Hugging Face's logo

Multi-Domain-Expert-Learning
/
draco

Text Generation
Transformers
PyTorch
gpt_bigcode
text-generation-inference
Inference Endpoints
Model card Files Community
Ontocord.AI commited on
Commit
4ff5d07
•
1 Parent(s): 99e3975

Upload GPTBigCodeForCausalLM

Browse files
Files changed (10) hide show
  1. config.json +39 -0
  2. generation_config.json +6 -0
  3. pytorch_model-00001-of-00007.bin +3 -0
  4. pytorch_model-00002-of-00007.bin +3 -0
  5. pytorch_model-00003-of-00007.bin +3 -0
  6. pytorch_model-00004-of-00007.bin +3 -0
  7. pytorch_model-00005-of-00007.bin +3 -0
  8. pytorch_model-00006-of-00007.bin +3 -0
  9. pytorch_model-00007-of-00007.bin +3 -0
  10. pytorch_model.bin.index.json +1176 -0
config.json ADDED
@@ -0,0 +1,39 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "HuggingFaceH4/starchat-beta",
3
+ "activation_function": "gelu",
4
+ "architectures": [
5
+ "GPTBigCodeForCausalLM"
6
+ ],
7
+ "attention_softmax_in_fp32": true,
8
+ "attn_pdrop": 0.1,
9
+ "bos_token_id": 0,
10
+ "embd_pdrop": 0.1,
11
+ "eos_token_id": 0,
12
+ "inference_runner": 0,
13
+ "initializer_range": 0.02,
14
+ "layer_norm_epsilon": 1e-05,
15
+ "max_batch_size": null,
16
+ "max_sequence_length": null,
17
+ "model_type": "gpt_bigcode",
18
+ "multi_query": true,
19
+ "n_embd": 6144,
20
+ "n_head": 48,
21
+ "n_inner": 24576,
22
+ "n_layer": 40,
23
+ "n_positions": 8192,
24
+ "pad_key_length": true,
25
+ "pre_allocate_kv_cache": false,
26
+ "resid_pdrop": 0.1,
27
+ "scale_attention_softmax_in_fp32": true,
28
+ "scale_attn_weights": true,
29
+ "summary_activation": null,
30
+ "summary_first_dropout": 0.1,
31
+ "summary_proj_to_labels": true,
32
+ "summary_type": "cls_index",
33
+ "summary_use_proj": true,
34
+ "torch_dtype": "bfloat16",
35
+ "transformers_version": "4.30.1",
36
+ "use_cache": true,
37
+ "validate_runner_input": true,
38
+ "vocab_size": 49156
39
+ }
generation_config.json ADDED
@@ -0,0 +1,6 @@
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 0,
4
+ "eos_token_id": 0,
5
+ "transformers_version": "4.30.1"
6
+ }
pytorch_model-00001-of-00007.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:74a23a51b2d6e99c2fbbc2400926d4a34c1e155f976b2c0e45b14c99fc285b16
3
+ size 9961127353
pytorch_model-00002-of-00007.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d3ce11824edc710527075457c22bb9a04b539dee091d62c2e637ace8b10f8341
3
+ size 9932923299
pytorch_model-00003-of-00007.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6129c9cb46324d496c56ee3249b8e1d0da3c9684456241d02ef7a5d8634ffde2
3
+ size 9857398311
pytorch_model-00004-of-00007.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3f842651be2ff317495736ccbec3e8184224a03bf9bcacc25798d650b06ae034
3
+ size 9857408581
pytorch_model-00005-of-00007.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f3713acba047bef991f0a0b4516946b61a80cd2abe561e6e0c626a0011e08e1c
3
+ size 9857381543
pytorch_model-00006-of-00007.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3231cc9068440b70753e6ec2910ce32f0cbc7bdef13911a8944fa0bd1f5aff9c
3
+ size 9857381479
pytorch_model-00007-of-00007.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:312b8a95f24202b3e6390514900bf6f0b98e55bda1d0db385b9f570c6e02034f
3
+ size 604043037
pytorch_model.bin.index.json ADDED
@@ -0,0 +1,1176 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metadata": {
3
+ "total_size": 59927325696
4
+ },
5
+ "weight_map": {
6
+ "base_expert.0.attn.c_attn.bias": "pytorch_model-00002-of-00007.bin",
7
+ "base_expert.0.attn.c_attn.weight": "pytorch_model-00002-of-00007.bin",
8
+ "base_expert.0.attn.c_proj.bias": "pytorch_model-00002-of-00007.bin",
9
+ "base_expert.0.attn.c_proj.weight": "pytorch_model-00002-of-00007.bin",
10
+ "base_expert.0.ln_1.bias": "pytorch_model-00002-of-00007.bin",
11
+ "base_expert.0.ln_1.weight": "pytorch_model-00002-of-00007.bin",
12
+ "base_expert.0.ln_2.bias": "pytorch_model-00002-of-00007.bin",
13
+ "base_expert.0.ln_2.weight": "pytorch_model-00002-of-00007.bin",
14
+ "base_expert.0.mlp.c_fc.bias": "pytorch_model-00002-of-00007.bin",
15
+ "base_expert.0.mlp.c_fc.weight": "pytorch_model-00002-of-00007.bin",
16
+ "base_expert.0.mlp.c_proj.bias": "pytorch_model-00002-of-00007.bin",
17
+ "base_expert.0.mlp.c_proj.weight": "pytorch_model-00002-of-00007.bin",
18
+ "base_expert.1.attn.c_attn.bias": "pytorch_model-00002-of-00007.bin",
19
+ "base_expert.1.attn.c_attn.weight": "pytorch_model-00002-of-00007.bin",
20
+ "base_expert.1.attn.c_proj.bias": "pytorch_model-00002-of-00007.bin",
21
+ "base_expert.1.attn.c_proj.weight": "pytorch_model-00002-of-00007.bin",
22
+ "base_expert.1.ln_1.bias": "pytorch_model-00002-of-00007.bin",
23
+ "base_expert.1.ln_1.weight": "pytorch_model-00002-of-00007.bin",
24
+ "base_expert.1.ln_2.bias": "pytorch_model-00002-of-00007.bin",
25
+ "base_expert.1.ln_2.weight": "pytorch_model-00002-of-00007.bin",
26
+ "base_expert.1.mlp.c_fc.bias": "pytorch_model-00002-of-00007.bin",
27
+ "base_expert.1.mlp.c_fc.weight": "pytorch_model-00002-of-00007.bin",
28
+ "base_expert.1.mlp.c_proj.bias": "pytorch_model-00002-of-00007.bin",
29
+ "base_expert.1.mlp.c_proj.weight": "pytorch_model-00002-of-00007.bin",
30
+ "base_expert.10.attn.c_attn.bias": "pytorch_model-00003-of-00007.bin",
31
+ "base_expert.10.attn.c_attn.weight": "pytorch_model-00003-of-00007.bin",
32
+ "base_expert.10.attn.c_proj.bias": "pytorch_model-00003-of-00007.bin",
33
+ "base_expert.10.attn.c_proj.weight": "pytorch_model-00003-of-00007.bin",
34
+ "base_expert.10.ln_1.bias": "pytorch_model-00003-of-00007.bin",
35
+ "base_expert.10.ln_1.weight": "pytorch_model-00003-of-00007.bin",
36
+ "base_expert.10.ln_2.bias": "pytorch_model-00003-of-00007.bin",
37
+ "base_expert.10.ln_2.weight": "pytorch_model-00003-of-00007.bin",
38
+ "base_expert.10.mlp.c_fc.bias": "pytorch_model-00003-of-00007.bin",
39
+ "base_expert.10.mlp.c_fc.weight": "pytorch_model-00003-of-00007.bin",
40
+ "base_expert.10.mlp.c_proj.bias": "pytorch_model-00003-of-00007.bin",
41
+ "base_expert.10.mlp.c_proj.weight": "pytorch_model-00003-of-00007.bin",
42
+ "base_expert.11.attn.c_attn.bias": "pytorch_model-00003-of-00007.bin",
43
+ "base_expert.11.attn.c_attn.weight": "pytorch_model-00003-of-00007.bin",
44
+ "base_expert.11.attn.c_proj.bias": "pytorch_model-00003-of-00007.bin",
45
+ "base_expert.11.attn.c_proj.weight": "pytorch_model-00003-of-00007.bin",
46
+ "base_expert.11.ln_1.bias": "pytorch_model-00003-of-00007.bin",
47
+ "base_expert.11.ln_1.weight": "pytorch_model-00003-of-00007.bin",
48
+ "base_expert.11.ln_2.bias": "pytorch_model-00003-of-00007.bin",
49
+ "base_expert.11.ln_2.weight": "pytorch_model-00003-of-00007.bin",
50
+ "base_expert.11.mlp.c_fc.bias": "pytorch_model-00003-of-00007.bin",
51
+ "base_expert.11.mlp.c_fc.weight": "pytorch_model-00003-of-00007.bin",
52
+ "base_expert.11.mlp.c_proj.bias": "pytorch_model-00003-of-00007.bin",
53
+ "base_expert.11.mlp.c_proj.weight": "pytorch_model-00003-of-00007.bin",
54
+ "base_expert.12.attn.c_attn.bias": "pytorch_model-00003-of-00007.bin",
55
+ "base_expert.12.attn.c_attn.weight": "pytorch_model-00003-of-00007.bin",
56
+ "base_expert.12.attn.c_proj.bias": "pytorch_model-00003-of-00007.bin",
57
+ "base_expert.12.attn.c_proj.weight": "pytorch_model-00003-of-00007.bin",
58
+ "base_expert.12.ln_1.bias": "pytorch_model-00003-of-00007.bin",
59
+ "base_expert.12.ln_1.weight": "pytorch_model-00003-of-00007.bin",
60
+ "base_expert.12.ln_2.bias": "pytorch_model-00003-of-00007.bin",
61
+ "base_expert.12.ln_2.weight": "pytorch_model-00003-of-00007.bin",
62
+ "base_expert.12.mlp.c_fc.bias": "pytorch_model-00003-of-00007.bin",
63
+ "base_expert.12.mlp.c_fc.weight": "pytorch_model-00003-of-00007.bin",
64
+ "base_expert.12.mlp.c_proj.bias": "pytorch_model-00003-of-00007.bin",
65
+ "base_expert.12.mlp.c_proj.weight": "pytorch_model-00003-of-00007.bin",
66
+ "base_expert.13.attn.c_attn.bias": "pytorch_model-00003-of-00007.bin",
67
+ "base_expert.13.attn.c_attn.weight": "pytorch_model-00003-of-00007.bin",
68
+ "base_expert.13.attn.c_proj.bias": "pytorch_model-00003-of-00007.bin",
69
+ "base_expert.13.attn.c_proj.weight": "pytorch_model-00003-of-00007.bin",
70
+ "base_expert.13.ln_1.bias": "pytorch_model-00003-of-00007.bin",
71
+ "base_expert.13.ln_1.weight": "pytorch_model-00003-of-00007.bin",
72
+ "base_expert.13.ln_2.bias": "pytorch_model-00003-of-00007.bin",
73
+ "base_expert.13.ln_2.weight": "pytorch_model-00003-of-00007.bin",
74
+ "base_expert.13.mlp.c_fc.bias": "pytorch_model-00003-of-00007.bin",
75
+ "base_expert.13.mlp.c_fc.weight": "pytorch_model-00003-of-00007.bin",
76
+ "base_expert.13.mlp.c_proj.bias": "pytorch_model-00003-of-00007.bin",
77
+ "base_expert.13.mlp.c_proj.weight": "pytorch_model-00003-of-00007.bin",
78
+ "base_expert.14.attn.c_attn.bias": "pytorch_model-00003-of-00007.bin",
79
+ "base_expert.14.attn.c_attn.weight": "pytorch_model-00003-of-00007.bin",
80
+ "base_expert.14.attn.c_proj.bias": "pytorch_model-00003-of-00007.bin",
81
+ "base_expert.14.attn.c_proj.weight": "pytorch_model-00003-of-00007.bin",
82
+ "base_expert.14.ln_1.bias": "pytorch_model-00003-of-00007.bin",
83
+ "base_expert.14.ln_1.weight": "pytorch_model-00003-of-00007.bin",
84
+ "base_expert.14.ln_2.bias": "pytorch_model-00003-of-00007.bin",
85
+ "base_expert.14.ln_2.weight": "pytorch_model-00003-of-00007.bin",
86
+ "base_expert.14.mlp.c_fc.bias": "pytorch_model-00003-of-00007.bin",
87
+ "base_expert.14.mlp.c_fc.weight": "pytorch_model-00003-of-00007.bin",
88
+ "base_expert.14.mlp.c_proj.bias": "pytorch_model-00003-of-00007.bin",
89
+ "base_expert.14.mlp.c_proj.weight": "pytorch_model-00003-of-00007.bin",
90
+ "base_expert.15.attn.c_attn.bias": "pytorch_model-00003-of-00007.bin",
91
+ "base_expert.15.attn.c_attn.weight": "pytorch_model-00003-of-00007.bin",
92
+ "base_expert.15.attn.c_proj.bias": "pytorch_model-00003-of-00007.bin",
93
+ "base_expert.15.attn.c_proj.weight": "pytorch_model-00003-of-00007.bin",
94
+ "base_expert.15.ln_1.bias": "pytorch_model-00003-of-00007.bin",
95
+ "base_expert.15.ln_1.weight": "pytorch_model-00003-of-00007.bin",
96
+ "base_expert.15.ln_2.bias": "pytorch_model-00003-of-00007.bin",
97
+ "base_expert.15.ln_2.weight": "pytorch_model-00003-of-00007.bin",
98
+ "base_expert.15.mlp.c_fc.bias": "pytorch_model-00003-of-00007.bin",
99
