers.21.self_attn.o_proj.weight": "model-00028-of-00042.safetensors",
"model.layers.21.self_attn.q_proj.weight": "model-00027-of-00042.safetensors",
"model.layers.21.self_attn.v_proj.weight": "model-00028-of-00042.safetensors",
"model.layers.22.input_layernorm.weight": "model-00030-of-00042.safetensors",
"model.layers.22.mlp.down_proj.weight": "model-00030-of-00042.safetensors",
"model.layers.22.mlp.gate_proj.weight": "model-00029-of-00042.safetensors",
"model.layers.22.mlp.up_proj.weight": "model-00029-of-00042.safetensors",
"model.layers.22.post_attention_layernorm.weight": "model-00030-of-00042.safetensors",
"model.layers.22.self_attn.k_proj.weight": "model-00029-of-00042.safetensors",
"model.layers.22.self_attn.o_proj.weight": "model-00029-of-00042.safetensors",
"model.layers.22.self_attn.q_proj.weight": "model-00029-of-00042.safetensors",
"model.layers.22.self_attn.v_proj.weight": "model-00029-of-00042.safetensors",
"model.layers.23.input_layernorm.weight": "model-00031-of-00042.safetensors",
"model.layers.23.mlp.down_proj.weight": "model-00031-of-00042.safetensors",
"model.layers.23.mlp.gate_proj.weight": "model-00030-of-00042.safetensors",
"model.layers.23.mlp.up_proj.weight": "model-00031-of-00042.safetensors",
"model.layers.23.post_attention_layernorm.weight": "model-00031-of-00042.safetensors",
"model.layers.23.self_attn.k_proj.weight": "model-00030-of-00042.safetensors",
"model.layers.23.self_attn.o_proj.weight": "model-00030-of-00042.safetensors",
"model.layers.23.self_attn.q_proj.weight": "model-00030-of-00042.safetensors",
"model.layers.23.self_attn.v_proj.weight": "model-00030-of-00042.safetensors",
"model.layers.24.input_layernorm.weight": "model-00032-of-00042.safetensors",
"model.layers.24.mlp.down_proj.weight": "model-00032-of-00042.safetensors",
"model.layers.24.mlp.gate_proj.weight": "model-00032-of-00042.safetensors",
"model.layers.24.mlp.up_proj.weight": "model-00032-of-00042.safetensors",
"model.layers.24.post_attention_layernorm.weight": "model-00032-of-00042.safetensors",
"model.layers.24.self_attn.k_proj.weight": "model-00031-of-00042.safetensors",
"model.layers.24.self_attn.o_proj.weight": "model-00031-of-00042.safetensors",
"model.layers.24.self_attn.q_proj.weight": "model-00031-of-00042.safetensors",
"model.layers.24.self_attn.v_proj.weight": "model-00031-of-00042.safetensors",
"model.layers.25.input_layernorm.weight": "model-00033-of-00042.safetensors",
"model.layers.25.mlp.down_proj.weight": "model-00033-of-00042.safetensors",
"model.layers.25.mlp.gate_proj.weight": "model-00033-of-00042.safetensors",
"model.layers.25.mlp.up_proj.weight": "model-00033-of-00042.safetensors",
"model.layers.25.post_attention_layernorm.weight": "model-00033-of-00042.safetensors",
"model.layers.25.self_attn.k_proj.weight": "model-00032-of-00042.safetensors",
"model.layers.25.self_attn.o_proj.weight": "model-00033-of-00042.safetensors",
"model.layers.25.self_attn.q_proj.weight": "model-00032-of-00042.safetensors",
"model.layers.25.self_attn.v_proj.weight": "model-00033-of-00042.safetensors",
"model.layers.26.input_layernorm.weight": "model-00035-of-00042.safetensors",
"model.layers.26.mlp.down_proj.weight": "model-00035-of-00042.safetensors",
"model.layers.26.mlp.gate_proj.weight": "model-00034-of-00042.safetensors",
"model.layers.26.mlp.up_proj.weight": "model-00034-of-00042.safetensors",
"model.layers.26.post_attention_layernorm.weight": "model-00035-of-00042.safetensors",
"model.layers.26.self_attn.k_proj.weight": "model-00034-of-00042.safetensors",
"model.layers.26.self_attn.o_proj.weight": "model-00034-of-00042.safetensors",
"model.layers.26.self_attn.q_proj.weight": "model-00034-of-00042.safetensors",
"model.layers.26.self_attn.v_proj.weight": "model-00034-of-00042.safetensors",
"model.layers.27.input_layernorm.weight": "model-00036-of-00042.safetensors",
"model.layers.27.mlp.down_proj.weight": "model-00036-of-00042.safetensors",
"model.layers.27.mlp.gate_proj.weight": "model-00035-of-00042.safetensors",
"model.layers.27.mlp.up_proj.weight": "model-00036-of-00042.safetensors",
"model.layers.27.post_attention_layernorm.weight": "model-00036-of-00042.safetensors",
"model.layers.27.self_attn.k_proj.weight": "model-00035-of-00042.safetensors",
"model.layers.27.self_attn.o_proj.weight": "model-00035-of-00042.safetensors",
"model.layers.27.self_attn.q_proj.weight": "model-00035-of-00042.safetensors",
"model.layers.27.self_attn.v_proj.weight": "model-00035-of-00042.safetensors",
"model.layers.28.input_layernorm.weight": "model-00037-of-00042.safetensors",
"model.layers.28.mlp.down_proj.weight": "model-00037-of-00042.safetensors",
"model.layers.28.mlp.gate_proj.weight": "model-00037-of-00042.safetensors",
"model.layers.28.mlp.up_proj.weight": "model-00037-of-00042.safetensors",
"model.layers.28.post_attention_layernorm.weight": "model-00037-of-00042.safetensors",
"model.layers.28.self_attn.k_proj.weight": "model-00036-of-00042.safetensors",
"model.layers.28.self_attn.o_proj.weight": "model-00036-of-00042.safetensors",
"model.layers.28.self_attn.q_proj.weight": "model-00036-of-00042.safetensors",
"model.layers.28.self_attn.v_proj.weight": "model-00036-of-00042.safetensors",
"model.layers.29.input_layernorm.weight": "model-00038-of-00042.safetensors",
"model.layers.29.mlp.down_proj.weight": "model-00038-of-00042.safetensors",
"model.layers.29.mlp.gate_proj.weight": "model-00038-of-00042.safetensors",
"model.layers.29.mlp.up_proj.weight": "model-00038-of-00042.safetensors",
"model.layers.29.post_attention_layernorm.weight": "model-00038-of-00042.safetensors",
"model.layers.29.self_attn.k_proj.weight": "model-00037-of-00042.safetensors",
"model.layers.29.self_attn.o_proj.weight": "model-00038-of-00042.safetensors",
"model.layers.29.self_attn.q_proj.weight": "model-00037-of-00042.safetensors",
"model.layers.29.self_attn.v_proj.weight": "model-00038-of-00042.safetensors",
"model.layers.3.input_layernorm.weight": "model-00006-of-00042.safetensors",
"model.layers.3.mlp.down_proj.weight": "model-00006-of-00042.safetensors",
"model.layers.3.mlp.gate_proj.weight": "model-00005-of-00042.safetensors",
