Upload LlamaForCausalLM
Browse files- config.json +24 -0
- generation_config.json +7 -0
- pytorch_model-00001-of-00035.bin +3 -0
- pytorch_model-00002-of-00035.bin +3 -0
- pytorch_model-00003-of-00035.bin +3 -0
- pytorch_model-00004-of-00035.bin +3 -0
- pytorch_model-00005-of-00035.bin +3 -0
- pytorch_model-00006-of-00035.bin +3 -0
- pytorch_model-00007-of-00035.bin +3 -0
- pytorch_model-00008-of-00035.bin +3 -0
- pytorch_model-00009-of-00035.bin +3 -0
- pytorch_model-00010-of-00035.bin +3 -0
- pytorch_model-00011-of-00035.bin +3 -0
- pytorch_model-00012-of-00035.bin +3 -0
- pytorch_model-00013-of-00035.bin +3 -0
- pytorch_model-00014-of-00035.bin +3 -0
- pytorch_model-00015-of-00035.bin +3 -0
- pytorch_model-00016-of-00035.bin +3 -0
- pytorch_model-00017-of-00035.bin +3 -0
- pytorch_model-00018-of-00035.bin +3 -0
- pytorch_model-00019-of-00035.bin +3 -0
- pytorch_model-00020-of-00035.bin +3 -0
- pytorch_model-00021-of-00035.bin +3 -0
- pytorch_model-00022-of-00035.bin +3 -0
- pytorch_model-00023-of-00035.bin +3 -0
- pytorch_model-00024-of-00035.bin +3 -0
- pytorch_model-00025-of-00035.bin +3 -0
- pytorch_model-00026-of-00035.bin +3 -0
- pytorch_model-00027-of-00035.bin +3 -0
- pytorch_model-00028-of-00035.bin +3 -0
- pytorch_model-00029-of-00035.bin +3 -0
- pytorch_model-00030-of-00035.bin +3 -0
- pytorch_model-00031-of-00035.bin +3 -0
- pytorch_model-00032-of-00035.bin +3 -0
- pytorch_model-00033-of-00035.bin +3 -0
- pytorch_model-00034-of-00035.bin +3 -0
- pytorch_model-00035-of-00035.bin +3 -0
- pytorch_model.bin.index.json +330 -0
config.json
ADDED
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "sharpbai/baichuan-vicuna-7b",
|
3 |
+
"architectures": [
|
4 |
+
"LlamaForCausalLM"
|
5 |
+
],
|
6 |
+
"bos_token_id": 1,
|
7 |
+
"eos_token_id": 2,
|
8 |
+
"hidden_act": "silu",
|
9 |
+
"hidden_size": 4096,
|
10 |
+
"initializer_range": 0.02,
|
11 |
+
"intermediate_size": 11008,
|
12 |
+
"max_position_embeddings": 4096,
|
13 |
+
"max_sequence_length": 4096,
|
14 |
+
"model_type": "llama",
|
15 |
+
"num_attention_heads": 32,
|
16 |
+
"num_hidden_layers": 32,
|
17 |
+
"pad_token_id": 0,
|
18 |
+
"rms_norm_eps": 1e-06,
|
19 |
+
"tie_word_embeddings": false,
|
20 |
+
"torch_dtype": "float16",
|
21 |
+
"transformers_version": "4.30.1",
|
22 |
+
"use_cache": true,
|
23 |
+
"vocab_size": 64000
|
24 |
+
}
|
generation_config.json
ADDED
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_from_model_config": true,
|
3 |
+
"bos_token_id": 1,
|
4 |
+
"eos_token_id": 2,
|
5 |
+
"pad_token_id": 0,
|
6 |
+
"transformers_version": "4.30.1"
|
7 |
+
}
|
pytorch_model-00001-of-00035.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:cf02c4d7b83fa495ef73f290168a195ea899d0744f0fe32194046846bd87efe0
|
3 |
+
size 537
|
pytorch_model-00002-of-00035.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:0a14e74fc0e1cfc74ae5f1483823f3f474901e68fdcb9a4e6f3ba0b0e56bcc18
|
3 |
+
size 524288938
|
pytorch_model-00003-of-00035.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:586b577f016867857c4a2d3b42310553787968bd9a99fbcbe40269fad75b9db5
|
3 |
+
size 404770755
|
pytorch_model-00004-of-00035.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:86a61d34452c76c4eace9bdd88afe6985c01f83e9e9a06a2119b899538e6e866
|
3 |
+
size 404770755
|
pytorch_model-00005-of-00035.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:de8139e15e5f087d55b0fe16a01a868d4e6ebac52ccc3f01d4d6ce06c053c298
|
3 |
+
size 404770755
|
pytorch_model-00006-of-00035.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:26bef9613c2df5f267f452467110e0351df1fdd82c48bd9eff8f746c87616543
|
3 |
+
size 404770755
|
pytorch_model-00007-of-00035.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f9faadef490dc8789d2ef06daa0a3f0b4d9f6f65160a181fd65342e39a324d14
|
3 |
+
size 404770755
|
pytorch_model-00008-of-00035.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:2ee8882e5d1217731a55349b1d0ba11f2ef5c6d6f86668ffec55f2aa2a020539
|
3 |
+
size 404770755
|
pytorch_model-00009-of-00035.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f4302a68194adb7062859418c28e1cbd612320022bf05839d7374dfb9868cf98
|
3 |
+
size 404770755
|
pytorch_model-00010-of-00035.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:7610228dbc29764dbdd0b000be2ec8b9c172b9755661185728e43353e16c91e3
|
3 |
+
size 404770755
|
pytorch_model-00011-of-00035.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:186677c629d54794e2560e410c239af1894e5a1b12a5051bd47b229304282532
|
3 |
+
size 404770755
|
pytorch_model-00012-of-00035.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:30b5052340822ec010c9158815de833eb14ab39c6fe6875d436ed5bd5278b806
|
3 |
+
size 404770755
|