+ "base_expert.15.mlp.c_fc.weight": "pytorch_model-00003-of-00007.bin",
100
+ "base_expert.15.mlp.c_proj.bias": "pytorch_model-00003-of-00007.bin",
101
+ "base_expert.15.mlp.c_proj.weight": "pytorch_model-00003-of-00007.bin",
102
+ "base_expert.16.attn.c_attn.bias": "pytorch_model-00003-of-00007.bin",
103
+ "base_expert.16.attn.c_attn.weight": "pytorch_model-00003-of-00007.bin",
104
+ "base_expert.16.attn.c_proj.bias": "pytorch_model-00003-of-00007.bin",
105
+ "base_expert.16.attn.c_proj.weight": "pytorch_model-00003-of-00007.bin",
106
+ "base_expert.16.ln_1.bias": "pytorch_model-00003-of-00007.bin",
107
+ "base_expert.16.ln_1.weight": "pytorch_model-00003-of-00007.bin",
108
+ "base_expert.16.ln_2.bias": "pytorch_model-00003-of-00007.bin",
109
+ "base_expert.16.ln_2.weight": "pytorch_model-00003-of-00007.bin",
110
+ "base_expert.16.mlp.c_fc.bias": "pytorch_model-00003-of-00007.bin",
111
+ "base_expert.16.mlp.c_fc.weight": "pytorch_model-00003-of-00007.bin",
112
+ "base_expert.16.mlp.c_proj.bias": "pytorch_model-00003-of-00007.bin",
113
+ "base_expert.16.mlp.c_proj.weight": "pytorch_model-00003-of-00007.bin",
114
+ "base_expert.17.attn.c_attn.bias": "pytorch_model-00003-of-00007.bin",
115
+ "base_expert.17.attn.c_attn.weight": "pytorch_model-00003-of-00007.bin",
116
+ "base_expert.17.attn.c_proj.bias": "pytorch_model-00003-of-00007.bin",
117
+ "base_expert.17.attn.c_proj.weight": "pytorch_model-00003-of-00007.bin",
118
+ "base_expert.17.ln_1.bias": "pytorch_model-00003-of-00007.bin",
119
+ "base_expert.17.ln_1.weight": "pytorch_model-00003-of-00007.bin",
120
+ "base_expert.17.ln_2.bias": "pytorch_model-00003-of-00007.bin",
121
+ "base_expert.17.ln_2.weight": "pytorch_model-00003-of-00007.bin",
122
+ "base_expert.17.mlp.c_fc.bias": "pytorch_model-00004-of-00007.bin",
123
+ "base_expert.17.mlp.c_fc.weight": "pytorch_model-00004-of-00007.bin",
124
+ "base_expert.17.mlp.c_proj.bias": "pytorch_model-00004-of-00007.bin",
125
+ "base_expert.17.mlp.c_proj.weight": "pytorch_model-00004-of-00007.bin",
126
+ "base_expert.18.attn.c_attn.bias": "pytorch_model-00004-of-00007.bin",
127
+ "base_expert.18.attn.c_attn.weight": "pytorch_model-00004-of-00007.bin",
128
+ "base_expert.18.attn.c_proj.bias": "pytorch_model-00004-of-00007.bin",
129
+ "base_expert.18.attn.c_proj.weight": "pytorch_model-00004-of-00007.bin",
130
+ "base_expert.18.ln_1.bias": "pytorch_model-00004-of-00007.bin",
131
+ "base_expert.18.ln_1.weight": "pytorch_model-00004-of-00007.bin",
132
+ "base_expert.18.ln_2.bias": "pytorch_model-00004-of-00007.bin",
133
+ "base_expert.18.ln_2.weight": "pytorch_model-00004-of-00007.bin",
134
+ "base_expert.18.mlp.c_fc.bias": "pytorch_model-00004-of-00007.bin",
135
+ "base_expert.18.mlp.c_fc.weight": "pytorch_model-00004-of-00007.bin",
136
+ "base_expert.18.mlp.c_proj.bias": "pytorch_model-00004-of-00007.bin",
137
+ "base_expert.18.mlp.c_proj.weight": "pytorch_model-00004-of-00007.bin",
138
+ "base_expert.2.attn.c_attn.bias": "pytorch_model-00002-of-00007.bin",
139
+ "base_expert.2.attn.c_attn.weight": "pytorch_model-00002-of-00007.bin",
140
+ "base_expert.2.attn.c_proj.bias": "pytorch_model-00002-of-00007.bin",
141
+ "base_expert.2.attn.c_proj.weight": "pytorch_model-00002-of-00007.bin",
142
+ "base_expert.2.ln_1.bias": "pytorch_model-00002-of-00007.bin",
143
+ "base_expert.2.ln_1.weight": "pytorch_model-00002-of-00007.bin",
144
+ "base_expert.2.ln_2.bias": "pytorch_model-00002-of-00007.bin",
145
+ "base_expert.2.ln_2.weight": "pytorch_model-00002-of-00007.bin",
146
+ "base_expert.2.mlp.c_fc.bias": "pytorch_model-00002-of-00007.bin",
147
+ "base_expert.2.mlp.c_fc.weight": "pytorch_model-00002-of-00007.bin",
148
+ "base_expert.2.mlp.c_proj.bias": "pytorch_model-00002-of-00007.bin",
149
+ "base_expert.2.mlp.c_proj.weight": "pytorch_model-00002-of-00007.bin",
150
+ "base_expert.3.attn.c_attn.bias": "pytorch_model-00002-of-00007.bin",
151
+ "base_expert.3.attn.c_attn.weight": "pytorch_model-00002-of-00007.bin",
152
+ "base_expert.3.attn.c_proj.bias": "pytorch_model-00002-of-00007.bin",
153
+ "base_expert.3.attn.c_proj.weight": "pytorch_model-00002-of-00007.bin",
154
+ "base_expert.3.ln_1.bias": "pytorch_model-00002-of-00007.bin",
155
+ "base_expert.3.ln_1.weight": "pytorch_model-00002-of-00007.bin",
156
+ "base_expert.3.ln_2.bias": "pytorch_model-00002-of-00007.bin",
157
+ "base_expert.3.ln_2.weight": "pytorch_model-00002-of-00007.bin",
158
+ "base_expert.3.mlp.c_fc.bias": "pytorch_model-00002-of-00007.bin",
159
+ "base_expert.3.mlp.c_fc.weight": "pytorch_model-00002-of-00007.bin",
160
+ "base_expert.3.mlp.c_proj.bias": "pytorch_model-00002-of-00007.bin",
161
+ "base_expert.3.mlp.c_proj.weight": "pytorch_model-00002-of-00007.bin",
162
+ "base_expert.4.attn.c_attn.bias": "pytorch_model-00002-of-00007.bin",
163
+ "base_expert.4.attn.c_attn.weight": "pytorch_model-00002-of-00007.bin",
164
+ "base_expert.4.attn.c_proj.bias": "pytorch_model-00002-of-00007.bin",
165
+ "base_expert.4.attn.c_proj.weight": "pytorch_model-00002-of-00007.bin",
166
+ "base_expert.4.ln_1.bias": "pytorch_model-00002-of-00007.bin",
167
+ "base_expert.4.ln_1.weight": "pytorch_model-00002-of-00007.bin",
168
+ "base_expert.4.ln_2.bias": "pytorch_model-00002-of-00007.bin",
169
+ "base_expert.4.ln_2.weight": "pytorch_model-00002-of-00007.bin",
170
+ "base_expert.4.mlp.c_fc.bias": "pytorch_model-00003-of-00007.bin",
171
+ "base_expert.4.mlp.c_fc.weight": "pytorch_model-00003-of-00007.bin",
172
+ "base_expert.4.mlp.c_proj.bias": "pytorch_model-00003-of-00007.bin",
173
+ "base_expert.4.mlp.c_proj.weight": "pytorch_model-00003-of-00007.bin",
174
+ "base_expert.5.attn.c_attn.bias": "pytorch_model-00003-of-00007.bin",
175
+ "base_expert.5.attn.c_attn.weight": "pytorch_model-00003-of-00007.bin",
176
+ "base_expert.5.attn.c_proj.bias": "pytorch_model-00003-of-00007.bin",
177
+ "base_expert.5.attn.c_proj.weight": "pytorch_model-00003-of-00007.bin",
178
+ "base_expert.5.ln_1.bias": "pytorch_model-00003-of-00007.bin",
179
+ "base_expert.5.ln_1.weight": "pytorch_model-00003-of-00007.bin",
180
+ "base_expert.5.ln_2.bias": "pytorch_model-00003-of-00007.bin",
181
+ "base_expert.5.ln_2.weight": "pytorch_model-00003-of-00007.bin",
182
+ "base_expert.5.mlp.c_fc.bias": "pytorch_model-00003-of-00007.bin",
183
+ "base_expert.5.mlp.c_fc.weight": "pytorch_model-00003-of-00007.bin",
184
+ "base_expert.5.mlp.c_proj.bias": "pytorch_model-00003-of-00007.bin",
185
+ "base_expert.5.mlp.c_proj.weight": "pytorch_model-00003-of-00007.bin",
186
+ "base_expert.6.attn.c_attn.bias": "pytorch_model-00003-of-00007.bin",
187
+ "base_expert.6.attn.c_attn.weight": "pytorch_model-00003-of-00007.bin",
188
+ "base_expert.6.attn.c_proj.bias": "pytorch_model-00003-of-00007.bin",
189
+ "base_expert.6.attn.c_proj.weight": "pytorch_model-00003-of-00007.bin",
190
+ "base_expert.6.ln_1.bias": "pytorch_model-00003-of-00007.bin",
191
+ "base_expert.6.ln_1.weight": "pytorch_model-00003-of-00007.bin",
192
+ "base_expert.6.ln_2.bias": "pytorch_model-00003-of-00007.bin",
193
+ "base_expert.6.ln_2.weight": "pytorch_model-00003-of-00007.bin",
194
+ "base_expert.6.mlp.c_fc.bias": "pytorch_model-00003-of-00007.bin",
195
+ "base_expert.6.mlp.c_fc.weight": "pytorch_model-00003-of-00007.bin",
196
+ "base_expert.6.mlp.c_proj.bias": "pytorch_model-00003-of-00007.bin",
197
+ "base_expert.6.mlp.c_proj.weight": "pytorch_model-00003-of-00007.bin",
198
+ "base_expert.7.attn.c_attn.bias": "pytorch_model-00003-of-00007.bin",
199
+ "base_expert.7.attn.c_attn.weight": "pytorch_model-00003-of-00007.bin",
200
+ "base_expert.7.attn.c_proj.bias": "pytorch_model-00003-of-00007.bin",
201
+ "base_expert.7.attn.c_proj.weight": "pytorch_model-00003-of-00007.bin",
202
+ "base_expert.7.ln_1.bias": "pytorch_model-00003-of-00007.bin",
203
+ "base_expert.7.ln_1.weight": "pytorch_model-00003-of-00007.bin",
204
+ "base_expert.7.ln_2.bias": "pytorch_model-00003-of-00007.bin",
205
+ "base_expert.7.ln_2.weight": "pytorch_model-00003-of-00007.bin",
206
+ "base_expert.7.mlp.c_fc.bias": "pytorch_model-00003-of-00007.bin",
207
+ "base_expert.7.mlp.c_fc.weight": "pytorch_model-00003-of-00007.bin",
208
+ "base_expert.7.mlp.c_proj.bias": "pytorch_model-00003-of-00007.bin",
209
+ "base_expert.7.mlp.c_proj.weight": "pytorch_model-00003-of-00007.bin",
210
+ "base_expert.8.attn.c_attn.bias": "pytorch_model-00003-of-00007.bin",
211
+ "base_expert.8.attn.c_attn.weight": "pytorch_model-00003-of-00007.bin",
212
+ "base_expert.8.attn.c_proj.bias": "pytorch_model-00003-of-00007.bin",
213
+ "base_expert.8.attn.c_proj.weight": "pytorch_model-00003-of-00007.bin",
214
+ "base_expert.8.ln_1.bias": "pytorch_model-00003-of-00007.bin",
215
+ "base_expert.8.ln_1.weight": "pytorch_model-00003-of-00007.bin",
216
+ "base_expert.8.ln_2.bias": "pytorch_model-00003-of-00007.bin",
217
+ "base_expert.8.ln_2.weight": "pytorch_model-00003-of-00007.bin",
218
+ "base_expert.8.mlp.c_fc.bias": "pytorch_model-00003-of-00007.bin",
219
+ "base_expert.8.mlp.c_fc.weight": "pytorch_model-00003-of-00007.bin",
220
+ "base_expert.8.mlp.c_proj.bias": "pytorch_model-00003-of-00007.bin",
221
+ "base_expert.8.mlp.c_proj.weight": "pytorch_model-00003-of-00007.bin",
222
+ "base_expert.9.attn.c_attn.bias": "pytorch_model-00003-of-00007.bin",
223
+ "base_expert.9.attn.c_attn.weight": "pytorch_model-00003-of-00007.bin",
224
+ "base_expert.9.attn.c_proj.bias": "pytorch_model-00003-of-00007.bin",
225
+ "base_expert.9.attn.c_proj.weight": "pytorch_model-00003-of-00007.bin",
226
+ "base_expert.9.ln_1.bias": "pytorch_model-00003-of-00007.bin",
227
+ "base_expert.9.ln_1.weight": "pytorch_model-00003-of-00007.bin",
228
+ "base_expert.9.ln_2.bias": "pytorch_model-00003-of-00007.bin",
229
+ "base_expert.9.ln_2.weight": "pytorch_model-00003-of-00007.bin",
230
+ "base_expert.9.mlp.c_fc.bias": "pytorch_model-00003-of-00007.bin",
231
+ "base_expert.9.mlp.c_fc.weight": "pytorch_model-00003-of-00007.bin",
232
+ "base_expert.9.mlp.c_proj.bias": "pytorch_model-00003-of-00007.bin",
233
+ "base_expert.9.mlp.c_proj.weight": "pytorch_model-00003-of-00007.bin",
234
+ "books_expert.0.attn.c_attn.bias": "pytorch_model-00005-of-00007.bin",
235
+ "books_expert.0.attn.c_attn.weight": "pytorch_model-00005-of-00007.bin",
236
+ "books_expert.0.attn.c_proj.bias": "pytorch_model-00005-of-00007.bin",
237