"model.layers.3.mlp.up_proj.weight": "model-00006-of-00042.safetensors",
"model.layers.3.post_attention_layernorm.weight": "model-00006-of-00042.safetensors",
"model.layers.3.self_attn.k_proj.weight": "model-00005-of-00042.safetensors",
"model.layers.3.self_attn.o_proj.weight": "model-00005-of-00042.safetensors",
"model.layers.3.self_attn.q_proj.weight": "model-00005-of-00042.safetensors",
"model.layers.3.self_attn.v_proj.weight": "model-00005-of-00042.safetensors",
"model.layers.30.input_layernorm.weight": "model-00040-of-00042.safetensors",
"model.layers.30.mlp.down_proj.weight": "model-00040-of-00042.safetensors",
"model.layers.30.mlp.gate_proj.weight": "model-00039-of-00042.safetensors",
"model.layers.30.mlp.up_proj.weight": "model-00039-of-00042.safetensors",
"model.layers.30.post_attention_layernorm.weight": "model-00040-of-00042.safetensors",
"model.layers.30.self_attn.k_proj.weight": "model-00039-of-00042.safetensors",
"model.layers.30.self_attn.o_proj.weight": "model-00039-of-00042.safetensors",
"model.layers.30.self_attn.q_proj.weight": "model-00039-of-00042.safetensors",
"model.layers.30.self_attn.v_proj.weight": "model-00039-of-00042.safetensors",
"model.layers.31.input_layernorm.weight": "model-00041-of-00042.safetensors",
"model.layers.31.mlp.down_proj.weight": "model-00041-of-00042.safetensors",
"model.layers.31.mlp.gate_proj.weight": "model-00040-of-00042.safetensors",
"model.layers.31.mlp.up_proj.weight": "model-00041-of-00042.safetensors",
"model.layers.31.post_attention_layernorm.weight": "model-00041-of-00042.safetensors",
"model.layers.31.self_attn.k_proj.weight": "model-00040-of-00042.safetensors",
"model.layers.31.self_attn.o_proj.weight": "model-00040-of-00042.safetensors",
"model.layers.31.self_attn.q_proj.weight": "model-00040-of-00042.safetensors",
"model.layers.31.self_attn.v_proj.weight": "model-00040-of-00042.safetensors",
"model.layers.4.input_layernorm.weight": "model-00007-of-00042.safetensors",
"model.layers.4.mlp.down_proj.weight": "model-00007-of-00042.safetensors",
"model.layers.4.mlp.gate_proj.weight": "model-00007-of-00042.safetensors",
"model.layers.4.mlp.up_proj.weight": "model-00007-of-00042.safetensors",
"model.layers.4.post_attention_layernorm.weight": "model-00007-of-00042.safetensors",
"model.layers.4.self_attn.k_proj.weight": "model-00006-of-00042.safetensors",
"model.layers.4.self_attn.o_proj.weight": "model-00006-of-00042.safetensors",
"model.layers.4.self_attn.q_proj.weight": "model-00006-of-00042.safetensors",
"model.layers.4.self_attn.v_proj.weight": "model-00006-of-00042.safetensors",
"model.layers.5.input_layernorm.weight": "model-00008-of-00042.safetensors",
"model.layers.5.mlp.down_proj.weight": "model-00008-of-00042.safetensors",
"model.layers.5.mlp.gate_proj.weight": "model-00008-of-00042.safetensors",
"model.layers.5.mlp.up_proj.weight": "model-00008-of-00042.safetensors",
"model.layers.5.post_attention_layernorm.weight": "model-00008-of-00042.safetensors",
"model.layers.5.self_attn.k_proj.weight": "model-00007-of-00042.safetensors",
"model.layers.5.self_attn.o_proj.weight": "model-00008-of-00042.safetensors",
"model.layers.5.self_attn.q_proj.weight": "model-00007-of-00042.safetensors",
"model.layers.5.self_attn.v_proj.weight": "model-00008-of-00042.safetensors",
"model.layers.6.input_layernorm.weight": "model-00010-of-00042.safetensors",
"model.layers.6.mlp.down_proj.weight": "model-00010-of-00042.safetensors",
"model.layers.6.mlp.gate_proj.weight": "model-00009-of-00042.safetensors",
"model.layers.6.mlp.up_proj.weight": "model-00009-of-00042.safetensors",
"model.layers.6.post_attention_layernorm.weight": "model-00010-of-00042.safetensors",
"model.layers.6.self_attn.k_proj.weight": "model-00009-of-00042.safetensors",
"model.layers.6.self_attn.o_proj.weight": "model-00009-of-00042.safetensors",
"model.layers.6.self_attn.q_proj.weight": "model-00009-of-00042.safetensors",
"model.layers.6.self_attn.v_proj.weight": "model-00009-of-00042.safetensors",
"model.layers.7.input_layernorm.weight": "model-00011-of-00042.safetensors",
"model.layers.7.mlp.down_proj.weight": "model-00011-of-00042.safetensors",
"model.layers.7.mlp.gate_proj.weight": "model-00010-of-00042.safetensors",
"model.layers.7.mlp.up_proj.weight": "model-00011-of-00042.safetensors",
"model.layers.7.post_attention_layernorm.weight": "model-00011-of-00042.safetensors",
"model.layers.7.self_attn.k_proj.weight": "model-00010-of-00042.safetensors",
"model.layers.7.self_attn.o_proj.weight": "model-00010-of-00042.safetensors",
"model.layers.7.self_attn.q_proj.weight": "model-00010-of-00042.safetensors",
"model.layers.7.self_attn.v_proj.weight": "model-00010-of-00042.safetensors",
"model.layers.8.input_layernorm.weight": "model-00012-of-00042.safetensors",
"model.layers.8.mlp.down_proj.weight": "model-00012-of-00042.safetensors",
"model.layers.8.mlp.gate_proj.weight": "model-00012-of-00042.safetensors",
"model.layers.8.mlp.up_proj.weight": "model-00012-of-00042.safetensors",
"model.layers.8.post_attention_layernorm.weight": "model-00012-of-00042.safetensors",
"model.layers.8.self_attn.k_proj.weight": "model-00011-of-00042.safetensors",
"model.layers.8.self_attn.o_proj.weight": "model-00011-of-00042.safetensors",
"model.layers.8.self_attn.q_proj.weight": "model-00011-of-00042.safetensors",
"model.layers.8.self_attn.v_proj.weight": "model-00011-of-00042.safetensors",
"model.layers.9.input_layernorm.weight": "model-00013-of-00042.safetensors",
"model.layers.9.mlp.down_proj.weight": "model-00013-of-00042.safetensors",
"model.layers.9.mlp.gate_proj.weight": "model-00013-of-00042.safetensors",
"model.layers.9.mlp.up_proj.weight": "model-00013-of-00042.safetensors",
"model.layers.9.post_attention_layernorm.weight": "model-00013-of-00042.safetensors",
"model.layers.9.self_attn.k_proj.weight": "model-00012-of-00042.safetensors",
"model.layers.9.self_attn.o_proj.weight": "model-00013-of-00042.safetensors",
"model.layers.9.self_attn.q_proj.weight": "model-00012-of-00042.safetensors",
"model.layers.9.self_attn.v_proj.weight": "model-00013-of-00042.safetensors",
"model.norm.weight": "model-00041-of-00042.safetensors"
}
}