pytorch_model-00013-of-00035.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:0199c7b844ad772efc0b84796737b8813e4fcc787fe064c5b2b1a2f7bfb6e579
|
3 |
+
size 404770755
|
pytorch_model-00014-of-00035.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:b855c5e3d8f35d6fe0a0e9c60425dd900341b9c1b3c83db2b219c91bd8393302
|
3 |
+
size 404770755
|
pytorch_model-00015-of-00035.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:8c674f3cd2f2fbaba294aa3e9183f046dddc92e3db25cf8b24cc94c1c5914d1f
|
3 |
+
size 404770755
|
pytorch_model-00016-of-00035.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:d8d3734975aca398e944b55a23f76a8336eb7ba4bd129bb70357bfa4bb5dad19
|
3 |
+
size 404770755
|
pytorch_model-00017-of-00035.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:26a2786c2aeefac4f879e17f68afa0a6af2a9728a2689859bad2488bdfff63ef
|
3 |
+
size 404770755
|
pytorch_model-00018-of-00035.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:7968fa439872bc73db517851f12d9c812056b12826549f609ade55476f3582ba
|
3 |
+
size 404770755
|
pytorch_model-00019-of-00035.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:a7a6ca7d9313fddd26942a798467e548609d6f31aeb47bef06268be0454e4f46
|
3 |
+
size 404770755
|
pytorch_model-00020-of-00035.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:10c421d668e7929a01f6ac69499b6b1b16749cbe4a9ff57720f691e7ba10862f
|
3 |
+
size 404770755
|
pytorch_model-00021-of-00035.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:a3d0b01667c594f7677aae9f9c308ffa3957e4c9398b9ae761f4bf086579eac6
|
3 |
+
size 404770755
|
pytorch_model-00022-of-00035.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:3a686a7f67ea2ce5095611e8333ea1ce3b74af029fd65c37a6b2819afa6bb0bd
|
3 |
+
size 404770755
|
pytorch_model-00023-of-00035.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:4231b436cedf2f26e5926e9de3a6d8d2751c3bfc433d6e56c0c136fa7abb8068
|
3 |
+
size 404770755
|
pytorch_model-00024-of-00035.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:47fa36966b78893e4080938225066551dc70c98350a55f22fe8c872622e78dee
|
3 |
+
size 404770755
|
pytorch_model-00025-of-00035.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:36b624d23f3aad75fa72e14421b022da00444c68daf3b6c0f8273b2dec9af2b3
|
3 |
+
size 404770755
|
pytorch_model-00026-of-00035.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:e90a83ff16f6c1356432a7bf589909fc0c8d0082aa0a4cf67e4837611dddaccf
|
3 |
+
size 404770755
|
pytorch_model-00027-of-00035.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:3ef9c35dcc7c65cf8eab93796587d475132bced892d0a4602e22e2e3c5f9bbb9
|
3 |
+
size 404770755
|
pytorch_model-00028-of-00035.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:392ebfdc9b7077dfcdaf516dcdb5e07e95d5b9b160736b562e20cdfbbe60034e
|
3 |
+
size 404770755
|
pytorch_model-00029-of-00035.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:41ac9f7929cd8940b9d9e723929074bee19eb7b325f0ba8563d9f05dc8421984
|
3 |
+
size 404770755
|
pytorch_model-00030-of-00035.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:d4afc6f2954e881928b3527668e7d8f080ac3382ecdc204db04293bbf597fc46
|
3 |
+
size 404770755
|
pytorch_model-00031-of-00035.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:daabfa599db583247ddf4e0385c49c100df286deaaf45be554a752c279ab2da5
|
3 |
+
size 404770755
|
pytorch_model-00032-of-00035.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:26c09597138e295592c9522b5098415c2a1026123bda3d1ebe49037d11bdafd1
|
3 |
+
size 404770755
|
pytorch_model-00033-of-00035.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:4454c6335449d6d74f8c29b998e53649cc9a731085095573073c1948b8c9f759
|
3 |
+
size 404770755
|
pytorch_model-00034-of-00035.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:a369fb0a848554948a56eed9ee4484c7e273c7289946d2b96d9179572c8c1300
|
3 |
+
size 404779221
|
pytorch_model-00035-of-00035.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:17998d264b07fdead64e9eaf895b0278df656eecdbe00ab7b4a764b06433e46d
|
3 |
+
size 524288938
|
pytorch_model.bin.index.json
ADDED
@@ -0,0 +1,330 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"metadata": {
|
3 |
+
"total_size": 14001127424
|
4 |
+
},
|
5 |
+
"weight_map": {
|
6 |
+
"lm_head.weight": "pytorch_model-00035-of-00035.bin",
|
7 |
+
"model.embed_tokens.weight": "pytorch_model-00002-of-00035.bin",
|
8 |
+
"model.layers.0.input_layernorm.weight": "pytorch_model-00003-of-00035.bin",
|
9 |
+
"model.layers.0.mlp.down_proj.weight": "pytorch_model-00003-of-00035.bin",
|
10 |
+
"model.layers.0.mlp.gate_proj.weight": "pytorch_model-00003-of-00035.bin",
|
11 |
+
"model.layers.0.mlp.up_proj.weight": "pytorch_model-00003-of-00035.bin",
|
12 |
+
"model.layers.0.post_attention_layernorm.weight": "pytorch_model-00003-of-00035.bin",