+ "books_expert.0.attn.c_proj.weight": "pytorch_model-00005-of-00007.bin",
238
+ "books_expert.0.ln_1.bias": "pytorch_model-00005-of-00007.bin",
239
+ "books_expert.0.ln_1.weight": "pytorch_model-00005-of-00007.bin",
240
+ "books_expert.0.ln_2.bias": "pytorch_model-00005-of-00007.bin",
241
+ "books_expert.0.ln_2.weight": "pytorch_model-00005-of-00007.bin",
242
+ "books_expert.0.mlp.c_fc.bias": "pytorch_model-00005-of-00007.bin",
243
+ "books_expert.0.mlp.c_fc.weight": "pytorch_model-00005-of-00007.bin",
244
+ "books_expert.0.mlp.c_proj.bias": "pytorch_model-00005-of-00007.bin",
245
+ "books_expert.0.mlp.c_proj.weight": "pytorch_model-00005-of-00007.bin",
246
+ "books_expert.1.attn.c_attn.bias": "pytorch_model-00005-of-00007.bin",
247
+ "books_expert.1.attn.c_attn.weight": "pytorch_model-00005-of-00007.bin",
248
+ "books_expert.1.attn.c_proj.bias": "pytorch_model-00005-of-00007.bin",
249
+ "books_expert.1.attn.c_proj.weight": "pytorch_model-00005-of-00007.bin",
250
+ "books_expert.1.ln_1.bias": "pytorch_model-00005-of-00007.bin",
251
+ "books_expert.1.ln_1.weight": "pytorch_model-00005-of-00007.bin",
252
+ "books_expert.1.ln_2.bias": "pytorch_model-00005-of-00007.bin",
253
+ "books_expert.1.ln_2.weight": "pytorch_model-00005-of-00007.bin",
254
+ "books_expert.1.mlp.c_fc.bias": "pytorch_model-00005-of-00007.bin",
255
+ "books_expert.1.mlp.c_fc.weight": "pytorch_model-00005-of-00007.bin",
256
+ "books_expert.1.mlp.c_proj.bias": "pytorch_model-00005-of-00007.bin",
257
+ "books_expert.1.mlp.c_proj.weight": "pytorch_model-00005-of-00007.bin",
258
+ "books_expert.10.attn.c_attn.bias": "pytorch_model-00006-of-00007.bin",
259
+ "books_expert.10.attn.c_attn.weight": "pytorch_model-00006-of-00007.bin",
260
+ "books_expert.10.attn.c_proj.bias": "pytorch_model-00006-of-00007.bin",
261
+ "books_expert.10.attn.c_proj.weight": "pytorch_model-00006-of-00007.bin",
262
+ "books_expert.10.ln_1.bias": "pytorch_model-00006-of-00007.bin",
263
+ "books_expert.10.ln_1.weight": "pytorch_model-00006-of-00007.bin",
264
+ "books_expert.10.ln_2.bias": "pytorch_model-00006-of-00007.bin",
265
+ "books_expert.10.ln_2.weight": "pytorch_model-00006-of-00007.bin",
266
+ "books_expert.10.mlp.c_fc.bias": "pytorch_model-00006-of-00007.bin",
267
+ "books_expert.10.mlp.c_fc.weight": "pytorch_model-00006-of-00007.bin",
268
+ "books_expert.10.mlp.c_proj.bias": "pytorch_model-00006-of-00007.bin",
269
+ "books_expert.10.mlp.c_proj.weight": "pytorch_model-00006-of-00007.bin",
270
+ "books_expert.11.attn.c_attn.bias": "pytorch_model-00006-of-00007.bin",
271
+ "books_expert.11.attn.c_attn.weight": "pytorch_model-00006-of-00007.bin",
272
+ "books_expert.11.attn.c_proj.bias": "pytorch_model-00006-of-00007.bin",
273
+ "books_expert.11.attn.c_proj.weight": "pytorch_model-00006-of-00007.bin",
274
+ "books_expert.11.ln_1.bias": "pytorch_model-00006-of-00007.bin",
275
+ "books_expert.11.ln_1.weight": "pytorch_model-00006-of-00007.bin",
276
+ "books_expert.11.ln_2.bias": "pytorch_model-00006-of-00007.bin",
277
+ "books_expert.11.ln_2.weight": "pytorch_model-00006-of-00007.bin",
278
+ "books_expert.11.mlp.c_fc.bias": "pytorch_model-00006-of-00007.bin",
279
+ "books_expert.11.mlp.c_fc.weight": "pytorch_model-00006-of-00007.bin",
280
+ "books_expert.11.mlp.c_proj.bias": "pytorch_model-00006-of-00007.bin",
281
+ "books_expert.11.mlp.c_proj.weight": "pytorch_model-00006-of-00007.bin",
282
+ "books_expert.12.attn.c_attn.bias": "pytorch_model-00006-of-00007.bin",
283
+ "books_expert.12.attn.c_attn.weight": "pytorch_model-00006-of-00007.bin",
284
+ "books_expert.12.attn.c_proj.bias": "pytorch_model-00006-of-00007.bin",
285
+ "books_expert.12.attn.c_proj.weight": "pytorch_model-00006-of-00007.bin",
286
+ "books_expert.12.ln_1.bias": "pytorch_model-00006-of-00007.bin",
287
+ "books_expert.12.ln_1.weight": "pytorch_model-00006-of-00007.bin",
288
+ "books_expert.12.ln_2.bias": "pytorch_model-00006-of-00007.bin",
289
+ "books_expert.12.ln_2.weight": "pytorch_model-00006-of-00007.bin",
290
+ "books_expert.12.mlp.c_fc.bias": "pytorch_model-00006-of-00007.bin",
291
+ "books_expert.12.mlp.c_fc.weight": "pytorch_model-00006-of-00007.bin",
292
+ "books_expert.12.mlp.c_proj.bias": "pytorch_model-00006-of-00007.bin",
293
+ "books_expert.12.mlp.c_proj.weight": "pytorch_model-00006-of-00007.bin",
294
+ "books_expert.13.attn.c_attn.bias": "pytorch_model-00006-of-00007.bin",
295
+ "books_expert.13.attn.c_attn.weight": "pytorch_model-00006-of-00007.bin",
296
+ "books_expert.13.attn.c_proj.bias": "pytorch_model-00006-of-00007.bin",
297
+ "books_expert.13.attn.c_proj.weight": "pytorch_model-00006-of-00007.bin",
298
+ "books_expert.13.ln_1.bias": "pytorch_model-00006-of-00007.bin",
299
+ "books_expert.13.ln_1.weight": "pytorch_model-00006-of-00007.bin",
300
+ "books_expert.13.ln_2.bias": "pytorch_model-00006-of-00007.bin",
301
+ "books_expert.13.ln_2.weight": "pytorch_model-00006-of-00007.bin",
302
+ "books_expert.13.mlp.c_fc.bias": "pytorch_model-00006-of-00007.bin",
303
+ "books_expert.13.mlp.c_fc.weight": "pytorch_model-00006-of-00007.bin",
304
+ "books_expert.13.mlp.c_proj.bias": "pytorch_model-00006-of-00007.bin",
305
+ "books_expert.13.mlp.c_proj.weight": "pytorch_model-00006-of-00007.bin",
306
+ "books_expert.14.attn.c_attn.bias": "pytorch_model-00006-of-00007.bin",
307
+ "books_expert.14.attn.c_attn.weight": "pytorch_model-00006-of-00007.bin",
308
+ "books_expert.14.attn.c_proj.bias": "pytorch_model-00006-of-00007.bin",
309
+ "books_expert.14.attn.c_proj.weight": "pytorch_model-00006-of-00007.bin",
310
+ "books_expert.14.ln_1.bias": "pytorch_model-00006-of-00007.bin",
311
+ "books_expert.14.ln_1.weight": "pytorch_model-00006-of-00007.bin",
312
+ "books_expert.14.ln_2.bias": "pytorch_model-00006-of-00007.bin",
313
+ "books_expert.14.ln_2.weight": "pytorch_model-00006-of-00007.bin",
314
+ "books_expert.14.mlp.c_fc.bias": "pytorch_model-00006-of-00007.bin",
315
+ "books_expert.14.mlp.c_fc.weight": "pytorch_model-00006-of-00007.bin",
316
+ "books_expert.14.mlp.c_proj.bias": "pytorch_model-00006-of-00007.bin",
317
+ "books_expert.14.mlp.c_proj.weight": "pytorch_model-00006-of-00007.bin",
318
+ "books_expert.15.attn.c_attn.bias": "pytorch_model-00006-of-00007.bin",
319
+ "books_expert.15.attn.c_attn.weight": "pytorch_model-00006-of-00007.bin",
320
+ "books_expert.15.attn.c_proj.bias": "pytorch_model-00006-of-00007.bin",
321
+ "books_expert.15.attn.c_proj.weight": "pytorch_model-00006-of-00007.bin",
322
+ "books_expert.15.ln_1.bias": "pytorch_model-00006-of-00007.bin",
323
+ "books_expert.15.ln_1.weight": "pytorch_model-00006-of-00007.bin",
324
+ "books_expert.15.ln_2.bias": "pytorch_model-00006-of-00007.bin",
325
+ "books_expert.15.ln_2.weight": "pytorch_model-00006-of-00007.bin",
326
+ "books_expert.15.mlp.c_fc.bias": "pytorch_model-00006-of-00007.bin",
327
+ "books_expert.15.mlp.c_fc.weight": "pytorch_model-00006-of-00007.bin",
328
+ "books_expert.15.mlp.c_proj.bias": "pytorch_model-00006-of-00007.bin",
329
+ "books_expert.15.mlp.c_proj.weight": "pytorch_model-00006-of-00007.bin",
330
+ "books_expert.16.attn.c_attn.bias": "pytorch_model-00006-of-00007.bin",
331
+ "books_expert.16.attn.c_attn.weight": "pytorch_model-00006-of-00007.bin",
332
+ "books_expert.16.attn.c_proj.bias": "pytorch_model-00006-of-00007.bin",
333
+ "books_expert.16.attn.c_proj.weight": "pytorch_model-00006-of-00007.bin",
334
+ "books_expert.16.ln_1.bias": "pytorch_model-00006-of-00007.bin",
335
+ "books_expert.16.ln_1.weight": "pytorch_model-00006-of-00007.bin",
336
+ "books_expert.16.ln_2.bias": "pytorch_model-00006-of-00007.bin",
337
+ "books_expert.16.ln_2.weight": "pytorch_model-00006-of-00007.bin",
338
+ "books_expert.16.mlp.c_fc.bias": "pytorch_model-00006-of-00007.bin",
339
+ "books_expert.16.mlp.c_fc.weight": "pytorch_model-00006-of-00007.bin",
340
+ "books_expert.16.mlp.c_proj.bias": "pytorch_model-00006-of-00007.bin",
341
+ "books_expert.16.mlp.c_proj.weight": "pytorch_model-00006-of-00007.bin",
342
+ "books_expert.17.attn.c_attn.bias": "pytorch_model-00006-of-00007.bin",
343
+ "books_expert.17.attn.c_attn.weight": "pytorch_model-00006-of-00007.bin",
344
+ "books_expert.17.attn.c_proj.bias": "pytorch_model-00006-of-00007.bin",
345
+ "books_expert.17.attn.c_proj.weight": "pytorch_model-00006-of-00007.bin",
346
+ "books_expert.17.ln_1.bias": "pytorch_model-00006-of-00007.bin",
347
+ "books_expert.17.ln_1.weight": "pytorch_model-00006-of-00007.bin",
348
+ "books_expert.17.ln_2.bias": "pytorch_model-00006-of-00007.bin",
349
+ "books_expert.17.ln_2.weight": "pytorch_model-00006-of-00007.bin",
350
+ "books_expert.17.mlp.c_fc.bias": "pytorch_model-00006-of-00007.bin",
351
+ "books_expert.17.mlp.c_fc.weight": "pytorch_model-00006-of-00007.bin",
352
+ "books_expert.17.mlp.c_proj.bias": "pytorch_model-00006-of-00007.bin",
353
+ "books_expert.17.mlp.c_proj.weight": "pytorch_model-00006-of-00007.bin",
354
+ "books_expert.18.attn.c_attn.bias": "pytorch_model-00006-of-00007.bin",
355
+ "books_expert.18.attn.c_attn.weight": "pytorch_model-00006-of-00007.bin",
356
+ "books_expert.18.attn.c_proj.bias": "pytorch_model-00006-of-00007.bin",
357
+ "books_expert.18.attn.c_proj.weight": "pytorch_model-00006-of-00007.bin",
358
+ "books_expert.18.ln_1.bias": "pytorch_model-00006-of-00007.bin",
359
+ "books_expert.18.ln_1.weight": "pytorch_model-00006-of-00007.bin",
360
+ "books_expert.18.ln_2.bias": "pytorch_model-00006-of-00007.bin",
361
+ "books_expert.18.ln_2.weight": "pytorch_model-00006-of-00007.bin",
362
+ "books_expert.18.mlp.c_fc.bias": "pytorch_model-00007-of-00007.bin",
363
+ "books_expert.18.mlp.c_fc.weight": "pytorch_model-00007-of-00007.bin",
364
+ "books_expert.18.mlp.c_proj.bias": "pytorch_model-00007-of-00007.bin",
365
+ "books_expert.18.mlp.c_proj.weight": "pytorch_model-00007-of-00007.bin",
366
+ "books_expert.2.attn.c_attn.bias": "pytorch_model-00005-of-00007.bin",
367
+ "books_expert.2.attn.c_attn.weight": "pytorch_model-00005-of-00007.bin",
368
+ "books_expert.2.attn.c_proj.bias": "pytorch_model-00005-of-00007.bin",
369
+ "books_expert.2.attn.c_proj.weight": "pytorch_model-00005-of-00007.bin",
370
+ "books_expert.2.ln_1.bias": "pytorch_model-00005-of-00007.bin",
371
+ "books_expert.2.ln_1.weight": "pytorch_model-00005-of-00007.bin",
372
+ "books_expert.2.ln_2.bias": "pytorch_model-00005-of-00007.bin",