3
pytorch_model-00001-of-00042.bin Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:edb01660ef4916e36a696565de9ec529c4701924110b0ab094c5734f5b5ffeb1
size 346032713

3
pytorch_model-00002-of-00042.bin Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:8d0d6e54fa699918fde76080c8a02a67a2c4aa15806b899f43e2723d43c95279
size 394284267

3
pytorch_model-00003-of-00042.bin Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:249b69253fe26489a12283ddcabea151cce789452bf1611d7273d6bcb38b9e93
size 394284267

3
pytorch_model-00004-of-00042.bin Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:bfcc388e55b3b58e44d15a57eba94e965ce458f875de5bd3b5b6dc5a331a800c
size 318770074

3
pytorch_model-00005-of-00042.bin Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:a02b6e482f45ff4bb1be3a71d4b8a230d9a75023d2e2839612c337d68966e586
size 318787068

3
pytorch_model-00006-of-00042.bin Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:d3b51300a12087cd7f62e4778de0525bb847bfaae48f954cd0fe5f79caa137d7
size 318787068

3
pytorch_model-00007-of-00042.bin Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:b5fc11db3dd9485b36602a8521d976d841d971a99ea231390d9fad2555fb0fb0
size 394284267

3
pytorch_model-00008-of-00042.bin Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:c55baa15ee7dbc5936d4556051c64ba5f1b349ea2c4d0d6b87e2f65125972dab
size 394284267

3
pytorch_model-00009-of-00042.bin Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:ae9afe708e7a6dde5f57089f9a88f435772b637892db56c2eeecaf5a49354b74
size 318770074

3
pytorch_model-00010-of-00042.bin Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:5f13884078f1cb4b917907e21b6d9bd674f5a9f624e31ec608c60187cb39b5ca
size 318787068

3
pytorch_model-00011-of-00042.bin Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:630d13f57dee95e9b7cf023369ff0fe152ea878b9e301f5d6ac10363b9e48c67
size 318787068

3
pytorch_model-00012-of-00042.bin Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:856ac3613f7ad070c11794fd47a1e772107f43afb3758b14e120bd8637a76712
size 394284267

3
pytorch_model-00013-of-00042.bin Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:6d41ddb47f347871bae5cfd750576c8dce01eaba6136d95c83937e9a05cf23c4
size 394284267

3
pytorch_model-00014-of-00042.bin Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:e2108878bac09ca9face0468aae6ebfbc9c043fd3723e4244e647ca69a9bbb93
size 318770074

3
pytorch_model-00015-of-00042.bin Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:b9561acefed07cd764476ccbf1355491647360a69bb294703cba44ed3a3a4cbf
size 318787068

3
pytorch_model-00016-of-00042.bin Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:ec102bc2116c3f9f9b114311f9bcf033e56921e359ee4cd0d3c082f89574b4fb
size 318787068

3
pytorch_model-00017-of-00042.bin Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:0a814544849be2516f28275c96d373ee2e45d546b685908cdf48e353808df96e
size 394284267

3
pytorch_model-00018-of-00042.bin Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:adca7e78aaef8818724e88dc4ab562aedf83fa4cc527cc853f7e4376211c07f4
size 394284267

3
pytorch_model-00019-of-00042.bin Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:126357d3c3b549e448ed4a2bea4d9b5c46fcc4fdfc00b738710605373c20a37a
size 318770074

3
pytorch_model-00020-of-00042.bin Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:72b7dca1d2750a2c1eb1be37894e42621b99f46d0717b3a3335aab1c0f602339
size 318787068

3
pytorch_model-00021-of-00042.bin Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:17163a60e8ea05dd4aeec038dcf641b69c2009064f3baddbf17b512eb354508a
size 318787068

3
pytorch_model-00022-of-00042.bin Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:6dbfd36a417bd830e50b5fc7f937ab191672b95bad74dd230db63abf8d43802a
size 394284267

3
pytorch_model-00023-of-00042.bin Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:4300a5f62846d21cfce3147ebbc4cf6e3a3f6ad189a1b5eff07cbec6f035e661
size 394284267

3
pytorch_model-00024-of-00042.bin Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:4b1aa078d8eccfd622354258ded773eb4f9684a20588cf05dff37d656c03ec05
size 318770074

3
pytorch_model-00025-of-00042.bin Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:fa80e6c346d64df7115dc5709537d6e0c69174afda6ad1b4d7b15e099e65a4fb
size 318787068

3
pytorch_model-00026-of-00042.bin Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:c8c16ae149337d2c7dd54f96f54c2beffba6cef5d728fdc3432f935fe636d42c
size 318787068

3
pytorch_model-00027-of-00042.bin Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:522894f4f003687daf901f7f8e38147357dc5686d7751811e99f0d25724b43c3
size 394284267

3
pytorch_model-00028-of-00042.bin Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:3f8ffb7150a9895ebe23b6881f40623adfa8463522aec671230add89f0e74fc7
size 394284267

3
pytorch_model-00029-of-00042.bin Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:81cd2ec4bd563235298d38201664be2ea88b2f1b2a9ff8cd0c4a45c077af7b1a
size 318770074