|
13 |
+
"model.layers.0.self_attn.k_proj.weight": "pytorch_model-00003-of-00035.bin",
|
14 |
+
"model.layers.0.self_attn.o_proj.weight": "pytorch_model-00003-of-00035.bin",
|
15 |
+
"model.layers.0.self_attn.q_proj.weight": "pytorch_model-00003-of-00035.bin",
|
16 |
+
"model.layers.0.self_attn.rotary_emb.inv_freq": "pytorch_model-00003-of-00035.bin",
|
17 |
+
"model.layers.0.self_attn.v_proj.weight": "pytorch_model-00003-of-00035.bin",
|
18 |
+
"model.layers.1.input_layernorm.weight": "pytorch_model-00004-of-00035.bin",
|
19 |
+
"model.layers.1.mlp.down_proj.weight": "pytorch_model-00004-of-00035.bin",
|
20 |
+
"model.layers.1.mlp.gate_proj.weight": "pytorch_model-00004-of-00035.bin",
|
21 |
+
"model.layers.1.mlp.up_proj.weight": "pytorch_model-00004-of-00035.bin",
|
22 |
+
"model.layers.1.post_attention_layernorm.weight": "pytorch_model-00004-of-00035.bin",
|
23 |
+
"model.layers.1.self_attn.k_proj.weight": "pytorch_model-00004-of-00035.bin",
|
24 |
+
"model.layers.1.self_attn.o_proj.weight": "pytorch_model-00004-of-00035.bin",
|
25 |
+
"model.layers.1.self_attn.q_proj.weight": "pytorch_model-00004-of-00035.bin",
|
26 |
+
"model.layers.1.self_attn.rotary_emb.inv_freq": "pytorch_model-00004-of-00035.bin",
|
27 |
+
"model.layers.1.self_attn.v_proj.weight": "pytorch_model-00004-of-00035.bin",
|
28 |
+
"model.layers.10.input_layernorm.weight": "pytorch_model-00013-of-00035.bin",
|
29 |
+
"model.layers.10.mlp.down_proj.weight": "pytorch_model-00013-of-00035.bin",
|
30 |
+
"model.layers.10.mlp.gate_proj.weight": "pytorch_model-00013-of-00035.bin",
|
31 |
+
"model.layers.10.mlp.up_proj.weight": "pytorch_model-00013-of-00035.bin",
|
32 |
+
"model.layers.10.post_attention_layernorm.weight": "pytorch_model-00013-of-00035.bin",
|
33 |
+
"model.layers.10.self_attn.k_proj.weight": "pytorch_model-00013-of-00035.bin",
|
34 |
+
"model.layers.10.self_attn.o_proj.weight": "pytorch_model-00013-of-00035.bin",
|
35 |
+
"model.layers.10.self_attn.q_proj.weight": "pytorch_model-00013-of-00035.bin",
|
36 |
+
"model.layers.10.self_attn.rotary_emb.inv_freq": "pytorch_model-00013-of-00035.bin",
|
37 |
+
"model.layers.10.self_attn.v_proj.weight": "pytorch_model-00013-of-00035.bin",
|
38 |
+
"model.layers.11.input_layernorm.weight": "pytorch_model-00014-of-00035.bin",
|
39 |
+
"model.layers.11.mlp.down_proj.weight": "pytorch_model-00014-of-00035.bin",
|
40 |
+
"model.layers.11.mlp.gate_proj.weight": "pytorch_model-00014-of-00035.bin",
|
41 |
+
"model.layers.11.mlp.up_proj.weight": "pytorch_model-00014-of-00035.bin",
|
42 |
+
"model.layers.11.post_attention_layernorm.weight": "pytorch_model-00014-of-00035.bin",
|
43 |
+
"model.layers.11.self_attn.k_proj.weight": "pytorch_model-00014-of-00035.bin",
|
44 |
+
"model.layers.11.self_attn.o_proj.weight": "pytorch_model-00014-of-00035.bin",
|
45 |
+
"model.layers.11.self_attn.q_proj.weight": "pytorch_model-00014-of-00035.bin",
|
46 |
+
"model.layers.11.self_attn.rotary_emb.inv_freq": "pytorch_model-00014-of-00035.bin",
|
47 |
+
"model.layers.11.self_attn.v_proj.weight": "pytorch_model-00014-of-00035.bin",
|
48 |
+
"model.layers.12.input_layernorm.weight": "pytorch_model-00015-of-00035.bin",
|
49 |
+
"model.layers.12.mlp.down_proj.weight": "pytorch_model-00015-of-00035.bin",
|
50 |
+
"model.layers.12.mlp.gate_proj.weight": "pytorch_model-00015-of-00035.bin",
|
51 |
+
"model.layers.12.mlp.up_proj.weight": "pytorch_model-00015-of-00035.bin",
|
52 |
+
"model.layers.12.post_attention_layernorm.weight": "pytorch_model-00015-of-00035.bin",
|
53 |
+
"model.layers.12.self_attn.k_proj.weight": "pytorch_model-00015-of-00035.bin",
|
54 |
+
"model.layers.12.self_attn.o_proj.weight": "pytorch_model-00015-of-00035.bin",
|
55 |
+
"model.layers.12.self_attn.q_proj.weight": "pytorch_model-00015-of-00035.bin",
|
56 |
+
"model.layers.12.self_attn.rotary_emb.inv_freq": "pytorch_model-00015-of-00035.bin",
|
57 |
+
"model.layers.12.self_attn.v_proj.weight": "pytorch_model-00015-of-00035.bin",
|
58 |
+
"model.layers.13.input_layernorm.weight": "pytorch_model-00016-of-00035.bin",
|
59 |
+
"model.layers.13.mlp.down_proj.weight": "pytorch_model-00016-of-00035.bin",
|
60 |
+
"model.layers.13.mlp.gate_proj.weight": "pytorch_model-00016-of-00035.bin",
|
61 |
+
"model.layers.13.mlp.up_proj.weight": "pytorch_model-00016-of-00035.bin",
|
62 |
+
"model.layers.13.post_attention_layernorm.weight": "pytorch_model-00016-of-00035.bin",
|
63 |
+
"model.layers.13.self_attn.k_proj.weight": "pytorch_model-00016-of-00035.bin",
|
64 |
+
"model.layers.13.self_attn.o_proj.weight": "pytorch_model-00016-of-00035.bin",
|
65 |
+
"model.layers.13.self_attn.q_proj.weight": "pytorch_model-00016-of-00035.bin",