373
+ "books_expert.2.ln_2.weight": "pytorch_model-00005-of-00007.bin",
374
+ "books_expert.2.mlp.c_fc.bias": "pytorch_model-00005-of-00007.bin",
375
+ "books_expert.2.mlp.c_fc.weight": "pytorch_model-00005-of-00007.bin",
376
+ "books_expert.2.mlp.c_proj.bias": "pytorch_model-00005-of-00007.bin",
377
+ "books_expert.2.mlp.c_proj.weight": "pytorch_model-00005-of-00007.bin",
378
+ "books_expert.3.attn.c_attn.bias": "pytorch_model-00005-of-00007.bin",
379
+ "books_expert.3.attn.c_attn.weight": "pytorch_model-00005-of-00007.bin",
380
+ "books_expert.3.attn.c_proj.bias": "pytorch_model-00005-of-00007.bin",
381
+ "books_expert.3.attn.c_proj.weight": "pytorch_model-00005-of-00007.bin",
382
+ "books_expert.3.ln_1.bias": "pytorch_model-00005-of-00007.bin",
383
+ "books_expert.3.ln_1.weight": "pytorch_model-00005-of-00007.bin",
384
+ "books_expert.3.ln_2.bias": "pytorch_model-00005-of-00007.bin",
385
+ "books_expert.3.ln_2.weight": "pytorch_model-00005-of-00007.bin",
386
+ "books_expert.3.mlp.c_fc.bias": "pytorch_model-00005-of-00007.bin",
387
+ "books_expert.3.mlp.c_fc.weight": "pytorch_model-00005-of-00007.bin",
388
+ "books_expert.3.mlp.c_proj.bias": "pytorch_model-00005-of-00007.bin",
389
+ "books_expert.3.mlp.c_proj.weight": "pytorch_model-00005-of-00007.bin",
390
+ "books_expert.4.attn.c_attn.bias": "pytorch_model-00005-of-00007.bin",
391
+ "books_expert.4.attn.c_attn.weight": "pytorch_model-00005-of-00007.bin",
392
+ "books_expert.4.attn.c_proj.bias": "pytorch_model-00005-of-00007.bin",
393
+ "books_expert.4.attn.c_proj.weight": "pytorch_model-00005-of-00007.bin",
394
+ "books_expert.4.ln_1.bias": "pytorch_model-00005-of-00007.bin",
395
+ "books_expert.4.ln_1.weight": "pytorch_model-00005-of-00007.bin",
396
+ "books_expert.4.ln_2.bias": "pytorch_model-00005-of-00007.bin",
397
+ "books_expert.4.ln_2.weight": "pytorch_model-00005-of-00007.bin",
398
+ "books_expert.4.mlp.c_fc.bias": "pytorch_model-00005-of-00007.bin",
399
+ "books_expert.4.mlp.c_fc.weight": "pytorch_model-00005-of-00007.bin",
400
+ "books_expert.4.mlp.c_proj.bias": "pytorch_model-00005-of-00007.bin",
401
+ "books_expert.4.mlp.c_proj.weight": "pytorch_model-00005-of-00007.bin",
402
+ "books_expert.5.attn.c_attn.bias": "pytorch_model-00005-of-00007.bin",
403
+ "books_expert.5.attn.c_attn.weight": "pytorch_model-00005-of-00007.bin",
404
+ "books_expert.5.attn.c_proj.bias": "pytorch_model-00005-of-00007.bin",
405
+ "books_expert.5.attn.c_proj.weight": "pytorch_model-00005-of-00007.bin",
406
+ "books_expert.5.ln_1.bias": "pytorch_model-00005-of-00007.bin",
407
+ "books_expert.5.ln_1.weight": "pytorch_model-00005-of-00007.bin",
408
+ "books_expert.5.ln_2.bias": "pytorch_model-00005-of-00007.bin",
409
+ "books_expert.5.ln_2.weight": "pytorch_model-00005-of-00007.bin",
410
+ "books_expert.5.mlp.c_fc.bias": "pytorch_model-00006-of-00007.bin",
411
+ "books_expert.5.mlp.c_fc.weight": "pytorch_model-00006-of-00007.bin",
412
+ "books_expert.5.mlp.c_proj.bias": "pytorch_model-00006-of-00007.bin",
413
+ "books_expert.5.mlp.c_proj.weight": "pytorch_model-00006-of-00007.bin",
414
+ "books_expert.6.attn.c_attn.bias": "pytorch_model-00006-of-00007.bin",
415
+ "books_expert.6.attn.c_attn.weight": "pytorch_model-00006-of-00007.bin",
416
+ "books_expert.6.attn.c_proj.bias": "pytorch_model-00006-of-00007.bin",
417
+ "books_expert.6.attn.c_proj.weight": "pytorch_model-00006-of-00007.bin",
418
+ "books_expert.6.ln_1.bias": "pytorch_model-00006-of-00007.bin",
419
+ "books_expert.6.ln_1.weight": "pytorch_model-00006-of-00007.bin",
420
+ "books_expert.6.ln_2.bias": "pytorch_model-00006-of-00007.bin",
421
+ "books_expert.6.ln_2.weight": "pytorch_model-00006-of-00007.bin",
422
+ "books_expert.6.mlp.c_fc.bias": "pytorch_model-00006-of-00007.bin",
423
+ "books_expert.6.mlp.c_fc.weight": "pytorch_model-00006-of-00007.bin",
424
+ "books_expert.6.mlp.c_proj.bias": "pytorch_model-00006-of-00007.bin",
425
+ "books_expert.6.mlp.c_proj.weight": "pytorch_model-00006-of-00007.bin",
426
+ "books_expert.7.attn.c_attn.bias": "pytorch_model-00006-of-00007.bin",
427
+ "books_expert.7.attn.c_attn.weight": "pytorch_model-00006-of-00007.bin",
428
+ "books_expert.7.attn.c_proj.bias": "pytorch_model-00006-of-00007.bin",
429
+ "books_expert.7.attn.c_proj.weight": "pytorch_model-00006-of-00007.bin",
430
+ "books_expert.7.ln_1.bias": "pytorch_model-00006-of-00007.bin",
431
+ "books_expert.7.ln_1.weight": "pytorch_model-00006-of-00007.bin",
432
+ "books_expert.7.ln_2.bias": "pytorch_model-00006-of-00007.bin",
433
+ "books_expert.7.ln_2.weight": "pytorch_model-00006-of-00007.bin",
434
+ "books_expert.7.mlp.c_fc.bias": "pytorch_model-00006-of-00007.bin",
435
+ "books_expert.7.mlp.c_fc.weight": "pytorch_model-00006-of-00007.bin",
436
+ "books_expert.7.mlp.c_proj.bias": "pytorch_model-00006-of-00007.bin",
437
+ "books_expert.7.mlp.c_proj.weight": "pytorch_model-00006-of-00007.bin",
438
+ "books_expert.8.attn.c_attn.bias": "pytorch_model-00006-of-00007.bin",
439
+ "books_expert.8.attn.c_attn.weight": "pytorch_model-00006-of-00007.bin",
440
+ "books_expert.8.attn.c_proj.bias": "pytorch_model-00006-of-00007.bin",
441
+ "books_expert.8.attn.c_proj.weight": "pytorch_model-00006-of-00007.bin",
442
+ "books_expert.8.ln_1.bias": "pytorch_model-00006-of-00007.bin",
443
+ "books_expert.8.ln_1.weight": "pytorch_model-00006-of-00007.bin",
444
+ "books_expert.8.ln_2.bias": "pytorch_model-00006-of-00007.bin",
445
+ "books_expert.8.ln_2.weight": "pytorch_model-00006-of-00007.bin",
446
+ "books_expert.8.mlp.c_fc.bias": "pytorch_model-00006-of-00007.bin",
447
+ "books_expert.8.mlp.c_fc.weight": "pytorch_model-00006-of-00007.bin",
448
+ "books_expert.8.mlp.c_proj.bias": "pytorch_model-00006-of-00007.bin",
449
+ "books_expert.8.mlp.c_proj.weight": "pytorch_model-00006-of-00007.bin",
450
+ "books_expert.9.attn.c_attn.bias": "pytorch_model-00006-of-00007.bin",
451
+ "books_expert.9.attn.c_attn.weight": "pytorch_model-00006-of-00007.bin",
452
+ "books_expert.9.attn.c_proj.bias": "pytorch_model-00006-of-00007.bin",
453
+ "books_expert.9.attn.c_proj.weight": "pytorch_model-00006-of-00007.bin",
454
+ "books_expert.9.ln_1.bias": "pytorch_model-00006-of-00007.bin",
455
+ "books_expert.9.ln_1.weight": "pytorch_model-00006-of-00007.bin",
456
+ "books_expert.9.ln_2.bias": "pytorch_model-00006-of-00007.bin",
457
+ "books_expert.9.ln_2.weight": "pytorch_model-00006-of-00007.bin",
458
+ "books_expert.9.mlp.c_fc.bias": "pytorch_model-00006-of-00007.bin",
459
+ "books_expert.9.mlp.c_fc.weight": "pytorch_model-00006-of-00007.bin",
460
+ "books_expert.9.mlp.c_proj.bias": "pytorch_model-00006-of-00007.bin",
461
+ "books_expert.9.mlp.c_proj.weight": "pytorch_model-00006-of-00007.bin",
462
+ "github_expert.0.attn.c_attn.bias": "pytorch_model-00004-of-00007.bin",
463
+ "github_expert.0.attn.c_attn.weight": "pytorch_model-00004-of-00007.bin",
464
+ "github_expert.0.attn.c_proj.bias": "pytorch_model-00004-of-00007.bin",
465
+ "github_expert.0.attn.c_proj.weight": "pytorch_model-00004-of-00007.bin",
466
+ "github_expert.0.ln_1.bias": "pytorch_model-00004-of-00007.bin",
467
+ "github_expert.0.ln_1.weight": "pytorch_model-00004-of-00007.bin",
468
+ "github_expert.0.ln_2.bias": "pytorch_model-00004-of-00007.bin",
469
+ "github_expert.0.ln_2.weight": "pytorch_model-00004-of-00007.bin",
470
+ "github_expert.0.mlp.c_fc.bias": "pytorch_model-00004-of-00007.bin",
471
+ "github_expert.0.mlp.c_fc.weight": "pytorch_model-00004-of-00007.bin",
472
+ "github_expert.0.mlp.c_proj.bias": "pytorch_model-00004-of-00007.bin",
473
+ "github_expert.0.mlp.c_proj.weight": "pytorch_model-00004-of-00007.bin",
474
+ "github_expert.1.attn.c_attn.bias": "pytorch_model-00004-of-00007.bin",
475
+ "github_expert.1.attn.c_attn.weight": "pytorch_model-00004-of-00007.bin",
476
+ "github_expert.1.attn.c_proj.bias": "pytorch_model-00004-of-00007.bin",
477
+ "github_expert.1.attn.c_proj.weight": "pytorch_model-00004-of-00007.bin",
478
+ "github_expert.1.ln_1.bias": "pytorch_model-00004-of-00007.bin",
479
+ "github_expert.1.ln_1.weight": "pytorch_model-00004-of-00007.bin",
480
+ "github_expert.1.ln_2.bias": "pytorch_model-00004-of-00007.bin",
481
+ "github_expert.1.ln_2.weight": "pytorch_model-00004-of-00007.bin",
482
+ "github_expert.1.mlp.c_fc.bias": "pytorch_model-00004-of-00007.bin",
483
+ "github_expert.1.mlp.c_fc.weight": "pytorch_model-00004-of-00007.bin",
484
+ "github_expert.1.mlp.c_proj.bias": "pytorch_model-00004-of-00007.bin",
485
+ "github_expert.1.mlp.c_proj.weight": "pytorch_model-00004-of-00007.bin",
486
+ "github_expert.10.attn.c_attn.bias": "pytorch_model-00004-of-00007.bin",
487
+ "github_expert.10.attn.c_attn.weight": "pytorch_model-00004-of-00007.bin",
488
+ "github_expert.10.attn.c_proj.bias": "pytorch_model-00004-of-00007.bin",
489
+ "github_expert.10.attn.c_proj.weight": "pytorch_model-00004-of-00007.bin",
490
+ "github_expert.10.ln_1.bias": "pytorch_model-00004-of-00007.bin",
491
+ "github_expert.10.ln_1.weight": "pytorch_model-00004-of-00007.bin",
492
+ "github_expert.10.ln_2.bias": "pytorch_model-00004-of-00007.bin",
493
+ "github_expert.10.ln_2.weight": "pytorch_model-00004-of-00007.bin",
494
+ "github_expert.10.mlp.c_fc.bias": "pytorch_model-00004-of-00007.bin",
495
+ "github_expert.10.mlp.c_fc.weight": "pytorch_model-00004-of-00007.bin",
496
+ "github_expert.10.mlp.c_proj.bias": "pytorch_model-00004-of-00007.bin",
497
+ "github_expert.10.mlp.c_proj.weight": "pytorch_model-00004-of-00007.bin",
498
+ "github_expert.11.attn.c_attn.bias": "pytorch_model-00004-of-00007.bin",
499
+ "github_expert.11.attn.c_attn.weight": "pytorch_model-00004-of-00007.bin",
500
+ "github_expert.11.attn.c_proj.bias": "pytorch_model-00004-of-00007.bin",
501
+ "github_expert.11.attn.c_proj.weight": "pytorch_model-00004-of-00007.bin",
502
+ "github_expert.11.ln_1.bias": "pytorch_model-00004-of-00007.bin",
503
+ "github_expert.11.ln_1.weight": "pytorch_model-00004-of-00007.bin",
504
+ "github_expert.11.ln_2.bias": "pytorch_model-00004-of-00007.bin",
505
+ "github_expert.11.ln_2.weight": "pytorch_model-00004-of-00007.bin",
506
+ "github_expert.11.mlp.c_fc.bias": "pytorch_model-00005-of-00007.bin",
507
+ "github_expert.11.mlp.c_fc.weight": "pytorch_model-00005-of-00007.bin",
508
+ "github_expert.11.mlp.c_proj.bias": "pytorch_model-00005-of-00007.bin",
509