3
pytorch_model-00030-of-00042.bin Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:76be51be2bcf7c972566b65609591a65d82489c2ede405eca2aa56e74dd19bf1
size 318787068

3
pytorch_model-00031-of-00042.bin Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:dc3063ebffc49c3abce3283a50dfdb2bdcb7ac07a3bd1019416eaba4f4cbaee4
size 318787068

3
pytorch_model-00032-of-00042.bin Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:ec236fcbf407315dfa4d4939e40f9e2e0909cd3610bdb01fe89dea82612b12af
size 394284267

3
pytorch_model-00033-of-00042.bin Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:2786abfbbc995832feb308855300c51c001f17f189f49188750baca152093edb
size 394284267

3
pytorch_model-00034-of-00042.bin Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:3e123964eaa9692dbfbb7af16ab8d39c8647fe7467e02767dc299357c8c7c088
size 318770074

3
pytorch_model-00035-of-00042.bin Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:3ba597adb38c9e2697f61ef8abe17e73ef4f85152c275c17ba9ee94851adc530
size 318787068

3
pytorch_model-00036-of-00042.bin Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:e76abc8b6dedd4c2efe3a4cfa8b8226d98b892660e5100349beb6c14cdd10a51
size 318787068

3
pytorch_model-00037-of-00042.bin Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:09b97b10a7a55b20a9a189cc84056b6576511dd93a942b1ac4a91915b59b2bc1
size 394284267

3
pytorch_model-00038-of-00042.bin Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:c541f2ba420a54c40d61a238395ef80f32468cd93d723becb4d18cf352ef206a
size 394284267

3
pytorch_model-00039-of-00042.bin Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:01b181d65e7b1f3623258a7a77e85dec66a39713f0ab0c24fa25485366a910a2
size 318770074

3
pytorch_model-00040-of-00042.bin Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:f81fcf33655b841eed3b21373f0974976357409db46eb857370b8809ef3a8cf7
size 318787068

3
pytorch_model-00041-of-00042.bin Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:6d465ba911b93ac63ef293378820ea5338b6f42337e98820df8ebd48b881866c
size 234908233

3
pytorch_model-00042-of-00042.bin Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:8d3745cd4d751227cc3179b3f06c0b84d80bb2a1436c638c854a4bfbc0712666
size 262145413