|
66 |
+
"model.layers.13.self_attn.rotary_emb.inv_freq": "pytorch_model-00016-of-00035.bin",
|
67 |
+
"model.layers.13.self_attn.v_proj.weight": "pytorch_model-00016-of-00035.bin",
|
68 |
+
"model.layers.14.input_layernorm.weight": "pytorch_model-00017-of-00035.bin",
|
69 |
+
"model.layers.14.mlp.down_proj.weight": "pytorch_model-00017-of-00035.bin",
|
70 |
+
"model.layers.14.mlp.gate_proj.weight": "pytorch_model-00017-of-00035.bin",
|
71 |
+
"model.layers.14.mlp.up_proj.weight": "pytorch_model-00017-of-00035.bin",
|
72 |
+
"model.layers.14.post_attention_layernorm.weight": "pytorch_model-00017-of-00035.bin",
|
73 |
+
"model.layers.14.self_attn.k_proj.weight": "pytorch_model-00017-of-00035.bin",
|
74 |
+
"model.layers.14.self_attn.o_proj.weight": "pytorch_model-00017-of-00035.bin",
|
75 |
+
"model.layers.14.self_attn.q_proj.weight": "pytorch_model-00017-of-00035.bin",
|
76 |
+
"model.layers.14.self_attn.rotary_emb.inv_freq": "pytorch_model-00017-of-00035.bin",
|
77 |
+
"model.layers.14.self_attn.v_proj.weight": "pytorch_model-00017-of-00035.bin",
|
78 |
+
"model.layers.15.input_layernorm.weight": "pytorch_model-00018-of-00035.bin",
|
79 |
+
"model.layers.15.mlp.down_proj.weight": "pytorch_model-00018-of-00035.bin",
|
80 |
+
"model.layers.15.mlp.gate_proj.weight": "pytorch_model-00018-of-00035.bin",
|
81 |
+
"model.layers.15.mlp.up_proj.weight": "pytorch_model-00018-of-00035.bin",
|
82 |
+
"model.layers.15.post_attention_layernorm.weight": "pytorch_model-00018-of-00035.bin",
|
83 |
+
"model.layers.15.self_attn.k_proj.weight": "pytorch_model-00018-of-00035.bin",
|
84 |
+
"model.layers.15.self_attn.o_proj.weight": "pytorch_model-00018-of-00035.bin",
|
85 |
+
"model.layers.15.self_attn.q_proj.weight": "pytorch_model-00018-of-00035.bin",
|
86 |
+
"model.layers.15.self_attn.rotary_emb.inv_freq": "pytorch_model-00018-of-00035.bin",
|
87 |
+
"model.layers.15.self_attn.v_proj.weight": "pytorch_model-00018-of-00035.bin",
|
88 |
+
"model.layers.16.input_layernorm.weight": "pytorch_model-00019-of-00035.bin",
|
89 |
+
"model.layers.16.mlp.down_proj.weight": "pytorch_model-00019-of-00035.bin",
|
90 |
+
"model.layers.16.mlp.gate_proj.weight": "pytorch_model-00019-of-00035.bin",
|
91 |
+
"model.layers.16.mlp.up_proj.weight": "pytorch_model-00019-of-00035.bin",
|
92 |
+
"model.layers.16.post_attention_layernorm.weight": "pytorch_model-00019-of-00035.bin",
|
93 |
+
"model.layers.16.self_attn.k_proj.weight": "pytorch_model-00019-of-00035.bin",
|
94 |
+
"model.layers.16.self_attn.o_proj.weight": "pytorch_model-00019-of-00035.bin",
|
95 |
+
"model.layers.16.self_attn.q_proj.weight": "pytorch_model-00019-of-00035.bin",
|
96 |
+
"model.layers.16.self_attn.rotary_emb.inv_freq": "pytorch_model-00019-of-00035.bin",
|
97 |
+
"model.layers.16.self_attn.v_proj.weight": "pytorch_model-00019-of-00035.bin",
|
98 |
+
"model.layers.17.input_layernorm.weight": "pytorch_model-00020-of-00035.bin",
|
99 |
+
"model.layers.17.mlp.down_proj.weight": "pytorch_model-00020-of-00035.bin",
|
100 |
+
"model.layers.17.mlp.gate_proj.weight": "pytorch_model-00020-of-00035.bin",
|
101 |
+
"model.layers.17.mlp.up_proj.weight": "pytorch_model-00020-of-00035.bin",
|
102 |
+
"model.layers.17.post_attention_layernorm.weight": "pytorch_model-00020-of-00035.bin",
|
103 |
+
"model.layers.17.self_attn.k_proj.weight": "pytorch_model-00020-of-00035.bin",
|
104 |
+
"model.layers.17.self_attn.o_proj.weight": "pytorch_model-00020-of-00035.bin",
|
105 |
+
"model.layers.17.self_attn.q_proj.weight": "pytorch_model-00020-of-00035.bin",
|
106 |
+
"model.layers.17.self_attn.rotary_emb.inv_freq": "pytorch_model-00020-of-00035.bin",
|
107 |
+
"model.layers.17.self_attn.v_proj.weight": "pytorch_model-00020-of-00035.bin",
|
108 |
+
"model.layers.18.input_layernorm.weight": "pytorch_model-00021-of-00035.bin",
|
109 |
+
"model.layers.18.mlp.down_proj.weight": "pytorch_model-00021-of-00035.bin",
|
110 |
+
"model.layers.18.mlp.gate_proj.weight": "pytorch_model-00021-of-00035.bin",
|
111 |
+
"model.layers.18.mlp.up_proj.weight": "pytorch_model-00021-of-00035.bin",
|
112 |
+
"model.layers.18.post_attention_layernorm.weight": "pytorch_model-00021-of-00035.bin",
|
113 |
+
"model.layers.18.self_attn.k_proj.weight": "pytorch_model-00021-of-00035.bin",
|
114 |
+
"model.layers.18.self_attn.o_proj.weight": "pytorch_model-00021-of-00035.bin",
|
115 |
+
"model.layers.18.self_attn.q_proj.weight": "pytorch_model-00021-of-00035.bin",
|
116 |
+
"model.layers.18.self_attn.rotary_emb.inv_freq": "pytorch_model-00021-of-00035.bin",
|
117 |
+
"model.layers.18.self_attn.v_proj.weight": "pytorch_model-00021-of-00035.bin",
|
118 |
+