+ "github_expert.11.mlp.c_proj.weight": "pytorch_model-00005-of-00007.bin",
510
+ "github_expert.12.attn.c_attn.bias": "pytorch_model-00005-of-00007.bin",
511
+ "github_expert.12.attn.c_attn.weight": "pytorch_model-00005-of-00007.bin",
512
+ "github_expert.12.attn.c_proj.bias": "pytorch_model-00005-of-00007.bin",
513
+ "github_expert.12.attn.c_proj.weight": "pytorch_model-00005-of-00007.bin",
514
+ "github_expert.12.ln_1.bias": "pytorch_model-00005-of-00007.bin",
515
+ "github_expert.12.ln_1.weight": "pytorch_model-00005-of-00007.bin",
516
+ "github_expert.12.ln_2.bias": "pytorch_model-00005-of-00007.bin",
517
+ "github_expert.12.ln_2.weight": "pytorch_model-00005-of-00007.bin",
518
+ "github_expert.12.mlp.c_fc.bias": "pytorch_model-00005-of-00007.bin",
519
+ "github_expert.12.mlp.c_fc.weight": "pytorch_model-00005-of-00007.bin",
520
+ "github_expert.12.mlp.c_proj.bias": "pytorch_model-00005-of-00007.bin",
521
+ "github_expert.12.mlp.c_proj.weight": "pytorch_model-00005-of-00007.bin",
522
+ "github_expert.13.attn.c_attn.bias": "pytorch_model-00005-of-00007.bin",
523
+ "github_expert.13.attn.c_attn.weight": "pytorch_model-00005-of-00007.bin",
524
+ "github_expert.13.attn.c_proj.bias": "pytorch_model-00005-of-00007.bin",
525
+ "github_expert.13.attn.c_proj.weight": "pytorch_model-00005-of-00007.bin",
526
+ "github_expert.13.ln_1.bias": "pytorch_model-00005-of-00007.bin",
527
+ "github_expert.13.ln_1.weight": "pytorch_model-00005-of-00007.bin",
528
+ "github_expert.13.ln_2.bias": "pytorch_model-00005-of-00007.bin",
529
+ "github_expert.13.ln_2.weight": "pytorch_model-00005-of-00007.bin",
530
+ "github_expert.13.mlp.c_fc.bias": "pytorch_model-00005-of-00007.bin",
531
+ "github_expert.13.mlp.c_fc.weight": "pytorch_model-00005-of-00007.bin",
532
+ "github_expert.13.mlp.c_proj.bias": "pytorch_model-00005-of-00007.bin",
533
+ "github_expert.13.mlp.c_proj.weight": "pytorch_model-00005-of-00007.bin",
534
+ "github_expert.14.attn.c_attn.bias": "pytorch_model-00005-of-00007.bin",
535
+ "github_expert.14.attn.c_attn.weight": "pytorch_model-00005-of-00007.bin",
536
+ "github_expert.14.attn.c_proj.bias": "pytorch_model-00005-of-00007.bin",
537
+ "github_expert.14.attn.c_proj.weight": "pytorch_model-00005-of-00007.bin",
538
+ "github_expert.14.ln_1.bias": "pytorch_model-00005-of-00007.bin",
539
+ "github_expert.14.ln_1.weight": "pytorch_model-00005-of-00007.bin",
540
+ "github_expert.14.ln_2.bias": "pytorch_model-00005-of-00007.bin",
541
+ "github_expert.14.ln_2.weight": "pytorch_model-00005-of-00007.bin",
542
+ "github_expert.14.mlp.c_fc.bias": "pytorch_model-00005-of-00007.bin",
543
+ "github_expert.14.mlp.c_fc.weight": "pytorch_model-00005-of-00007.bin",
544
+ "github_expert.14.mlp.c_proj.bias": "pytorch_model-00005-of-00007.bin",
545
+ "github_expert.14.mlp.c_proj.weight": "pytorch_model-00005-of-00007.bin",
546
+ "github_expert.15.attn.c_attn.bias": "pytorch_model-00005-of-00007.bin",
547
+ "github_expert.15.attn.c_attn.weight": "pytorch_model-00005-of-00007.bin",
548
+ "github_expert.15.attn.c_proj.bias": "pytorch_model-00005-of-00007.bin",
549
+ "github_expert.15.attn.c_proj.weight": "pytorch_model-00005-of-00007.bin",
550
+ "github_expert.15.ln_1.bias": "pytorch_model-00005-of-00007.bin",
551
+ "github_expert.15.ln_1.weight": "pytorch_model-00005-of-00007.bin",
552
+ "github_expert.15.ln_2.bias": "pytorch_model-00005-of-00007.bin",
553
+ "github_expert.15.ln_2.weight": "pytorch_model-00005-of-00007.bin",
554
+ "github_expert.15.mlp.c_fc.bias": "pytorch_model-00005-of-00007.bin",
555
+ "github_expert.15.mlp.c_fc.weight": "pytorch_model-00005-of-00007.bin",
556
+ "github_expert.15.mlp.c_proj.bias": "pytorch_model-00005-of-00007.bin",
557
+ "github_expert.15.mlp.c_proj.weight": "pytorch_model-00005-of-00007.bin",
558
+ "github_expert.16.attn.c_attn.bias": "pytorch_model-00005-of-00007.bin",
559
+ "github_expert.16.attn.c_attn.weight": "pytorch_model-00005-of-00007.bin",
560
+ "github_expert.16.attn.c_proj.bias": "pytorch_model-00005-of-00007.bin",
561
+ "github_expert.16.attn.c_proj.weight": "pytorch_model-00005-of-00007.bin",
562
+ "github_expert.16.ln_1.bias": "pytorch_model-00005-of-00007.bin",
563
+ "github_expert.16.ln_1.weight": "pytorch_model-00005-of-00007.bin",
564
+ "github_expert.16.ln_2.bias": "pytorch_model-00005-of-00007.bin",
565
+ "github_expert.16.ln_2.weight": "pytorch_model-00005-of-00007.bin",
566
+ "github_expert.16.mlp.c_fc.bias": "pytorch_model-00005-of-00007.bin",
567
+ "github_expert.16.mlp.c_fc.weight": "pytorch_model-00005-of-00007.bin",
568
+ "github_expert.16.mlp.c_proj.bias": "pytorch_model-00005-of-00007.bin",
569
+ "github_expert.16.mlp.c_proj.weight": "pytorch_model-00005-of-00007.bin",
570
+ "github_expert.17.attn.c_attn.bias": "pytorch_model-00005-of-00007.bin",
571
+ "github_expert.17.attn.c_attn.weight": "pytorch_model-00005-of-00007.bin",
572
+ "github_expert.17.attn.c_proj.bias": "pytorch_model-00005-of-00007.bin",
573
+ "github_expert.17.attn.c_proj.weight": "pytorch_model-00005-of-00007.bin",
574
+ "github_expert.17.ln_1.bias": "pytorch_model-00005-of-00007.bin",
575
+ "github_expert.17.ln_1.weight": "pytorch_model-00005-of-00007.bin",
576
+ "github_expert.17.ln_2.bias": "pytorch_model-00005-of-00007.bin",
577
+ "github_expert.17.ln_2.weight": "pytorch_model-00005-of-00007.bin",
578
+ "github_expert.17.mlp.c_fc.bias": "pytorch_model-00005-of-00007.bin",
579
+ "github_expert.17.mlp.c_fc.weight": "pytorch_model-00005-of-00007.bin",
580
+ "github_expert.17.mlp.c_proj.bias": "pytorch_model-00005-of-00007.bin",
581
+ "github_expert.17.mlp.c_proj.weight": "pytorch_model-00005-of-00007.bin",
582
+ "github_expert.18.attn.c_attn.bias": "pytorch_model-00005-of-00007.bin",
583
+ "github_expert.18.attn.c_attn.weight": "pytorch_model-00005-of-00007.bin",
584
+ "github_expert.18.attn.c_proj.bias": "pytorch_model-00005-of-00007.bin",
585
+ "github_expert.18.attn.c_proj.weight": "pytorch_model-00005-of-00007.bin",
586
+ "github_expert.18.ln_1.bias": "pytorch_model-00005-of-00007.bin",
587
+ "github_expert.18.ln_1.weight": "pytorch_model-00005-of-00007.bin",
588
+ "github_expert.18.ln_2.bias": "pytorch_model-00005-of-00007.bin",
589
+ "github_expert.18.ln_2.weight": "pytorch_model-00005-of-00007.bin",
590
+ "github_expert.18.mlp.c_fc.bias": "pytorch_model-00005-of-00007.bin",
591
+ "github_expert.18.mlp.c_fc.weight": "pytorch_model-00005-of-00007.bin",
592
+ "github_expert.18.mlp.c_proj.bias": "pytorch_model-00005-of-00007.bin",
593
+ "github_expert.18.mlp.c_proj.weight": "pytorch_model-00005-of-00007.bin",
594
+ "github_expert.2.attn.c_attn.bias": "pytorch_model-00004-of-00007.bin",
595
+ "github_expert.2.attn.c_attn.weight": "pytorch_model-00004-of-00007.bin",
596
+ "github_expert.2.attn.c_proj.bias": "pytorch_model-00004-of-00007.bin",
597
+ "github_expert.2.attn.c_proj.weight": "pytorch_model-00004-of-00007.bin",
598
+ "github_expert.2.ln_1.bias": "pytorch_model-00004-of-00007.bin",
599
+ "github_expert.2.ln_1.weight": "pytorch_model-00004-of-00007.bin",
600
+ "github_expert.2.ln_2.bias": "pytorch_model-00004-of-00007.bin",
601
+ "github_expert.2.ln_2.weight": "pytorch_model-00004-of-00007.bin",
602
+ "github_expert.2.mlp.c_fc.bias": "pytorch_model-00004-of-00007.bin",
603
+ "github_expert.2.mlp.c_fc.weight": "pytorch_model-00004-of-00007.bin",
604
+ "github_expert.2.mlp.c_proj.bias": "pytorch_model-00004-of-00007.bin",
605
+ "github_expert.2.mlp.c_proj.weight": "pytorch_model-00004-of-00007.bin",
606
+ "github_expert.3.attn.c_attn.bias": "pytorch_model-00004-of-00007.bin",
607
+ "github_expert.3.attn.c_attn.weight": "pytorch_model-00004-of-00007.bin",
608
+ "github_expert.3.attn.c_proj.bias": "pytorch_model-00004-of-00007.bin",
609
+ "github_expert.3.attn.c_proj.weight": "pytorch_model-00004-of-00007.bin",
610
+ "github_expert.3.ln_1.bias": "pytorch_model-00004-of-00007.bin",
611
+ "github_expert.3.ln_1.weight": "pytorch_model-00004-of-00007.bin",
612
+ "github_expert.3.ln_2.bias": "pytorch_model-00004-of-00007.bin",
613
+ "github_expert.3.ln_2.weight": "pytorch_model-00004-of-00007.bin",
614
+ "github_expert.3.mlp.c_fc.bias": "pytorch_model-00004-of-00007.bin",
615
+ "github_expert.3.mlp.c_fc.weight": "pytorch_model-00004-of-00007.bin",
616
+ "github_expert.3.mlp.c_proj.bias": "pytorch_model-00004-of-00007.bin",
617
+ "github_expert.3.mlp.c_proj.weight": "pytorch_model-00004-of-00007.bin",
618
+ "github_expert.4.attn.c_attn.bias": "pytorch_model-00004-of-00007.bin",
619
+ "github_expert.4.attn.c_attn.weight": "pytorch_model-00004-of-00007.bin",
620
+ "github_expert.4.attn.c_proj.bias": "pytorch_model-00004-of-00007.bin",
621
+ "github_expert.4.attn.c_proj.weight": "pytorch_model-00004-of-00007.bin",
622
+ "github_expert.4.ln_1.bias": "pytorch_model-00004-of-00007.bin",
623
+ "github_expert.4.ln_1.weight": "pytorch_model-00004-of-00007.bin",
624
+ "github_expert.4.ln_2.bias": "pytorch_model-00004-of-00007.bin",
625
+ "github_expert.4.ln_2.weight": "pytorch_model-00004-of-00007.bin",
626
+ "github_expert.4.mlp.c_fc.bias": "pytorch_model-00004-of-00007.bin",
627
+ "github_expert.4.mlp.c_fc.weight": "pytorch_model-00004-of-00007.bin",
628
+ "github_expert.4.mlp.c_proj.bias": "pytorch_model-00004-of-00007.bin",
629
+ "github_expert.4.mlp.c_proj.weight": "pytorch_model-00004-of-00007.bin",
630
+ "github_expert.5.attn.c_attn.bias": "pytorch_model-00004-of-00007.bin",
631
+ "github_expert.5.attn.c_attn.weight": "pytorch_model-00004-of-00007.bin",
632
+ "github_expert.5.attn.c_proj.bias": "pytorch_model-00004-of-00007.bin",
633
+ "github_expert.5.attn.c_proj.weight": "pytorch_model-00004-of-00007.bin",
634
+ "github_expert.5.ln_1.bias": "pytorch_model-00004-of-00007.bin",
635
+ "github_expert.5.ln_1.weight": "pytorch_model-00004-of-00007.bin",
636
+ "github_expert.5.ln_2.bias": "pytorch_model-00004-of-00007.bin",
637
+ "github_expert.5.ln_2.weight": "pytorch_model-00004-of-00007.bin",
638
+ "github_expert.5.mlp.c_fc.bias": "pytorch_model-00004-of-00007.bin",
639
+ "github_expert.5.mlp.c_fc.weight": "pytorch_model-00004-of-00007.bin",
640
+ "github_expert.5.mlp.c_proj.bias": "pytorch_model-00004-of-00007.bin",
641
+ "github_expert.5.mlp.c_proj.weight": "pytorch_model-00004-of-00007.bin",
642
+ "github_expert.6.attn.c_attn.bias": "pytorch_model-00004-of-00007.bin",
643