298
pytorch_model.bin.index.json Normal file
View File

@@ -0,0 +1,298 @@
{
"metadata": {
"total_size": 14483464192
},
"weight_map": {
"lm_head.weight": "pytorch_model-00042-of-00042.bin",
"model.embed_tokens.weight": "pytorch_model-00001-of-00042.bin",
"model.layers.0.input_layernorm.weight": "pytorch_model-00002-of-00042.bin",
"model.layers.0.mlp.down_proj.weight": "pytorch_model-00002-of-00042.bin",
"model.layers.0.mlp.gate_proj.weight": "pytorch_model-00002-of-00042.bin",
"model.layers.0.mlp.up_proj.weight": "pytorch_model-00002-of-00042.bin",
"model.layers.0.post_attention_layernorm.weight": "pytorch_model-00002-of-00042.bin",
"model.layers.0.self_attn.k_proj.weight": "pytorch_model-00001-of-00042.bin",
"model.layers.0.self_attn.o_proj.weight": "pytorch_model-00001-of-00042.bin",
"model.layers.0.self_attn.q_proj.weight": "pytorch_model-00001-of-00042.bin",
"model.layers.0.self_attn.v_proj.weight": "pytorch_model-00001-of-00042.bin",
"model.layers.1.input_layernorm.weight": "pytorch_model-00003-of-00042.bin",
"model.layers.1.mlp.down_proj.weight": "pytorch_model-00003-of-00042.bin",
"model.layers.1.mlp.gate_proj.weight": "pytorch_model-00003-of-00042.bin",
"model.layers.1.mlp.up_proj.weight": "pytorch_model-00003-of-00042.bin",
"model.layers.1.post_attention_layernorm.weight": "pytorch_model-00003-of-00042.bin",
"model.layers.1.self_attn.k_proj.weight": "pytorch_model-00002-of-00042.bin",
"model.layers.1.self_attn.o_proj.weight": "pytorch_model-00003-of-00042.bin",
"model.layers.1.self_attn.q_proj.weight": "pytorch_model-00002-of-00042.bin",
"model.layers.1.self_attn.v_proj.weight": "pytorch_model-00003-of-00042.bin",
"model.layers.10.input_layernorm.weight": "pytorch_model-00015-of-00042.bin",
"model.layers.10.mlp.down_proj.weight": "pytorch_model-00015-of-00042.bin",
"model.layers.10.mlp.gate_proj.weight": "pytorch_model-00014-of-00042.bin",
"model.layers.10.mlp.up_proj.weight": "pytorch_model-00014-of-00042.bin",
"model.layers.10.post_attention_layernorm.weight": "pytorch_model-00015-of-00042.bin",
"model.layers.10.self_attn.k_proj.weight": "pytorch_model-00014-of-00042.bin",
"model.layers.10.self_attn.o_proj.weight": "pytorch_model-00014-of-00042.bin",
"model.layers.10.self_attn.q_proj.weight": "pytorch_model-00014-of-00042.bin",
"model.layers.10.self_attn.v_proj.weight": "pytorch_model-00014-of-00042.bin",
"model.layers.11.input_layernorm.weight": "pytorch_model-00016-of-00042.bin",
"model.layers.11.mlp.down_proj.weight": "pytorch_model-00016-of-00042.bin",
"model.layers.11.mlp.gate_proj.weight": "pytorch_model-00015-of-00042.bin",
"model.layers.11.mlp.up_proj.weight": "pytorch_model-00016-of-00042.bin",
"model.layers.11.post_attention_layernorm.weight": "pytorch_model-00016-of-00042.bin",
"model.layers.11.self_attn.k_proj.weight": "pytorch_model-00015-of-00042.bin",
"model.layers.11.self_attn.o_proj.weight": "pytorch_model-00015-of-00042.bin",
"model.layers.11.self_attn.q_proj.weight": "pytorch_model-00015-of-00042.bin",
"model.layers.11.self_attn.v_proj.weight": "pytorch_model-00015-of-00042.bin",
"model.layers.12.input_layernorm.weight": "pytorch_model-00017-of-00042.bin",
"model.layers.12.mlp.down_proj.weight": "pytorch_model-00017-of-00042.bin",
"model.layers.12.mlp.gate_proj.weight": "pytorch_model-00017-of-00042.bin",
"model.layers.12.mlp.up_proj.weight": "pytorch_model-00017-of-00042.bin",
"model.layers.12.post_attention_layernorm.weight": "pytorch_model-00017-of-00042.bin",
"model.layers.12.self_attn.k_proj.weight": "pytorch_model-00016-of-00042.bin",
"model.layers.12.self_attn.o_proj.weight": "pytorch_model-00016-of-00042.bin",
"model.layers.12.self_attn.q_proj.weight": "pytorch_model-00016-of-00042.bin",
"model.layers.12.self_attn.v_proj.weight": "pytorch_model-00016-of-00042.bin",
"model.layers.13.input_layernorm.weight": "pytorch_model-00018-of-00042.bin",
"model.layers.13.mlp.down_proj.weight": "pytorch_model-00018-of-00042.bin",
"model.layers.13.mlp.gate_proj.weight": "pytorch_model-00018-of-00042.bin",
"model.layers.13.mlp.up_proj.weight": "pytorch_model-00018-of-00042.bin",
"model.layers.13.post_attention_layernorm.weight": "pytorch_model-00018-of-00042.bin",
"model.layers.13.self_attn.k_proj.weight": "pytorch_model-00017-of-00042.bin",
"model.layers.13.self_attn.o_proj.weight": "pytorch_model-00018-of-00042.bin",
"model.layers.13.self_attn.q_proj.weight": "pytorch_model-00017-of-00042.bin",
"model.layers.13.self_attn.v_proj.weight": "pytorch_model-00018-of-00042.bin",
"model.layers.14.input_layernorm.weight": "pytorch_model-00020-of-00042.bin",
"model.layers.14.mlp.down_proj.weight": "pytorch_model-00020-of-00042.bin",
"model.layers.14.mlp.gate_proj.weight": "pytorch_model-00019-of-00042.bin",
"model.layers.14.mlp.up_proj.weight": "pytorch_model-00019-of-00042.bin",
"model.layers.14.post_attention_layernorm.weight": "pytorch_model-00020-of-00042.bin",
"model.layers.14.self_attn.k_proj.weight": "pytorch_model-00019-of-00042.bin",
"model.layers.14.self_attn.o_proj.weight": "pytorch_model-00019-of-00042.bin",
"model.layers.14.self_attn.q_proj.weight": "pytorch_model-00019-of-00042.bin",
"model.layers.14.self_attn.v_proj.weight": "pytorch_model-00019-of-00042.bin",
"model.layers.15.input_layernorm.weight": "pytorch_model-00021-of-00042.bin",
"model.layers.15.mlp.down_proj.weight": "pytorch_model-00021-of-00042.bin",
"model.layers.15.mlp.gate_proj.weight": "pytorch_model-00020-of-00042.bin",
"model.layers.15.mlp.up_proj.weight": "pytorch_model-00021-of-00042.bin",
"model.layers.15.post_attention_layernorm.weight": "pytorch_model-00021-of-00042.bin",
"model.layers.15.self_attn.k_proj.weight": "pytorch_model-00020-of-00042.bin",
"model.layers.15.self_attn.o_proj.weight": "pytorch_model-00020-of-00042.bin",
"model.layers.15.self_attn.q_proj.weight": "pytorch_model-00020-of-00042.bin",
"model.layers.15.self_attn.v_proj.weight": "pytorch_model-00020-of-00042.bin",