"model.layers.19.input_layernorm.weight": "pytorch_model-00022-of-00035.bin",
|
119 |
+
"model.layers.19.mlp.down_proj.weight": "pytorch_model-00022-of-00035.bin",
|
120 |
+
"model.layers.19.mlp.gate_proj.weight": "pytorch_model-00022-of-00035.bin",
|
121 |
+
"model.layers.19.mlp.up_proj.weight": "pytorch_model-00022-of-00035.bin",
|
122 |
+
"model.layers.19.post_attention_layernorm.weight": "pytorch_model-00022-of-00035.bin",
|
123 |
+
"model.layers.19.self_attn.k_proj.weight": "pytorch_model-00022-of-00035.bin",
|
124 |
+
"model.layers.19.self_attn.o_proj.weight": "pytorch_model-00022-of-00035.bin",
|
125 |
+
"model.layers.19.self_attn.q_proj.weight": "pytorch_model-00022-of-00035.bin",
|
126 |
+
"model.layers.19.self_attn.rotary_emb.inv_freq": "pytorch_model-00022-of-00035.bin",
|
127 |
+
"model.layers.19.self_attn.v_proj.weight": "pytorch_model-00022-of-00035.bin",
|
128 |
+
"model.layers.2.input_layernorm.weight": "pytorch_model-00005-of-00035.bin",
|
129 |
+
"model.layers.2.mlp.down_proj.weight": "pytorch_model-00005-of-00035.bin",
|
130 |
+
"model.layers.2.mlp.gate_proj.weight": "pytorch_model-00005-of-00035.bin",
|
131 |
+
"model.layers.2.mlp.up_proj.weight": "pytorch_model-00005-of-00035.bin",
|
132 |
+
"model.layers.2.post_attention_layernorm.weight": "pytorch_model-00005-of-00035.bin",
|
133 |
+
"model.layers.2.self_attn.k_proj.weight": "pytorch_model-00005-of-00035.bin",
|
134 |
+
"model.layers.2.self_attn.o_proj.weight": "pytorch_model-00005-of-00035.bin",
|
135 |
+
"model.layers.2.self_attn.q_proj.weight": "pytorch_model-00005-of-00035.bin",
|
136 |
+
"model.layers.2.self_attn.rotary_emb.inv_freq": "pytorch_model-00005-of-00035.bin",
|
137 |
+
"model.layers.2.self_attn.v_proj.weight": "pytorch_model-00005-of-00035.bin",
|
138 |
+
"model.layers.20.input_layernorm.weight": "pytorch_model-00023-of-00035.bin",
|
139 |
+
"model.layers.20.mlp.down_proj.weight": "pytorch_model-00023-of-00035.bin",
|
140 |
+
"model.layers.20.mlp.gate_proj.weight": "pytorch_model-00023-of-00035.bin",
|
141 |
+
"model.layers.20.mlp.up_proj.weight": "pytorch_model-00023-of-00035.bin",
|
142 |
+
"model.layers.20.post_attention_layernorm.weight": "pytorch_model-00023-of-00035.bin",
|
143 |
+
"model.layers.20.self_attn.k_proj.weight": "pytorch_model-00023-of-00035.bin",
|
144 |
+
"model.layers.20.self_attn.o_proj.weight": "pytorch_model-00023-of-00035.bin",
|
145 |
+
"model.layers.20.self_attn.q_proj.weight": "pytorch_model-00023-of-00035.bin",
|
146 |
+
"model.layers.20.self_attn.rotary_emb.inv_freq": "pytorch_model-00023-of-00035.bin",
|
147 |
+
"model.layers.20.self_attn.v_proj.weight": "pytorch_model-00023-of-00035.bin",
|
148 |
+
"model.layers.21.input_layernorm.weight": "pytorch_model-00024-of-00035.bin",
|
149 |
+
"model.layers.21.mlp.down_proj.weight": "pytorch_model-00024-of-00035.bin",
|
150 |
+
"model.layers.21.mlp.gate_proj.weight": "pytorch_model-00024-of-00035.bin",
|
151 |
+
"model.layers.21.mlp.up_proj.weight": "pytorch_model-00024-of-00035.bin",
|
152 |
+
"model.layers.21.post_attention_layernorm.weight": "pytorch_model-00024-of-00035.bin",
|
153 |
+
"model.layers.21.self_attn.k_proj.weight": "pytorch_model-00024-of-00035.bin",
|
154 |
+
"model.layers.21.self_attn.o_proj.weight": "pytorch_model-00024-of-00035.bin",
|
155 |
+
"model.layers.21.self_attn.q_proj.weight": "pytorch_model-00024-of-00035.bin",
|
156 |
+
"model.layers.21.self_attn.rotary_emb.inv_freq": "pytorch_model-00024-of-00035.bin",
|
157 |
+
"model.layers.21.self_attn.v_proj.weight": "pytorch_model-00024-of-00035.bin",
|
158 |
+
"model.layers.22.input_layernorm.weight": "pytorch_model-00025-of-00035.bin",
|
159 |
+
"model.layers.22.mlp.down_proj.weight": "pytorch_model-00025-of-00035.bin",
|
160 |
+
"model.layers.22.mlp.gate_proj.weight": "pytorch_model-00025-of-00035.bin",
|
161 |
+
"model.layers.22.mlp.up_proj.weight": "pytorch_model-00025-of-00035.bin",
|
162 |
+
"model.layers.22.post_attention_layernorm.weight": "pytorch_model-00025-of-00035.bin",
|
163 |
+
"model.layers.22.self_attn.k_proj.weight": "pytorch_model-00025-of-00035.bin",
|
164 |
+
"model.layers.22.self_attn.o_proj.weight": "pytorch_model-00025-of-00035.bin",
|
165 |
+
"model.layers.22.self_attn.q_proj.weight": "pytorch_model-00025-of-00035.bin",
|
166 |
+
"model.layers.22.self_attn.rotary_emb.inv_freq": "pytorch_model-00025-of-00035.bin",
|
167 |
+
"model.layers.22.self_attn.v_proj.weight": "pytorch_model-00025-of-00035.bin",
|
168 |
+
"model.layers.23.input_layernorm.weight": "pytorch_model-00026-of-00035.bin",
|
169 |
+
"model.layers.23.mlp.down_proj.weight": "pytorch_model-00026-of-00035.bin",
|
170 |
+
"model.layers.23.mlp.gate_proj.weight": "pytorch_model-00026-of-00035.bin",
|
171 |
+