+ "github_expert.6.attn.c_attn.weight": "pytorch_model-00004-of-00007.bin",
644
+ "github_expert.6.attn.c_proj.bias": "pytorch_model-00004-of-00007.bin",
645
+ "github_expert.6.attn.c_proj.weight": "pytorch_model-00004-of-00007.bin",
646
+ "github_expert.6.ln_1.bias": "pytorch_model-00004-of-00007.bin",
647
+ "github_expert.6.ln_1.weight": "pytorch_model-00004-of-00007.bin",
648
+ "github_expert.6.ln_2.bias": "pytorch_model-00004-of-00007.bin",
649
+ "github_expert.6.ln_2.weight": "pytorch_model-00004-of-00007.bin",
650
+ "github_expert.6.mlp.c_fc.bias": "pytorch_model-00004-of-00007.bin",
651
+ "github_expert.6.mlp.c_fc.weight": "pytorch_model-00004-of-00007.bin",
652
+ "github_expert.6.mlp.c_proj.bias": "pytorch_model-00004-of-00007.bin",
653
+ "github_expert.6.mlp.c_proj.weight": "pytorch_model-00004-of-00007.bin",
654
+ "github_expert.7.attn.c_attn.bias": "pytorch_model-00004-of-00007.bin",
655
+ "github_expert.7.attn.c_attn.weight": "pytorch_model-00004-of-00007.bin",
656
+ "github_expert.7.attn.c_proj.bias": "pytorch_model-00004-of-00007.bin",
657
+ "github_expert.7.attn.c_proj.weight": "pytorch_model-00004-of-00007.bin",
658
+ "github_expert.7.ln_1.bias": "pytorch_model-00004-of-00007.bin",
659
+ "github_expert.7.ln_1.weight": "pytorch_model-00004-of-00007.bin",
660
+ "github_expert.7.ln_2.bias": "pytorch_model-00004-of-00007.bin",
661
+ "github_expert.7.ln_2.weight": "pytorch_model-00004-of-00007.bin",
662
+ "github_expert.7.mlp.c_fc.bias": "pytorch_model-00004-of-00007.bin",
663
+ "github_expert.7.mlp.c_fc.weight": "pytorch_model-00004-of-00007.bin",
664
+ "github_expert.7.mlp.c_proj.bias": "pytorch_model-00004-of-00007.bin",
665
+ "github_expert.7.mlp.c_proj.weight": "pytorch_model-00004-of-00007.bin",
666
+ "github_expert.8.attn.c_attn.bias": "pytorch_model-00004-of-00007.bin",
667
+ "github_expert.8.attn.c_attn.weight": "pytorch_model-00004-of-00007.bin",
668
+ "github_expert.8.attn.c_proj.bias": "pytorch_model-00004-of-00007.bin",
669
+ "github_expert.8.attn.c_proj.weight": "pytorch_model-00004-of-00007.bin",
670
+ "github_expert.8.ln_1.bias": "pytorch_model-00004-of-00007.bin",
671
+ "github_expert.8.ln_1.weight": "pytorch_model-00004-of-00007.bin",
672
+ "github_expert.8.ln_2.bias": "pytorch_model-00004-of-00007.bin",
673
+ "github_expert.8.ln_2.weight": "pytorch_model-00004-of-00007.bin",
674
+ "github_expert.8.mlp.c_fc.bias": "pytorch_model-00004-of-00007.bin",
675
+ "github_expert.8.mlp.c_fc.weight": "pytorch_model-00004-of-00007.bin",
676
+ "github_expert.8.mlp.c_proj.bias": "pytorch_model-00004-of-00007.bin",
677
+ "github_expert.8.mlp.c_proj.weight": "pytorch_model-00004-of-00007.bin",
678
+ "github_expert.9.attn.c_attn.bias": "pytorch_model-00004-of-00007.bin",
679
+ "github_expert.9.attn.c_attn.weight": "pytorch_model-00004-of-00007.bin",
680
+ "github_expert.9.attn.c_proj.bias": "pytorch_model-00004-of-00007.bin",
681
+ "github_expert.9.attn.c_proj.weight": "pytorch_model-00004-of-00007.bin",
682
+ "github_expert.9.ln_1.bias": "pytorch_model-00004-of-00007.bin",
683
+ "github_expert.9.ln_1.weight": "pytorch_model-00004-of-00007.bin",
684
+ "github_expert.9.ln_2.bias": "pytorch_model-00004-of-00007.bin",
685
+ "github_expert.9.ln_2.weight": "pytorch_model-00004-of-00007.bin",
686
+ "github_expert.9.mlp.c_fc.bias": "pytorch_model-00004-of-00007.bin",
687
+ "github_expert.9.mlp.c_fc.weight": "pytorch_model-00004-of-00007.bin",
688
+ "github_expert.9.mlp.c_proj.bias": "pytorch_model-00004-of-00007.bin",
689
+ "github_expert.9.mlp.c_proj.weight": "pytorch_model-00004-of-00007.bin",
690
+ "lm_head.weight": "pytorch_model-00001-of-00007.bin",
691
+ "transformer.h.0.attn.c_attn.bias": "pytorch_model-00001-of-00007.bin",
692
+ "transformer.h.0.attn.c_attn.weight": "pytorch_model-00001-of-00007.bin",
693
+ "transformer.h.0.attn.c_proj.bias": "pytorch_model-00001-of-00007.bin",
694
+ "transformer.h.0.attn.c_proj.weight": "pytorch_model-00001-of-00007.bin",
695
+ "transformer.h.0.ln_1.bias": "pytorch_model-00001-of-00007.bin",
696
+ "transformer.h.0.ln_1.weight": "pytorch_model-00001-of-00007.bin",
697
+ "transformer.h.0.ln_2.bias": "pytorch_model-00001-of-00007.bin",
698
+ "transformer.h.0.ln_2.weight": "pytorch_model-00001-of-00007.bin",
699
+ "transformer.h.0.mlp.c_fc.bias": "pytorch_model-00001-of-00007.bin",
700
+ "transformer.h.0.mlp.c_fc.weight": "pytorch_model-00001-of-00007.bin",
701
+ "transformer.h.0.mlp.c_proj.bias": "pytorch_model-00001-of-00007.bin",
702
+ "transformer.h.0.mlp.c_proj.weight": "pytorch_model-00001-of-00007.bin",
703
+ "transformer.h.1.attn.c_attn.bias": "pytorch_model-00001-of-00007.bin",
704
+ "transformer.h.1.attn.c_attn.weight": "pytorch_model-00001-of-00007.bin",
705
+ "transformer.h.1.attn.c_proj.bias": "pytorch_model-00001-of-00007.bin",
706
+ "transformer.h.1.attn.c_proj.weight": "pytorch_model-00001-of-00007.bin",
707
+ "transformer.h.1.ln_1.bias": "pytorch_model-00001-of-00007.bin",
708
+ "transformer.h.1.ln_1.weight": "pytorch_model-00001-of-00007.bin",
709
+ "transformer.h.1.ln_2.bias": "pytorch_model-00001-of-00007.bin",
710
+ "transformer.h.1.ln_2.weight": "pytorch_model-00001-of-00007.bin",
711
+ "transformer.h.1.mlp.c_fc.bias": "pytorch_model-00001-of-00007.bin",
712
+ "transformer.h.1.mlp.c_fc.weight": "pytorch_model-00001-of-00007.bin",
713
+ "transformer.h.1.mlp.c_proj.bias": "pytorch_model-00001-of-00007.bin",
714
+ "transformer.h.1.mlp.c_proj.weight": "pytorch_model-00001-of-00007.bin",
715
+ "transformer.h.10.attn.c_attn.bias": "pytorch_model-00001-of-00007.bin",
716
+ "transformer.h.10.attn.c_attn.weight": "pytorch_model-00001-of-00007.bin",
717
+ "transformer.h.10.attn.c_proj.bias": "pytorch_model-00001-of-00007.bin",
718
+ "transformer.h.10.attn.c_proj.weight": "pytorch_model-00001-of-00007.bin",
719
+ "transformer.h.10.ln_1.bias": "pytorch_model-00001-of-00007.bin",
720
+ "transformer.h.10.ln_1.weight": "pytorch_model-00001-of-00007.bin",
721
+ "transformer.h.10.ln_2.bias": "pytorch_model-00001-of-00007.bin",
722
+ "transformer.h.10.ln_2.weight": "pytorch_model-00001-of-00007.bin",
723
+ "transformer.h.10.mlp.c_fc.bias": "pytorch_model-00001-of-00007.bin",
724
+ "transformer.h.10.mlp.c_fc.weight": "pytorch_model-00001-of-00007.bin",
725
+ "transformer.h.10.mlp.c_proj.bias": "pytorch_model-00001-of-00007.bin",
726
+ "transformer.h.10.mlp.c_proj.weight": "pytorch_model-00001-of-00007.bin",
727
+ "transformer.h.11.attn.c_attn.bias": "pytorch_model-00001-of-00007.bin",
728
+ "transformer.h.11.attn.c_attn.weight": "pytorch_model-00001-of-00007.bin",
729
+ "transformer.h.11.attn.c_proj.bias": "pytorch_model-00001-of-00007.bin",
730
+ "transformer.h.11.attn.c_proj.weight": "pytorch_model-00001-of-00007.bin",
731
+ "transformer.h.11.ln_1.bias": "pytorch_model-00001-of-00007.bin",
732
+ "transformer.h.11.ln_1.weight": "pytorch_model-00001-of-00007.bin",
733
+ "transformer.h.11.ln_2.bias": "pytorch_model-00001-of-00007.bin",
734
+ "transformer.h.11.ln_2.weight": "pytorch_model-00001-of-00007.bin",
735
+ "transformer.h.11.mlp.c_fc.bias": "pytorch_model-00001-of-00007.bin",
736
+ "transformer.h.11.mlp.c_fc.weight": "pytorch_model-00001-of-00007.bin",
737
+ "transformer.h.11.mlp.c_proj.bias": "pytorch_model-00001-of-00007.bin",
738
+ "transformer.h.11.mlp.c_proj.weight": "pytorch_model-00001-of-00007.bin",
739
+ "transformer.h.12.attn.c_attn.bias": "pytorch_model-00001-of-00007.bin",
740
+ "transformer.h.12.attn.c_attn.weight": "pytorch_model-00001-of-00007.bin",
741
+ "transformer.h.12.attn.c_proj.bias": "pytorch_model-00002-of-00007.bin",
742
+ "transformer.h.12.attn.c_proj.weight": "pytorch_model-00002-of-00007.bin",
743
+ "transformer.h.12.ln_1.bias": "pytorch_model-00001-of-00007.bin",
744
+ "transformer.h.12.ln_1.weight": "pytorch_model-00001-of-00007.bin",
745
+ "transformer.h.12.ln_2.bias": "pytorch_model-00002-of-00007.bin",
746
+ "transformer.h.12.ln_2.weight": "pytorch_model-00002-of-00007.bin",
747
+ "transformer.h.12.mlp.c_fc.bias": "pytorch_model-00002-of-00007.bin",
748
+ "transformer.h.12.mlp.c_fc.weight": "pytorch_model-00002-of-00007.bin",
749
+ "transformer.h.12.mlp.c_proj.bias": "pytorch_model-00002-of-00007.bin",
750
+ "transformer.h.12.mlp.c_proj.weight": "pytorch_model-00002-of-00007.bin",
751
+ "transformer.h.13.attn.c_attn.bias": "pytorch_model-00002-of-00007.bin",
752
+ "transformer.h.13.attn.c_attn.weight": "pytorch_model-00002-of-00007.bin",
753
+ "transformer.h.13.attn.c_proj.bias": "pytorch_model-00002-of-00007.bin",
754
+ "transformer.h.13.attn.c_proj.weight": "pytorch_model-00002-of-00007.bin",
755
+ "transformer.h.13.ln_1.bias": "pytorch_model-00002-of-00007.bin",
756
+ "transformer.h.13.ln_1.weight": "pytorch_model-00002-of-00007.bin",
757
+ "transformer.h.13.ln_2.bias": "pytorch_model-00002-of-00007.bin",
758
+ "transformer.h.13.ln_2.weight": "pytorch_model-00002-of-00007.bin",
759
+ "transformer.h.13.mlp.c_fc.bias": "pytorch_model-00002-of-00007.bin",
760
+ "transformer.h.13.mlp.c_fc.weight": "pytorch_model-00002-of-00007.bin",
761
+ "transformer.h.13.mlp.c_proj.bias": "pytorch_model-00002-of-00007.bin",
762
+ "transformer.h.13.mlp.c_proj.weight": "pytorch_model-00002-of-00007.bin",
763
+ "transformer.h.14.attn.c_attn.bias": "pytorch_model-00002-of-00007.bin",
764
+ "transformer.h.14.attn.c_attn.weight": "pytorch_model-00002-of-00007.bin",
765
+ "transformer.h.14.attn.c_proj.bias": "pytorch_model-00002-of-00007.bin",
766
+ "transformer.h.14.attn.c_proj.weight": "pytorch_model-00002-of-00007.bin",
767
+ "transformer.h.14.ln_1.bias": "pytorch_model-00002-of-00007.bin",
768
+ "transformer.h.14.ln_1.weight": "pytorch_model-00002-of-00007.bin",
769
+ "transformer.h.14.ln_2.bias": "pytorch_model-00002-of-00007.bin",
770
+ "transformer.h.14.ln_2.weight": "pytorch_model-00002-of-00007.bin",
771
+ "transformer.h.14.mlp.c_fc.bias": "pytorch_model-00002-of-00007.bin",
772
+ "transformer.h.14.mlp.c_fc.weight": "pytorch_model-00002-of-00007.bin",
773
+ "transformer.h.14.mlp.c_proj.bias": "pytorch_model-00002-of-00007.bin",
774
+ "transformer.h.14.mlp.c_proj.weight": "pytorch_model-00002-of-00007.bin",
775
+ "transformer.h.15.attn.c_attn.bias": "pytorch_model-00002-of-00007.bin",
776
+ "transformer.h.15.attn.c_attn.weight": "pytorch_model-00002-of-00007.bin",
777
+ "transformer.h.15.attn.c_proj.bias": "pytorch_model-00002-of-00007.bin",