"model.layers.16.input_layernorm.weight": "pytorch_model-00022-of-00042.bin",
"model.layers.16.mlp.down_proj.weight": "pytorch_model-00022-of-00042.bin",
"model.layers.16.mlp.gate_proj.weight": "pytorch_model-00022-of-00042.bin",
"model.layers.16.mlp.up_proj.weight": "pytorch_model-00022-of-00042.bin",
"model.layers.16.post_attention_layernorm.weight": "pytorch_model-00022-of-00042.bin",
"model.layers.16.self_attn.k_proj.weight": "pytorch_model-00021-of-00042.bin",
"model.layers.16.self_attn.o_proj.weight": "pytorch_model-00021-of-00042.bin",
"model.layers.16.self_attn.q_proj.weight": "pytorch_model-00021-of-00042.bin",
"model.layers.16.self_attn.v_proj.weight": "pytorch_model-00021-of-00042.bin",
"model.layers.17.input_layernorm.weight": "pytorch_model-00023-of-00042.bin",
"model.layers.17.mlp.down_proj.weight": "pytorch_model-00023-of-00042.bin",
"model.layers.17.mlp.gate_proj.weight": "pytorch_model-00023-of-00042.bin",
"model.layers.17.mlp.up_proj.weight": "pytorch_model-00023-of-00042.bin",
"model.layers.17.post_attention_layernorm.weight": "pytorch_model-00023-of-00042.bin",
"model.layers.17.self_attn.k_proj.weight": "pytorch_model-00022-of-00042.bin",
"model.layers.17.self_attn.o_proj.weight": "pytorch_model-00023-of-00042.bin",
"model.layers.17.self_attn.q_proj.weight": "pytorch_model-00022-of-00042.bin",
"model.layers.17.self_attn.v_proj.weight": "pytorch_model-00023-of-00042.bin",
"model.layers.18.input_layernorm.weight": "pytorch_model-00025-of-00042.bin",
"model.layers.18.mlp.down_proj.weight": "pytorch_model-00025-of-00042.bin",
"model.layers.18.mlp.gate_proj.weight": "pytorch_model-00024-of-00042.bin",
"model.layers.18.mlp.up_proj.weight": "pytorch_model-00024-of-00042.bin",
"model.layers.18.post_attention_layernorm.weight": "pytorch_model-00025-of-00042.bin",
"model.layers.18.self_attn.k_proj.weight": "pytorch_model-00024-of-00042.bin",
"model.layers.18.self_attn.o_proj.weight": "pytorch_model-00024-of-00042.bin",
"model.layers.18.self_attn.q_proj.weight": "pytorch_model-00024-of-00042.bin",
"model.layers.18.self_attn.v_proj.weight": "pytorch_model-00024-of-00042.bin",
"model.layers.19.input_layernorm.weight": "pytorch_model-00026-of-00042.bin",
"model.layers.19.mlp.down_proj.weight": "pytorch_model-00026-of-00042.bin",
"model.layers.19.mlp.gate_proj.weight": "pytorch_model-00025-of-00042.bin",
"model.layers.19.mlp.up_proj.weight": "pytorch_model-00026-of-00042.bin",
"model.layers.19.post_attention_layernorm.weight": "pytorch_model-00026-of-00042.bin",
"model.layers.19.self_attn.k_proj.weight": "pytorch_model-00025-of-00042.bin",
"model.layers.19.self_attn.o_proj.weight": "pytorch_model-00025-of-00042.bin",
"model.layers.19.self_attn.q_proj.weight": "pytorch_model-00025-of-00042.bin",
"model.layers.19.self_attn.v_proj.weight": "pytorch_model-00025-of-00042.bin",
"model.layers.2.input_layernorm.weight": "pytorch_model-00005-of-00042.bin",
"model.layers.2.mlp.down_proj.weight": "pytorch_model-00005-of-00042.bin",
"model.layers.2.mlp.gate_proj.weight": "pytorch_model-00004-of-00042.bin",
"model.layers.2.mlp.up_proj.weight": "pytorch_model-00004-of-00042.bin",
"model.layers.2.post_attention_layernorm.weight": "pytorch_model-00005-of-00042.bin",
"model.layers.2.self_attn.k_proj.weight": "pytorch_model-00004-of-00042.bin",
"model.layers.2.self_attn.o_proj.weight": "pytorch_model-00004-of-00042.bin",
"model.layers.2.self_attn.q_proj.weight": "pytorch_model-00004-of-00042.bin",
"model.layers.2.self_attn.v_proj.weight": "pytorch_model-00004-of-00042.bin",
"model.layers.20.input_layernorm.weight": "pytorch_model-00027-of-00042.bin",
"model.layers.20.mlp.down_proj.weight": "pytorch_model-00027-of-00042.bin",
"model.layers.20.mlp.gate_proj.weight": "pytorch_model-00027-of-00042.bin",
"model.layers.20.mlp.up_proj.weight": "pytorch_model-00027-of-00042.bin",
"model.layers.20.post_attention_layernorm.weight": "pytorch_model-00027-of-00042.bin",
"model.layers.20.self_attn.k_proj.weight": "pytorch_model-00026-of-00042.bin",
"model.layers.20.self_attn.o_proj.weight": "pytorch_model-00026-of-00042.bin",
"model.layers.20.self_attn.q_proj.weight": "pytorch_model-00026-of-00042.bin",
"model.layers.20.self_attn.v_proj.weight": "pytorch_model-00026-of-00042.bin",
"model.layers.21.input_layernorm.weight": "pytorch_model-00028-of-00042.bin",
"model.layers.21.mlp.down_proj.weight": "pytorch_model-00028-of-00042.bin",
"model.layers.21.mlp.gate_proj.weight": "pytorch_model-00028-of-00042.bin",
"model.layers.21.mlp.up_proj.weight": "pytorch_model-00028-of-00042.bin",
"model.layers.21.post_attention_layernorm.weight": "pytorch_model-00028-of-00042.bin",
"model.layers.21.self_attn.k_proj.weight": "pytorch_model-00027-of-00042.bin",
"model.layers.21.self_attn.o_proj.weight": "pytorch_model-00028-of-00042.bin",
"model.layers.21.self_attn.q_proj.weight": "pytorch_model-00027-of-00042.bin",
"model.layers.21.self_attn.v_proj.weight": "pytorch_model-00028-of-00042.bin",
"model.layers.22.input_layernorm.weight": "pytorch_model-00030-of-00042.bin",
"model.layers.22.mlp.down_proj.weight": "pytorch_model-00030-of-00042.bin",
"model.layers.22.mlp.gate_proj.weight": "pytorch_model-00029-of-00042.bin",
"model.layers.22.mlp.up_proj.weight": "pytorch_model-00029-of-00042.bin",
"model.layers.22.post_attention_layernorm.weight": "pytorch_model-00030-of-00042.bin",
"model.layers.22.self_attn.k_proj.weight": "pytorch_model-00029-of-00042.bin",
"model.layers.22.self_attn.o_proj.weight": "pytorch_model-00029-of-00042.bin",
"model.layers.22.self_attn.q_proj.weight": "pytorch_model-00029-of-00042.bin",
"model.layers.22.self_attn.v_proj.weight": "pytorch_model-00029-of-00042.bin",
"model.layers.23.input_layernorm.weight": "pytorch_model-00031-of-00042.bin",
"model.layers.23.mlp.down_proj.weight": "pytorch_model-00031-of-00042.bin",
"model.layers.23.mlp.gate_proj.weight": "pytorch_model-00030-of-00042.bin",
"model.layers.23.mlp.up_proj.weight": "pytorch_model-00031-of-00042.bin",