"model.layers.23.mlp.up_proj.weight": "pytorch_model-00026-of-00035.bin",
|
172 |
+
"model.layers.23.post_attention_layernorm.weight": "pytorch_model-00026-of-00035.bin",
|
173 |
+
"model.layers.23.self_attn.k_proj.weight": "pytorch_model-00026-of-00035.bin",
|
174 |
+
"model.layers.23.self_attn.o_proj.weight": "pytorch_model-00026-of-00035.bin",
|
175 |
+
"model.layers.23.self_attn.q_proj.weight": "pytorch_model-00026-of-00035.bin",
|
176 |
+
"model.layers.23.self_attn.rotary_emb.inv_freq": "pytorch_model-00026-of-00035.bin",
|
177 |
+
"model.layers.23.self_attn.v_proj.weight": "pytorch_model-00026-of-00035.bin",
|
178 |
+
"model.layers.24.input_layernorm.weight": "pytorch_model-00027-of-00035.bin",
|
179 |
+
"model.layers.24.mlp.down_proj.weight": "pytorch_model-00027-of-00035.bin",
|
180 |
+
"model.layers.24.mlp.gate_proj.weight": "pytorch_model-00027-of-00035.bin",
|
181 |
+
"model.layers.24.mlp.up_proj.weight": "pytorch_model-00027-of-00035.bin",
|
182 |
+
"model.layers.24.post_attention_layernorm.weight": "pytorch_model-00027-of-00035.bin",
|
183 |
+
"model.layers.24.self_attn.k_proj.weight": "pytorch_model-00027-of-00035.bin",
|
184 |
+
"model.layers.24.self_attn.o_proj.weight": "pytorch_model-00027-of-00035.bin",
|
185 |
+
"model.layers.24.self_attn.q_proj.weight": "pytorch_model-00027-of-00035.bin",
|
186 |
+
"model.layers.24.self_attn.rotary_emb.inv_freq": "pytorch_model-00027-of-00035.bin",
|
187 |
+
"model.layers.24.self_attn.v_proj.weight": "pytorch_model-00027-of-00035.bin",
|
188 |
+
"model.layers.25.input_layernorm.weight": "pytorch_model-00028-of-00035.bin",
|
189 |
+
"model.layers.25.mlp.down_proj.weight": "pytorch_model-00028-of-00035.bin",
|
190 |
+
"model.layers.25.mlp.gate_proj.weight": "pytorch_model-00028-of-00035.bin",
|
191 |
+
"model.layers.25.mlp.up_proj.weight": "pytorch_model-00028-of-00035.bin",
|
192 |
+
"model.layers.25.post_attention_layernorm.weight": "pytorch_model-00028-of-00035.bin",
|
193 |
+
"model.layers.25.self_attn.k_proj.weight": "pytorch_model-00028-of-00035.bin",
|
194 |
+
"model.layers.25.self_attn.o_proj.weight": "pytorch_model-00028-of-00035.bin",
|
195 |
+
"model.layers.25.self_attn.q_proj.weight": "pytorch_model-00028-of-00035.bin",
|
196 |
+
"model.layers.25.self_attn.rotary_emb.inv_freq": "pytorch_model-00028-of-00035.bin",
|
197 |
+
"model.layers.25.self_attn.v_proj.weight": "pytorch_model-00028-of-00035.bin",
|
198 |
+
"model.layers.26.input_layernorm.weight": "pytorch_model-00029-of-00035.bin",
|
199 |
+
"model.layers.26.mlp.down_proj.weight": "pytorch_model-00029-of-00035.bin",
|
200 |
+
"model.layers.26.mlp.gate_proj.weight": "pytorch_model-00029-of-00035.bin",
|
201 |
+
"model.layers.26.mlp.up_proj.weight": "pytorch_model-00029-of-00035.bin",
|
202 |
+
"model.layers.26.post_attention_layernorm.weight": "pytorch_model-00029-of-00035.bin",
|
203 |
+
"model.layers.26.self_attn.k_proj.weight": "pytorch_model-00029-of-00035.bin",
|
204 |
+
"model.layers.26.self_attn.o_proj.weight": "pytorch_model-00029-of-00035.bin",
|
205 |
+
"model.layers.26.self_attn.q_proj.weight": "pytorch_model-00029-of-00035.bin",
|
206 |
+
"model.layers.26.self_attn.rotary_emb.inv_freq": "pytorch_model-00029-of-00035.bin",
|
207 |
+
"model.layers.26.self_attn.v_proj.weight": "pytorch_model-00029-of-00035.bin",
|
208 |
+
"model.layers.27.input_layernorm.weight": "pytorch_model-00030-of-00035.bin",
|
209 |
+
"model.layers.27.mlp.down_proj.weight": "pytorch_model-00030-of-00035.bin",
|
210 |
+
"model.layers.27.mlp.gate_proj.weight": "pytorch_model-00030-of-00035.bin",
|
211 |
+
"model.layers.27.mlp.up_proj.weight": "pytorch_model-00030-of-00035.bin",
|
212 |
+
"model.layers.27.post_attention_layernorm.weight": "pytorch_model-00030-of-00035.bin",
|
213 |
+
"model.layers.27.self_attn.k_proj.weight": "pytorch_model-00030-of-00035.bin",
|
214 |
+
"model.layers.27.self_attn.o_proj.weight": "pytorch_model-00030-of-00035.bin",
|
215 |
+
"model.layers.27.self_attn.q_proj.weight": "pytorch_model-00030-of-00035.bin",
|
216 |
+
"model.layers.27.self_attn.rotary_emb.inv_freq": "pytorch_model-00030-of-00035.bin",
|
217 |
+
"model.layers.27.self_attn.v_proj.weight": "pytorch_model-00030-of-00035.bin",
|
218 |
+
"model.layers.28.input_layernorm.weight": "pytorch_model-00031-of-00035.bin",
|
219 |
+
"model.layers.28.mlp.down_proj.weight": "pytorch_model-00031-of-00035.bin",
|
220 |
+
"model.layers.28.mlp.gate_proj.weight": "pytorch_model-00031-of-00035.bin",
|
221 |
+
"model.layers.28.mlp.up_proj.weight": "pytorch_model-00031-of-00035.bin",
|
222 |
+
"model.layers.28.post_attention_layernorm.weight": "pytorch_model-00031-of-00035.bin",
|
223 |
+