778
+ "transformer.h.15.attn.c_proj.weight": "pytorch_model-00002-of-00007.bin",
779
+ "transformer.h.15.ln_1.bias": "pytorch_model-00002-of-00007.bin",
780
+ "transformer.h.15.ln_1.weight": "pytorch_model-00002-of-00007.bin",
781
+ "transformer.h.15.ln_2.bias": "pytorch_model-00002-of-00007.bin",
782
+ "transformer.h.15.ln_2.weight": "pytorch_model-00002-of-00007.bin",
783
+ "transformer.h.15.mlp.c_fc.bias": "pytorch_model-00002-of-00007.bin",
784
+ "transformer.h.15.mlp.c_fc.weight": "pytorch_model-00002-of-00007.bin",
785
+ "transformer.h.15.mlp.c_proj.bias": "pytorch_model-00002-of-00007.bin",
786
+ "transformer.h.15.mlp.c_proj.weight": "pytorch_model-00002-of-00007.bin",
787
+ "transformer.h.16.attn.c_attn.bias": "pytorch_model-00002-of-00007.bin",
788
+ "transformer.h.16.attn.c_attn.weight": "pytorch_model-00002-of-00007.bin",
789
+ "transformer.h.16.attn.c_proj.bias": "pytorch_model-00002-of-00007.bin",
790
+ "transformer.h.16.attn.c_proj.weight": "pytorch_model-00002-of-00007.bin",
791
+ "transformer.h.16.ln_1.bias": "pytorch_model-00002-of-00007.bin",
792
+ "transformer.h.16.ln_1.weight": "pytorch_model-00002-of-00007.bin",
793
+ "transformer.h.16.ln_2.bias": "pytorch_model-00002-of-00007.bin",
794
+ "transformer.h.16.ln_2.weight": "pytorch_model-00002-of-00007.bin",
795
+ "transformer.h.16.mlp.c_fc.bias": "pytorch_model-00002-of-00007.bin",
796
+ "transformer.h.16.mlp.c_fc.weight": "pytorch_model-00002-of-00007.bin",
797
+ "transformer.h.16.mlp.c_proj.bias": "pytorch_model-00002-of-00007.bin",
798
+ "transformer.h.16.mlp.c_proj.weight": "pytorch_model-00002-of-00007.bin",
799
+ "transformer.h.17.attn.c_attn.bias": "pytorch_model-00002-of-00007.bin",
800
+ "transformer.h.17.attn.c_attn.weight": "pytorch_model-00002-of-00007.bin",
801
+ "transformer.h.17.attn.c_proj.bias": "pytorch_model-00002-of-00007.bin",
802
+ "transformer.h.17.attn.c_proj.weight": "pytorch_model-00002-of-00007.bin",
803
+ "transformer.h.17.ln_1.bias": "pytorch_model-00002-of-00007.bin",
804
+ "transformer.h.17.ln_1.weight": "pytorch_model-00002-of-00007.bin",
805
+ "transformer.h.17.ln_2.bias": "pytorch_model-00002-of-00007.bin",
806
+ "transformer.h.17.ln_2.weight": "pytorch_model-00002-of-00007.bin",
807
+ "transformer.h.17.mlp.c_fc.bias": "pytorch_model-00002-of-00007.bin",
808
+ "transformer.h.17.mlp.c_fc.weight": "pytorch_model-00002-of-00007.bin",
809
+ "transformer.h.17.mlp.c_proj.bias": "pytorch_model-00002-of-00007.bin",
810
+ "transformer.h.17.mlp.c_proj.weight": "pytorch_model-00002-of-00007.bin",
811
+ "transformer.h.18.attn.c_attn.bias": "pytorch_model-00002-of-00007.bin",
812
+ "transformer.h.18.attn.c_attn.weight": "pytorch_model-00002-of-00007.bin",
813
+ "transformer.h.18.attn.c_proj.bias": "pytorch_model-00002-of-00007.bin",
814
+ "transformer.h.18.attn.c_proj.weight": "pytorch_model-00002-of-00007.bin",
815
+ "transformer.h.18.ln_1.bias": "pytorch_model-00002-of-00007.bin",
816
+ "transformer.h.18.ln_1.weight": "pytorch_model-00002-of-00007.bin",
817
+ "transformer.h.18.ln_2.bias": "pytorch_model-00002-of-00007.bin",
818
+ "transformer.h.18.ln_2.weight": "pytorch_model-00002-of-00007.bin",
819
+ "transformer.h.18.mlp.c_fc.bias": "pytorch_model-00002-of-00007.bin",
820
+ "transformer.h.18.mlp.c_fc.weight": "pytorch_model-00002-of-00007.bin",
821
+ "transformer.h.18.mlp.c_proj.bias": "pytorch_model-00002-of-00007.bin",
822
+ "transformer.h.18.mlp.c_proj.weight": "pytorch_model-00002-of-00007.bin",
823
+ "transformer.h.19.attn.c_attn.bias": "pytorch_model-00002-of-00007.bin",
824
+ "transformer.h.19.attn.c_attn.weight": "pytorch_model-00002-of-00007.bin",
825
+ "transformer.h.19.attn.c_proj.bias": "pytorch_model-00002-of-00007.bin",
826
+ "transformer.h.19.attn.c_proj.weight": "pytorch_model-00002-of-00007.bin",
827
+ "transformer.h.19.ln_1.bias": "pytorch_model-00002-of-00007.bin",
828
+ "transformer.h.19.ln_1.weight": "pytorch_model-00002-of-00007.bin",
829
+ "transformer.h.19.ln_2.bias": "pytorch_model-00002-of-00007.bin",
830
+ "transformer.h.19.ln_2.weight": "pytorch_model-00002-of-00007.bin",
831
+ "transformer.h.19.mlp.c_fc.bias": "pytorch_model-00002-of-00007.bin",
832
+ "transformer.h.19.mlp.c_fc.weight": "pytorch_model-00002-of-00007.bin",
833
+ "transformer.h.19.mlp.c_proj.bias": "pytorch_model-00002-of-00007.bin",
834
+ "transformer.h.19.mlp.c_proj.weight": "pytorch_model-00002-of-00007.bin",
835
+ "transformer.h.2.attn.c_attn.bias": "pytorch_model-00001-of-00007.bin",
836
+ "transformer.h.2.attn.c_attn.weight": "pytorch_model-00001-of-00007.bin",
837
+ "transformer.h.2.attn.c_proj.bias": "pytorch_model-00001-of-00007.bin",
838
+ "transformer.h.2.attn.c_proj.weight": "pytorch_model-00001-of-00007.bin",
839
+ "transformer.h.2.ln_1.bias": "pytorch_model-00001-of-00007.bin",
840
+ "transformer.h.2.ln_1.weight": "pytorch_model-00001-of-00007.bin",
841
+ "transformer.h.2.ln_2.bias": "pytorch_model-00001-of-00007.bin",
842
+ "transformer.h.2.ln_2.weight": "pytorch_model-00001-of-00007.bin",
843
+ "transformer.h.2.mlp.c_fc.bias": "pytorch_model-00001-of-00007.bin",
844
+ "transformer.h.2.mlp.c_fc.weight": "pytorch_model-00001-of-00007.bin",
845
+ "transformer.h.2.mlp.c_proj.bias": "pytorch_model-00001-of-00007.bin",
846
+ "transformer.h.2.mlp.c_proj.weight": "pytorch_model-00001-of-00007.bin",
847
+ "transformer.h.20.attn.c_attn.bias": "pytorch_model-00002-of-00007.bin",
848
+ "transformer.h.20.attn.c_attn.weight": "pytorch_model-00002-of-00007.bin",
849
+ "transformer.h.20.attn.c_proj.bias": "pytorch_model-00002-of-00007.bin",
850
+ "transformer.h.20.attn.c_proj.weight": "pytorch_model-00002-of-00007.bin",
851
+ "transformer.h.20.ln_1.bias": "pytorch_model-00002-of-00007.bin",
852
+ "transformer.h.20.ln_1.weight": "pytorch_model-00002-of-00007.bin",
853
+ "transformer.h.20.ln_2.bias": "pytorch_model-00002-of-00007.bin",
854
+ "transformer.h.20.ln_2.weight": "pytorch_model-00002-of-00007.bin",
855
+ "transformer.h.20.mlp.c_fc.bias": "pytorch_model-00002-of-00007.bin",
856
+ "transformer.h.20.mlp.c_fc.weight": "pytorch_model-00002-of-00007.bin",
857
+ "transformer.h.20.mlp.c_proj.bias": "pytorch_model-00002-of-00007.bin",
858
+ "transformer.h.20.mlp.c_proj.weight": "pytorch_model-00002-of-00007.bin",
859
+ "transformer.h.21.attn.c_attn.bias": "pytorch_model-00002-of-00007.bin",
860
+ "transformer.h.21.attn.c_attn.weight": "pytorch_model-00002-of-00007.bin",
861
+ "transformer.h.21.attn.c_proj.bias": "pytorch_model-00002-of-00007.bin",
862
+ "transformer.h.21.attn.c_proj.weight": "pytorch_model-00002-of-00007.bin",
863
+ "transformer.h.21.ln_1.bias": "pytorch_model-00002-of-00007.bin",
864
+ "transformer.h.21.ln_1.weight": "pytorch_model-00002-of-00007.bin",
865
+ "transformer.h.21.ln_2.bias": "pytorch_model-00002-of-00007.bin",
866
+ "transformer.h.21.ln_2.weight": "pytorch_model-00002-of-00007.bin",
867
+ "transformer.h.21.mlp.c_fc.bias": "pytorch_model-00002-of-00007.bin",
868
+ "transformer.h.21.mlp.c_fc.weight": "pytorch_model-00002-of-00007.bin",
869
+ "transformer.h.21.mlp.c_proj.bias": "pytorch_model-00002-of-00007.bin",
870
+ "transformer.h.21.mlp.c_proj.weight": "pytorch_model-00002-of-00007.bin",
871
+ "transformer.h.22.attn.c_attn.bias": "pytorch_model-00002-of-00007.bin",
872
+ "transformer.h.22.attn.c_attn.weight": "pytorch_model-00002-of-00007.bin",
873
+ "transformer.h.22.attn.c_proj.bias": "pytorch_model-00002-of-00007.bin",
874
+ "transformer.h.22.attn.c_proj.weight": "pytorch_model-00002-of-00007.bin",
875
+ "transformer.h.22.ln_1.bias": "pytorch_model-00002-of-00007.bin",
876
+ "transformer.h.22.ln_1.weight": "pytorch_model-00002-of-00007.bin",
877
+ "transformer.h.22.ln_2.bias": "pytorch_model-00002-of-00007.bin",
878
+ "transformer.h.22.ln_2.weight": "pytorch_model-00002-of-00007.bin",
879
+ "transformer.h.22.mlp.c_fc.bias": "pytorch_model-00002-of-00007.bin",
880
+ "transformer.h.22.mlp.c_fc.weight": "pytorch_model-00002-of-00007.bin",
881
+ "transformer.h.22.mlp.c_proj.bias": "pytorch_model-00002-of-00007.bin",
882
+ "transformer.h.22.mlp.c_proj.weight": "pytorch_model-00002-of-00007.bin",
883
+ "transformer.h.23.attn.c_attn.bias": "pytorch_model-00002-of-00007.bin",
884
+ "transformer.h.23.attn.c_attn.weight": "pytorch_model-00002-of-00007.bin",
885
+ "transformer.h.23.attn.c_proj.bias": "pytorch_model-00002-of-00007.bin",
886
+ "transformer.h.23.attn.c_proj.weight": "pytorch_model-00002-of-00007.bin",
887
+ "transformer.h.23.ln_1.bias": "pytorch_model-00002-of-00007.bin",
888
+ "transformer.h.23.ln_1.weight": "pytorch_model-00002-of-00007.bin",
889
+ "transformer.h.23.ln_2.bias": "pytorch_model-00002-of-00007.bin",
890
+ "transformer.h.23.ln_2.weight": "pytorch_model-00002-of-00007.bin",
891
+ "transformer.h.23.mlp.c_fc.bias": "pytorch_model-00002-of-00007.bin",
892
+ "transformer.h.23.mlp.c_fc.weight": "pytorch_model-00002-of-00007.bin",
893
+ "transformer.h.23.mlp.c_proj.bias": "pytorch_model-00002-of-00007.bin",
894
+ "transformer.h.23.mlp.c_proj.weight": "pytorch_model-00002-of-00007.bin",
895
+ "transformer.h.24.attn.c_attn.bias": "pytorch_model-00002-of-00007.bin",
896
+ "transformer.h.24.attn.c_attn.weight": "pytorch_model-00002-of-00007.bin",
897
+ "transformer.h.24.attn.c_proj.bias": "pytorch_model-00002-of-00007.bin",
898
+ "transformer.h.24.attn.c_proj.weight": "pytorch_model-00002-of-00007.bin",
899
+ "transformer.h.24.ln_1.bias": "pytorch_model-00002-of-00007.bin",
900
+ "transformer.h.24.ln_1.weight": "pytorch_model-00002-of-00007.bin",
901
+ "transformer.h.24.ln_2.bias": "pytorch_model-00002-of-00007.bin",
902
+ "transformer.h.24.ln_2.weight": "pytorch_model-00002-of-00007.bin",
903
+ "transformer.h.24.mlp.c_fc.bias": "pytorch_model-00002-of-00007.bin",
904
+ "transformer.h.24.mlp.c_fc.weight": "pytorch_model-00002-of-00007.bin",
905
+ "transformer.h.24.mlp.c_proj.bias": "pytorch_model-00002-of-00007.bin",
906
+ "transformer.h.24.mlp.c_proj.weight": "pytorch_model-00002-of-00007.bin",
907
+ "transformer.h.25.attn.c_attn.bias": "pytorch_model-00002-of-00007.bin",
908
+ "transformer.h.25.attn.c_attn.weight": "pytorch_model-00002-of-00007.bin",
909
+ "transformer.h.25.attn.c_proj.bias": "pytorch_model-00002-of-00007.bin",
910
+ "transformer.h.25.attn.c_proj.weight": "pytorch_model-00002-of-00007.bin",
911