"model.layers.23.post_attention_layernorm.weight": "pytorch_model-00031-of-00042.bin",
"model.layers.23.self_attn.k_proj.weight": "pytorch_model-00030-of-00042.bin",
"model.layers.23.self_attn.o_proj.weight": "pytorch_model-00030-of-00042.bin",
"model.layers.23.self_attn.q_proj.weight": "pytorch_model-00030-of-00042.bin",
"model.layers.23.self_attn.v_proj.weight": "pytorch_model-00030-of-00042.bin",
"model.layers.24.input_layernorm.weight": "pytorch_model-00032-of-00042.bin",
"model.layers.24.mlp.down_proj.weight": "pytorch_model-00032-of-00042.bin",
"model.layers.24.mlp.gate_proj.weight": "pytorch_model-00032-of-00042.bin",
"model.layers.24.mlp.up_proj.weight": "pytorch_model-00032-of-00042.bin",
"model.layers.24.post_attention_layernorm.weight": "pytorch_model-00032-of-00042.bin",
"model.layers.24.self_attn.k_proj.weight": "pytorch_model-00031-of-00042.bin",
"model.layers.24.self_attn.o_proj.weight": "pytorch_model-00031-of-00042.bin",
"model.layers.24.self_attn.q_proj.weight": "pytorch_model-00031-of-00042.bin",
"model.layers.24.self_attn.v_proj.weight": "pytorch_model-00031-of-00042.bin",
"model.layers.25.input_layernorm.weight": "pytorch_model-00033-of-00042.bin",
"model.layers.25.mlp.down_proj.weight": "pytorch_model-00033-of-00042.bin",
"model.layers.25.mlp.gate_proj.weight": "pytorch_model-00033-of-00042.bin",
"model.layers.25.mlp.up_proj.weight": "pytorch_model-00033-of-00042.bin",
"model.layers.25.post_attention_layernorm.weight": "pytorch_model-00033-of-00042.bin",
"model.layers.25.self_attn.k_proj.weight": "pytorch_model-00032-of-00042.bin",
"model.layers.25.self_attn.o_proj.weight": "pytorch_model-00033-of-00042.bin",
"model.layers.25.self_attn.q_proj.weight": "pytorch_model-00032-of-00042.bin",
"model.layers.25.self_attn.v_proj.weight": "pytorch_model-00033-of-00042.bin",
"model.layers.26.input_layernorm.weight": "pytorch_model-00035-of-00042.bin",
"model.layers.26.mlp.down_proj.weight": "pytorch_model-00035-of-00042.bin",
"model.layers.26.mlp.gate_proj.weight": "pytorch_model-00034-of-00042.bin",
"model.layers.26.mlp.up_proj.weight": "pytorch_model-00034-of-00042.bin",
"model.layers.26.post_attention_layernorm.weight": "pytorch_model-00035-of-00042.bin",
"model.layers.26.self_attn.k_proj.weight": "pytorch_model-00034-of-00042.bin",
"model.layers.26.self_attn.o_proj.weight": "pytorch_model-00034-of-00042.bin",
"model.layers.26.self_attn.q_proj.weight": "pytorch_model-00034-of-00042.bin",
"model.layers.26.self_attn.v_proj.weight": "pytorch_model-00034-of-00042.bin",
"model.layers.27.input_layernorm.weight": "pytorch_model-00036-of-00042.bin",
"model.layers.27.mlp.down_proj.weight": "pytorch_model-00036-of-00042.bin",
"model.layers.27.mlp.gate_proj.weight": "pytorch_model-00035-of-00042.bin",
"model.layers.27.mlp.up_proj.weight": "pytorch_model-00036-of-00042.bin",
"model.layers.27.post_attention_layernorm.weight": "pytorch_model-00036-of-00042.bin",
"model.layers.27.self_attn.k_proj.weight": "pytorch_model-00035-of-00042.bin",
"model.layers.27.self_attn.o_proj.weight": "pytorch_model-00035-of-00042.bin",
"model.layers.27.self_attn.q_proj.weight": "pytorch_model-00035-of-00042.bin",
"model.layers.27.self_attn.v_proj.weight": "pytorch_model-00035-of-00042.bin",
"model.layers.28.input_layernorm.weight": "pytorch_model-00037-of-00042.bin",
"model.layers.28.mlp.down_proj.weight": "pytorch_model-00037-of-00042.bin",
"model.layers.28.mlp.gate_proj.weight": "pytorch_model-00037-of-00042.bin",
"model.layers.28.mlp.up_proj.weight": "pytorch_model-00037-of-00042.bin",
"model.layers.28.post_attention_layernorm.weight": "pytorch_model-00037-of-00042.bin",
"model.layers.28.self_attn.k_proj.weight": "pytorch_model-00036-of-00042.bin",
"model.layers.28.self_attn.o_proj.weight": "pytorch_model-00036-of-00042.bin",
"model.layers.28.self_attn.q_proj.weight": "pytorch_model-00036-of-00042.bin",
"model.layers.28.self_attn.v_proj.weight": "pytorch_model-00036-of-00042.bin",
"model.layers.29.input_layernorm.weight": "pytorch_model-00038-of-00042.bin",
"model.layers.29.mlp.down_proj.weight": "pytorch_model-00038-of-00042.bin",
"model.layers.29.mlp.gate_proj.weight": "pytorch_model-00038-of-00042.bin",
"model.layers.29.mlp.up_proj.weight": "pytorch_model-00038-of-00042.bin",
"model.layers.29.post_attention_layernorm.weight": "pytorch_model-00038-of-00042.bin",
"model.layers.29.self_attn.k_proj.weight": "pytorch_model-00037-of-00042.bin",
"model.layers.29.self_attn.o_proj.weight": "pytorch_model-00038-of-00042.bin",
"model.layers.29.self_attn.q_proj.weight": "pytorch_model-00037-of-00042.bin",
"model.layers.29.self_attn.v_proj.weight": "pytorch_model-00038-of-00042.bin",
"model.layers.3.input_layernorm.weight": "pytorch_model-00006-of-00042.bin",
"model.layers.3.mlp.down_proj.weight": "pytorch_model-00006-of-00042.bin",
"model.layers.3.mlp.gate_proj.weight": "pytorch_model-00005-of-00042.bin",
"model.layers.3.mlp.up_proj.weight": "pytorch_model-00006-of-00042.bin",
"model.layers.3.post_attention_layernorm.weight": "pytorch_model-00006-of-00042.bin",
"model.layers.3.self_attn.k_proj.weight": "pytorch_model-00005-of-00042.bin",
"model.layers.3.self_attn.o_proj.weight": "pytorch_model-00005-of-00042.bin",
"model.layers.3.self_attn.q_proj.weight": "pytorch_model-00005-of-00042.bin",
"model.layers.3.self_attn.v_proj.weight": "pytorch_model-00005-of-00042.bin",
"model.layers.30.input_layernorm.weight": "pytorch_model-00040-of-00042.bin",
"model.layers.30.mlp.down_proj.weight": "pytorch_model-00040-of-00042.bin",
"model.layers.30.mlp.gate_proj.weight": "pytorch_model-00039-of-00042.bin",
"model.layers.30.mlp.up_proj.weight": "pytorch_model-00039-of-00042.bin",
"model.layers.30.post_attention_layernorm.weight": "pytorch_model-00040-of-00042.bin",
"model.layers.30.self_attn.k_proj.weight": "pytorch_model-00039-of-00042.bin",
"model.layers.30.self_attn.o_proj.weight": "pytorch_model-00039-of-00042.bin",
"model.layers.30.self_attn.q_proj.weight": "pytorch_model-00039-of-00042.bin",