"model.layers.28.self_attn.k_proj.weight": "pytorch_model-00031-of-00035.bin",
|
224 |
+
"model.layers.28.self_attn.o_proj.weight": "pytorch_model-00031-of-00035.bin",
|
225 |
+
"model.layers.28.self_attn.q_proj.weight": "pytorch_model-00031-of-00035.bin",
|
226 |
+
"model.layers.28.self_attn.rotary_emb.inv_freq": "pytorch_model-00031-of-00035.bin",
|
227 |
+
"model.layers.28.self_attn.v_proj.weight": "pytorch_model-00031-of-00035.bin",
|
228 |
+
"model.layers.29.input_layernorm.weight": "pytorch_model-00032-of-00035.bin",
|
229 |
+
"model.layers.29.mlp.down_proj.weight": "pytorch_model-00032-of-00035.bin",
|
230 |
+
"model.layers.29.mlp.gate_proj.weight": "pytorch_model-00032-of-00035.bin",
|
231 |
+
"model.layers.29.mlp.up_proj.weight": "pytorch_model-00032-of-00035.bin",
|
232 |
+
"model.layers.29.post_attention_layernorm.weight": "pytorch_model-00032-of-00035.bin",
|
233 |
+
"model.layers.29.self_attn.k_proj.weight": "pytorch_model-00032-of-00035.bin",
|
234 |
+
"model.layers.29.self_attn.o_proj.weight": "pytorch_model-00032-of-00035.bin",
|
235 |
+
"model.layers.29.self_attn.q_proj.weight": "pytorch_model-00032-of-00035.bin",
|
236 |
+
"model.layers.29.self_attn.rotary_emb.inv_freq": "pytorch_model-00032-of-00035.bin",
|
237 |
+
"model.layers.29.self_attn.v_proj.weight": "pytorch_model-00032-of-00035.bin",
|
238 |
+
"model.layers.3.input_layernorm.weight": "pytorch_model-00006-of-00035.bin",
|
239 |
+
"model.layers.3.mlp.down_proj.weight": "pytorch_model-00006-of-00035.bin",
|
240 |
+
"model.layers.3.mlp.gate_proj.weight": "pytorch_model-00006-of-00035.bin",
|
241 |
+
"model.layers.3.mlp.up_proj.weight": "pytorch_model-00006-of-00035.bin",
|
242 |
+
"model.layers.3.post_attention_layernorm.weight": "pytorch_model-00006-of-00035.bin",
|
243 |
+
"model.layers.3.self_attn.k_proj.weight": "pytorch_model-00006-of-00035.bin",
|
244 |
+
"model.layers.3.self_attn.o_proj.weight": "pytorch_model-00006-of-00035.bin",
|
245 |
+
"model.layers.3.self_attn.q_proj.weight": "pytorch_model-00006-of-00035.bin",
|
246 |
+
"model.layers.3.self_attn.rotary_emb.inv_freq": "pytorch_model-00006-of-00035.bin",
|
247 |
+
"model.layers.3.self_attn.v_proj.weight": "pytorch_model-00006-of-00035.bin",
|
248 |
+
"model.layers.30.input_layernorm.weight": "pytorch_model-00033-of-00035.bin",
|
249 |
+
"model.layers.30.mlp.down_proj.weight": "pytorch_model-00033-of-00035.bin",
|
250 |
+
"model.layers.30.mlp.gate_proj.weight": "pytorch_model-00033-of-00035.bin",
|
251 |
+
"model.layers.30.mlp.up_proj.weight": "pytorch_model-00033-of-00035.bin",
|
252 |
+
"model.layers.30.post_attention_layernorm.weight": "pytorch_model-00033-of-00035.bin",
|
253 |
+
"model.layers.30.self_attn.k_proj.weight": "pytorch_model-00033-of-00035.bin",
|
254 |
+
"model.layers.30.self_attn.o_proj.weight": "pytorch_model-00033-of-00035.bin",
|
255 |
+
"model.layers.30.self_attn.q_proj.weight": "pytorch_model-00033-of-00035.bin",
|
256 |
+
"model.layers.30.self_attn.rotary_emb.inv_freq": "pytorch_model-00033-of-00035.bin",
|
257 |
+
"model.layers.30.self_attn.v_proj.weight": "pytorch_model-00033-of-00035.bin",
|
258 |
+
"model.layers.31.input_layernorm.weight": "pytorch_model-00034-of-00035.bin",
|
259 |
+
"model.layers.31.mlp.down_proj.weight": "pytorch_model-00034-of-00035.bin",
|
260 |
+
"model.layers.31.mlp.gate_proj.weight": "pytorch_model-00034-of-00035.bin",
|
261 |
+
"model.layers.31.mlp.up_proj.weight": "pytorch_model-00034-of-00035.bin",
|
262 |
+
"model.layers.31.post_attention_layernorm.weight": "pytorch_model-00034-of-00035.bin",
|
263 |
+
"model.layers.31.self_attn.k_proj.weight": "pytorch_model-00034-of-00035.bin",
|
264 |
+
"model.layers.31.self_attn.o_proj.weight": "pytorch_model-00034-of-00035.bin",
|
265 |
+
"model.layers.31.self_attn.q_proj.weight": "pytorch_model-00034-of-00035.bin",
|
266 |
+
"model.layers.31.self_attn.rotary_emb.inv_freq": "pytorch_model-00034-of-00035.bin",
|
267 |
+
"model.layers.31.self_attn.v_proj.weight": "pytorch_model-00034-of-00035.bin",
|
268 |
+
"model.layers.4.input_layernorm.weight": "pytorch_model-00007-of-00035.bin",
|
269 |
+
"model.layers.4.mlp.down_proj.weight": "pytorch_model-00007-of-00035.bin",
|
270 |
+
"model.layers.4.mlp.gate_proj.weight": "pytorch_model-00007-of-00035.bin",
|
271 |
+
"model.layers.4.mlp.up_proj.weight": "pytorch_model-00007-of-00035.bin",
|
272 |
+
"model.layers.4.post_attention_layernorm.weight": "pytorch_model-00007-of-00035.bin",
|
273 |
+
"model.layers.4.self_attn.k_proj.weight": "pytorch_model-00007-of-00035.bin",
|
274 |
+
"model.layers.4.self_attn.o_proj.weight": "pytorch_model-00007-of-00035.bin",
|
275 |
+
"model.layers.4.self_attn.q_proj.weight": "pytorch_model-00007-of-00035.bin",
|
276 |
+
"model.layers.4.self_attn.rotary_emb.inv_freq": "pytorch_model-00007-of-00035.bin",
|
277 |
+
"model.layers.4.self_attn.v_proj.weight": "pytorch_model-00007-of-00035.bin",
|
278 |
+
"model.layers.5.input_layernorm.weight": "pytorch_model-00008-of-00035.bin",
|
279 |
+
"model.layers.5.mlp.down_proj.weight": "pytorch_model-00008-of-00035.bin",
|
280 |
+
"model.layers.5.mlp.gate_proj.weight": "pytorch_model-00008-of-00035.bin",
|
281 |
+
"model.layers.5.mlp.up_proj.weight": "pytorch_model-00008-of-00035.bin",
|
282 |
+
"model.layers.5.post_attention_layernorm.weight": "pytorch_model-00008-of-00035.bin",
|
283 |
+
"model.layers.5.self_attn.k_proj.weight": "pytorch_model-00008-of-00035.bin",
|
284 |
+
"model.layers.5.self_attn.o_proj.weight": "pytorch_model-00008-of-00035.bin",
|
285 |
+
"model.layers.5.self_attn.q_proj.weight": "pytorch_model-00008-of-00035.bin",
|
286 |
+
"model.layers.5.self_attn.rotary_emb.inv_freq": "pytorch_model-00008-of-00035.bin",
|
287 |
+
"model.layers.5.self_attn.v_proj.weight": "pytorch_model-00008-of-00035.bin",
|
288 |
+
"model.layers.6.input_layernorm.weight": "pytorch_model-00009-of-00035.bin",
|
289 |
+
"model.layers.6.mlp.down_proj.weight": "pytorch_model-00009-of-00035.bin",
|
290 |
+
"model.layers.6.mlp.gate_proj.weight": "pytorch_model-00009-of-00035.bin",
|
291 |
+
"model.layers.6.mlp.up_proj.weight": "pytorch_model-00009-of-00035.bin",
|
292 |
+
"model.layers.6.post_attention_layernorm.weight": "pytorch_model-00009-of-00035.bin",
|
293 |
+
"model.layers.6.self_attn.k_proj.weight": "pytorch_model-00009-of-00035.bin",
|
294 |
+
"model.layers.6.self_attn.o_proj.weight": "pytorch_model-00009-of-00035.bin",
|
295 |
+
"model.layers.6.self_attn.q_proj.weight": "pytorch_model-00009-of-00035.bin",
|
296 |
+
"model.layers.6.self_attn.rotary_emb.inv_freq": "pytorch_model-00009-of-00035.bin",
|
297 |
+
"model.layers.6.self_attn.v_proj.weight": "pytorch_model-00009-of-00035.bin",
|
298 |
+
"model.layers.7.input_layernorm.weight": "pytorch_model-00010-of-00035.bin",
|
299 |
+
"model.layers.7.mlp.down_proj.weight": "pytorch_model-00010-of-00035.bin",
|
300 |
+
"model.layers.7.mlp.gate_proj.weight": "pytorch_model-00010-of-00035.bin",
|
301 |
+
"model.layers.7.mlp.up_proj.weight": "pytorch_model-00010-of-00035.bin",
|
302 |
+
"model.layers.7.post_attention_layernorm.weight": "pytorch_model-00010-of-00035.bin",
|
303 |
+
"model.layers.7.self_attn.k_proj.weight": "pytorch_model-00010-of-00035.bin",
|
304 |
+
"model.layers.7.self_attn.o_proj.weight": "pytorch_model-00010-of-00035.bin",
|
305 |
+
"model.layers.7.self_attn.q_proj.weight": "pytorch_model-00010-of-00035.bin",
|
306 |
+
"model.layers.7.self_attn.rotary_emb.inv_freq": "pytorch_model-00010-of-00035.bin",
|
307 |
+
"model.layers.7.self_attn.v_proj.weight": "pytorch_model-00010-of-00035.bin",
|
308 |
+
"model.layers.8.input_layernorm.weight": "pytorch_model-00011-of-00035.bin",
|
309 |
+
"model.layers.8.mlp.down_proj.weight": "pytorch_model-00011-of-00035.bin",
|
310 |
+
"model.layers.8.mlp.gate_proj.weight": "pytorch_model-00011-of-00035.bin",
|
311 |
+
"model.layers.8.mlp.up_proj.weight": "pytorch_model-00011-of-00035.bin",
|
312 |
+
"model.layers.8.post_attention_layernorm.weight": "pytorch_model-00011-of-00035.bin",
|
313 |
+
"model.layers.8.self_attn.k_proj.weight": "pytorch_model-00011-of-00035.bin",
|
314 |
+
"model.layers.8.self_attn.o_proj.weight": "pytorch_model-00011-of-00035.bin",
|
315 |
+
"model.layers.8.self_attn.q_proj.weight": "pytorch_model-00011-of-00035.bin",
|
316 |
+
"model.layers.8.self_attn.rotary_emb.inv_freq": "pytorch_model-00011-of-00035.bin",
|
317 |
+
"model.layers.8.self_attn.v_proj.weight": "pytorch_model-00011-of-00035.bin",
|
318 |
+
"model.layers.9.input_layernorm.weight": "pytorch_model-00012-of-00035.bin",
|
319 |
+
"model.layers.9.mlp.down_proj.weight": "pytorch_model-00012-of-00035.bin",
|
320 |
+
"model.layers.9.mlp.gate_proj.weight": "pytorch_model-00012-of-00035.bin",
|
321 |
+
"model.layers.9.mlp.up_proj.weight": "pytorch_model-00012-of-00035.bin",
|
322 |
+
"model.layers.9.post_attention_layernorm.weight": "pytorch_model-00012-of-00035.bin",
|
323 |
+
"model.layers.9.self_attn.k_proj.weight": "pytorch_model-00012-of-00035.bin",
|
324 |
+
"model.layers.9.self_attn.o_proj.weight": "pytorch_model-00012-of-00035.bin",
|
325 |
+
"model.layers.9.self_attn.q_proj.weight": "pytorch_model-00012-of-00035.bin",
|
326 |
+
"model.layers.9.self_attn.rotary_emb.inv_freq": "pytorch_model-00012-of-00035.bin",
|
327 |
+
"model.layers.9.self_attn.v_proj.weight": "pytorch_model-00012-of-00035.bin",
|
328 |
+
"model.norm.weight": "pytorch_model-00034-of-00035.bin"
|
329 |
+
}
|
330 |
+
}
|