"model.layers.30.self_attn.v_proj.weight": "pytorch_model-00039-of-00042.bin",
"model.layers.31.input_layernorm.weight": "pytorch_model-00041-of-00042.bin",
"model.layers.31.mlp.down_proj.weight": "pytorch_model-00041-of-00042.bin",
"model.layers.31.mlp.gate_proj.weight": "pytorch_model-00040-of-00042.bin",
"model.layers.31.mlp.up_proj.weight": "pytorch_model-00041-of-00042.bin",
"model.layers.31.post_attention_layernorm.weight": "pytorch_model-00041-of-00042.bin",
"model.layers.31.self_attn.k_proj.weight": "pytorch_model-00040-of-00042.bin",
"model.layers.31.self_attn.o_proj.weight": "pytorch_model-00040-of-00042.bin",
"model.layers.31.self_attn.q_proj.weight": "pytorch_model-00040-of-00042.bin",
"model.layers.31.self_attn.v_proj.weight": "pytorch_model-00040-of-00042.bin",
"model.layers.4.input_layernorm.weight": "pytorch_model-00007-of-00042.bin",
"model.layers.4.mlp.down_proj.weight": "pytorch_model-00007-of-00042.bin",
"model.layers.4.mlp.gate_proj.weight": "pytorch_model-00007-of-00042.bin",
"model.layers.4.mlp.up_proj.weight": "pytorch_model-00007-of-00042.bin",
"model.layers.4.post_attention_layernorm.weight": "pytorch_model-00007-of-00042.bin",
"model.layers.4.self_attn.k_proj.weight": "pytorch_model-00006-of-00042.bin",
"model.layers.4.self_attn.o_proj.weight": "pytorch_model-00006-of-00042.bin",
"model.layers.4.self_attn.q_proj.weight": "pytorch_model-00006-of-00042.bin",
"model.layers.4.self_attn.v_proj.weight": "pytorch_model-00006-of-00042.bin",
"model.layers.5.input_layernorm.weight": "pytorch_model-00008-of-00042.bin",
"model.layers.5.mlp.down_proj.weight": "pytorch_model-00008-of-00042.bin",
"model.layers.5.mlp.gate_proj.weight": "pytorch_model-00008-of-00042.bin",
"model.layers.5.mlp.up_proj.weight": "pytorch_model-00008-of-00042.bin",
"model.layers.5.post_attention_layernorm.weight": "pytorch_model-00008-of-00042.bin",
"model.layers.5.self_attn.k_proj.weight": "pytorch_model-00007-of-00042.bin",
"model.layers.5.self_attn.o_proj.weight": "pytorch_model-00008-of-00042.bin",
"model.layers.5.self_attn.q_proj.weight": "pytorch_model-00007-of-00042.bin",
"model.layers.5.self_attn.v_proj.weight": "pytorch_model-00008-of-00042.bin",
"model.layers.6.input_layernorm.weight": "pytorch_model-00010-of-00042.bin",
"model.layers.6.mlp.down_proj.weight": "pytorch_model-00010-of-00042.bin",
"model.layers.6.mlp.gate_proj.weight": "pytorch_model-00009-of-00042.bin",
"model.layers.6.mlp.up_proj.weight": "pytorch_model-00009-of-00042.bin",
"model.layers.6.post_attention_layernorm.weight": "pytorch_model-00010-of-00042.bin",
"model.layers.6.self_attn.k_proj.weight": "pytorch_model-00009-of-00042.bin",
"model.layers.6.self_attn.o_proj.weight": "pytorch_model-00009-of-00042.bin",
"model.layers.6.self_attn.q_proj.weight": "pytorch_model-00009-of-00042.bin",
"model.layers.6.self_attn.v_proj.weight": "pytorch_model-00009-of-00042.bin",
"model.layers.7.input_layernorm.weight": "pytorch_model-00011-of-00042.bin",
"model.layers.7.mlp.down_proj.weight": "pytorch_model-00011-of-00042.bin",
"model.layers.7.mlp.gate_proj.weight": "pytorch_model-00010-of-00042.bin",
"model.layers.7.mlp.up_proj.weight": "pytorch_model-00011-of-00042.bin",
"model.layers.7.post_attention_layernorm.weight": "pytorch_model-00011-of-00042.bin",
"model.layers.7.self_attn.k_proj.weight": "pytorch_model-00010-of-00042.bin",
"model.layers.7.self_attn.o_proj.weight": "pytorch_model-00010-of-00042.bin",
"model.layers.7.self_attn.q_proj.weight": "pytorch_model-00010-of-00042.bin",
"model.layers.7.self_attn.v_proj.weight": "pytorch_model-00010-of-00042.bin",
"model.layers.8.input_layernorm.weight": "pytorch_model-00012-of-00042.bin",
"model.layers.8.mlp.down_proj.weight": "pytorch_model-00012-of-00042.bin",
"model.layers.8.mlp.gate_proj.weight": "pytorch_model-00012-of-00042.bin",
"model.layers.8.mlp.up_proj.weight": "pytorch_model-00012-of-00042.bin",
"model.layers.8.post_attention_layernorm.weight": "pytorch_model-00012-of-00042.bin",
"model.layers.8.self_attn.k_proj.weight": "pytorch_model-00011-of-00042.bin",
"model.layers.8.self_attn.o_proj.weight": "pytorch_model-00011-of-00042.bin",
"model.layers.8.self_attn.q_proj.weight": "pytorch_model-00011-of-00042.bin",
"model.layers.8.self_attn.v_proj.weight": "pytorch_model-00011-of-00042.bin",
"model.layers.9.input_layernorm.weight": "pytorch_model-00013-of-00042.bin",
"model.layers.9.mlp.down_proj.weight": "pytorch_model-00013-of-00042.bin",
"model.layers.9.mlp.gate_proj.weight": "pytorch_model-00013-of-00042.bin",
"model.layers.9.mlp.up_proj.weight": "pytorch_model-00013-of-00042.bin",
"model.layers.9.post_attention_layernorm.weight": "pytorch_model-00013-of-00042.bin",
"model.layers.9.self_attn.k_proj.weight": "pytorch_model-00012-of-00042.bin",
"model.layers.9.self_attn.o_proj.weight": "pytorch_model-00013-of-00042.bin",
"model.layers.9.self_attn.q_proj.weight": "pytorch_model-00012-of-00042.bin",
"model.layers.9.self_attn.v_proj.weight": "pytorch_model-00013-of-00042.bin",
"model.norm.weight": "pytorch_model-00041-of-00042.bin"
}
}

91122
tokenizer.json Normal file
View File

File diff suppressed because it is too large Load Diff

BIN
tokenizer.model (Stored with Git LFS) Normal file
View File

Binary file not shown.

40
tokenizer_config.json Normal file
View File

@@ -0,0 +1,40 @@
{
"added_tokens_decoder": {
"0": {
"content": "<unk>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"1": {
"content": "<s>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"2": {
"content": "</s>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
}
},
"additional_special_tokens": [],
"bos_token": "<s>",
"clean_up_tokenization_spaces": false,
"eos_token": "</s>",
"legacy": true,
"model_max_length": 1000000000000000019884624838656,
"pad_token": null,
"sp_model_kwargs": {},
"spaces_between_special_tokens": false,
"tokenizer_class": "LlamaTokenizer",
"unk_token": "<unk>",
"use_default_system_prompt": true
}