+ "transformer.h.25.ln_1.bias": "pytorch_model-00002-of-00007.bin",
912
+ "transformer.h.25.ln_1.weight": "pytorch_model-00002-of-00007.bin",
913
+ "transformer.h.25.ln_2.bias": "pytorch_model-00002-of-00007.bin",
914
+ "transformer.h.25.ln_2.weight": "pytorch_model-00002-of-00007.bin",
915
+ "transformer.h.25.mlp.c_fc.bias": "pytorch_model-00003-of-00007.bin",
916
+ "transformer.h.25.mlp.c_fc.weight": "pytorch_model-00003-of-00007.bin",
917
+ "transformer.h.25.mlp.c_proj.bias": "pytorch_model-00003-of-00007.bin",
918
+ "transformer.h.25.mlp.c_proj.weight": "pytorch_model-00003-of-00007.bin",
919
+ "transformer.h.26.attn.c_attn.bias": "pytorch_model-00003-of-00007.bin",
920
+ "transformer.h.26.attn.c_attn.weight": "pytorch_model-00003-of-00007.bin",
921
+ "transformer.h.26.attn.c_proj.bias": "pytorch_model-00003-of-00007.bin",
922
+ "transformer.h.26.attn.c_proj.weight": "pytorch_model-00003-of-00007.bin",
923
+ "transformer.h.26.ln_1.bias": "pytorch_model-00003-of-00007.bin",
924
+ "transformer.h.26.ln_1.weight": "pytorch_model-00003-of-00007.bin",
925
+ "transformer.h.26.ln_2.bias": "pytorch_model-00003-of-00007.bin",
926
+ "transformer.h.26.ln_2.weight": "pytorch_model-00003-of-00007.bin",
927
+ "transformer.h.26.mlp.c_fc.bias": "pytorch_model-00003-of-00007.bin",
928
+ "transformer.h.26.mlp.c_fc.weight": "pytorch_model-00003-of-00007.bin",
929
+ "transformer.h.26.mlp.c_proj.bias": "pytorch_model-00003-of-00007.bin",
930
+ "transformer.h.26.mlp.c_proj.weight": "pytorch_model-00003-of-00007.bin",
931
+ "transformer.h.27.attn.c_attn.bias": "pytorch_model-00003-of-00007.bin",
932
+ "transformer.h.27.attn.c_attn.weight": "pytorch_model-00003-of-00007.bin",
933
+ "transformer.h.27.attn.c_proj.bias": "pytorch_model-00003-of-00007.bin",
934
+ "transformer.h.27.attn.c_proj.weight": "pytorch_model-00003-of-00007.bin",
935
+ "transformer.h.27.ln_1.bias": "pytorch_model-00003-of-00007.bin",
936
+ "transformer.h.27.ln_1.weight": "pytorch_model-00003-of-00007.bin",
937
+ "transformer.h.27.ln_2.bias": "pytorch_model-00003-of-00007.bin",
938
+ "transformer.h.27.ln_2.weight": "pytorch_model-00003-of-00007.bin",
939
+ "transformer.h.27.mlp.c_fc.bias": "pytorch_model-00003-of-00007.bin",
940
+ "transformer.h.27.mlp.c_fc.weight": "pytorch_model-00003-of-00007.bin",
941
+ "transformer.h.27.mlp.c_proj.bias": "pytorch_model-00003-of-00007.bin",
942
+ "transformer.h.27.mlp.c_proj.weight": "pytorch_model-00003-of-00007.bin",
943
+ "transformer.h.28.attn.c_attn.bias": "pytorch_model-00003-of-00007.bin",
944
+ "transformer.h.28.attn.c_attn.weight": "pytorch_model-00003-of-00007.bin",
945
+ "transformer.h.28.attn.c_proj.bias": "pytorch_model-00003-of-00007.bin",
946
+ "transformer.h.28.attn.c_proj.weight": "pytorch_model-00003-of-00007.bin",
947
+ "transformer.h.28.ln_1.bias": "pytorch_model-00003-of-00007.bin",
948
+ "transformer.h.28.ln_1.weight": "pytorch_model-00003-of-00007.bin",
949
+ "transformer.h.28.ln_2.bias": "pytorch_model-00003-of-00007.bin",
950
+ "transformer.h.28.ln_2.weight": "pytorch_model-00003-of-00007.bin",
951
+ "transformer.h.28.mlp.c_fc.bias": "pytorch_model-00003-of-00007.bin",
952
+ "transformer.h.28.mlp.c_fc.weight": "pytorch_model-00003-of-00007.bin",
953
+ "transformer.h.28.mlp.c_proj.bias": "pytorch_model-00003-of-00007.bin",
954
+ "transformer.h.28.mlp.c_proj.weight": "pytorch_model-00003-of-00007.bin",
955
+ "transformer.h.29.attn.c_attn.bias": "pytorch_model-00003-of-00007.bin",
956
+ "transformer.h.29.attn.c_attn.weight": "pytorch_model-00003-of-00007.bin",
957
+ "transformer.h.29.attn.c_proj.bias": "pytorch_model-00003-of-00007.bin",
958
+ "transformer.h.29.attn.c_proj.weight": "pytorch_model-00003-of-00007.bin",
959
+ "transformer.h.29.ln_1.bias": "pytorch_model-00003-of-00007.bin",
960
+ "transformer.h.29.ln_1.weight": "pytorch_model-00003-of-00007.bin",
961
+ "transformer.h.29.ln_2.bias": "pytorch_model-00003-of-00007.bin",
962
+ "transformer.h.29.ln_2.weight": "pytorch_model-00003-of-00007.bin",
963
+ "transformer.h.29.mlp.c_fc.bias": "pytorch_model-00003-of-00007.bin",
964
+ "transformer.h.29.mlp.c_fc.weight": "pytorch_model-00003-of-00007.bin",
965
+ "transformer.h.29.mlp.c_proj.bias": "pytorch_model-00003-of-00007.bin",
966
+ "transformer.h.29.mlp.c_proj.weight": "pytorch_model-00003-of-00007.bin",
967
+ "transformer.h.3.attn.c_attn.bias": "pytorch_model-00001-of-00007.bin",
968
+ "transformer.h.3.attn.c_attn.weight": "pytorch_model-00001-of-00007.bin",
969
+ "transformer.h.3.attn.c_proj.bias": "pytorch_model-00001-of-00007.bin",
970
+ "transformer.h.3.attn.c_proj.weight": "pytorch_model-00001-of-00007.bin",
971
+ "transformer.h.3.ln_1.bias": "pytorch_model-00001-of-00007.bin",
972
+ "transformer.h.3.ln_1.weight": "pytorch_model-00001-of-00007.bin",
973
+ "transformer.h.3.ln_2.bias": "pytorch_model-00001-of-00007.bin",
974
+ "transformer.h.3.ln_2.weight": "pytorch_model-00001-of-00007.bin",
975
+ "transformer.h.3.mlp.c_fc.bias": "pytorch_model-00001-of-00007.bin",
976
+ "transformer.h.3.mlp.c_fc.weight": "pytorch_model-00001-of-00007.bin",
977
+ "transformer.h.3.mlp.c_proj.bias": "pytorch_model-00001-of-00007.bin",
978
+ "transformer.h.3.mlp.c_proj.weight": "pytorch_model-00001-of-00007.bin",
979
+ "transformer.h.30.attn.c_attn.bias": "pytorch_model-00003-of-00007.bin",
980
+ "transformer.h.30.attn.c_attn.weight": "pytorch_model-00003-of-00007.bin",
981
+ "transformer.h.30.attn.c_proj.bias": "pytorch_model-00003-of-00007.bin",
982
+ "transformer.h.30.attn.c_proj.weight": "pytorch_model-00003-of-00007.bin",
983
+ "transformer.h.30.ln_1.bias": "pytorch_model-00003-of-00007.bin",
984
+ "transformer.h.30.ln_1.weight": "pytorch_model-00003-of-00007.bin",
985
+ "transformer.h.30.ln_2.bias": "pytorch_model-00003-of-00007.bin",
986
+ "transformer.h.30.ln_2.weight": "pytorch_model-00003-of-00007.bin",
987
+ "transformer.h.30.mlp.c_fc.bias": "pytorch_model-00003-of-00007.bin",
988
+ "transformer.h.30.mlp.c_fc.weight": "pytorch_model-00003-of-00007.bin",
989
+ "transformer.h.30.mlp.c_proj.bias": "pytorch_model-00003-of-00007.bin",
990
+ "transformer.h.30.mlp.c_proj.weight": "pytorch_model-00003-of-00007.bin",
991
+ "transformer.h.31.attn.c_attn.bias": "pytorch_model-00003-of-00007.bin",
992
+ "transformer.h.31.attn.c_attn.weight": "pytorch_model-00003-of-00007.bin",
993
+ "transformer.h.31.attn.c_proj.bias": "pytorch_model-00003-of-00007.bin",
994
+ "transformer.h.31.attn.c_proj.weight": "pytorch_model-00003-of-00007.bin",
995
+ "transformer.h.31.ln_1.bias": "pytorch_model-00003-of-00007.bin",
996
+ "transformer.h.31.ln_1.weight": "pytorch_model-00003-of-00007.bin",
997
+ "transformer.h.31.ln_2.bias": "pytorch_model-00003-of-00007.bin",
998
+ "transformer.h.31.ln_2.weight": "pytorch_model-00003-of-00007.bin",
999
+ "transformer.h.31.mlp.c_fc.bias": "pytorch_model-00003-of-00007.bin",
1000
+ "transformer.h.31.mlp.c_fc.weight": "pytorch_model-00003-of-00007.bin",
1001
+ "transformer.h.31.mlp.c_proj.bias": "pytorch_model-00003-of-00007.bin",
1002
+ "transformer.h.31.mlp.c_proj.weight": "pytorch_model-00003-of-00007.bin",
1003
+ "transformer.h.32.attn.c_attn.bias": "pytorch_model-00003-of-00007.bin",
1004
+ "transformer.h.32.attn.c_attn.weight": "pytorch_model-00003-of-00007.bin",
1005
+ "transformer.h.32.attn.c_proj.bias": "pytorch_model-00003-of-00007.bin",
1006
+ "transformer.h.32.attn.c_proj.weight": "pytorch_model-00003-of-00007.bin",
1007
+ "transformer.h.32.ln_1.bias": "pytorch_model-00003-of-00007.bin",
1008
+ "transformer.h.32.ln_1.weight": "pytorch_model-00003-of-00007.bin",
1009
+ "transformer.h.32.ln_2.bias": "pytorch_model-00003-of-00007.bin",
1010
+ "transformer.h.32.ln_2.weight": "pytorch_model-00003-of-00007.bin",
1011
+ "transformer.h.32.mlp.c_fc.bias": "pytorch_model-00003-of-00007.bin",
1012
+ "transformer.h.32.mlp.c_fc.weight": "pytorch_model-00003-of-00007.bin",
1013
+ "transformer.h.32.mlp.c_proj.bias": "pytorch_model-00003-of-00007.bin",
1014
+ "transformer.h.32.mlp.c_proj.weight": "pytorch_model-00003-of-00007.bin",
1015
+ "transformer.h.33.attn.c_attn.bias": "pytorch_model-00003-of-00007.bin",
1016
+ "transformer.h.33.attn.c_attn.weight": "pytorch_model-00003-of-00007.bin",
1017
+ "transformer.h.33.attn.c_proj.bias": "pytorch_model-00003-of-00007.bin",
1018
+ "transformer.h.33.attn.c_proj.weight": "pytorch_model-00003-of-00007.bin",
1019
+ "transformer.h.33.ln_1.bias": "pytorch_model-00003-of-00007.bin",
1020
+ "transformer.h.33.ln_1.weight": "pytorch_model-00003-of-00007.bin",
1021
+ "transformer.h.33.ln_2.bias": "pytorch_model-00003-of-00007.bin",
1022
+ "transformer.h.33.ln_2.weight": "pytorch_model-00003-of-00007.bin",
1023
+ "transformer.h.33.mlp.c_fc.bias": "pytorch_model-00003-of-00007.bin",
1024
+ "transformer.h.33.mlp.c_fc.weight": "pytorch_model-00003-of-00007.bin",
1025
+ "transformer.h.33.mlp.c_proj.bias": "pytorch_model-00003-of-00007.bin",
1026
+ "transformer.h.33.mlp.c_proj.weight": "pytorch_model-00003-of-00007.bin",
1027
+ "transformer.h.34.attn.c_attn.bias": "pytorch_model-00003-of-00007.bin",
1028
+ "transformer.h.34.attn.c_attn.weight": "pytorch_model-00003-of-00007.bin",
1029
+ "transformer.h.34.attn.c_proj.bias": "pytorch_model-00003-of-00007.bin",
1030
+ "transformer.h.34.attn.c_proj.weight": "pytorch_model-00003-of-00007.bin",
1031
+ "transformer.h.34.ln_1.bias": "pytorch_model-00003-of-00007.bin",
1032
+ "transformer.h.34.ln_1.weight": "pytorch_model-00003-of-00007.bin",
1033
+ "transformer.h.34.ln_2.bias": "pytorch_model-00003-of-00007.bin",
1034
+ "transformer.h.34.ln_2.weight": "pytorch_model-00003-of-00007.bin",
1035
+ "transformer.h.34.mlp.c_fc.bias": "pytorch_model-00003-of-00007.bin",
1036
+ "transformer.h.34.mlp.c_fc.weight": "pytorch_model-00003-of-00007.bin",
1037
+ "transformer.h.34.mlp.c_proj.bias": "pytorch_model-00003-of-00007.bin",
1038
+ "transformer.h.34.mlp.c_proj.weight": "pytorch_model-00003-of-00007.bin",
1039
+ "transformer.h.35.attn.c_attn.bias": "pytorch_model-00003-of-00007.bin",
1040
+ "transformer.h.35.attn.c_attn.weight": "pytorch_model-00003-of-00007.bin",
1041
+ "transformer.h.35.attn.c_proj.bias": "pytorch_model-00003-of-00007.bin",
1042
+ "transformer.h.35.attn.c_proj.weight": "pytorch_model-00003-of-00007.bin",
1043
+ "transformer.h.35.ln_1.bias": "pytorch_model-00003-of-00007.bin",
1044
+ "transformer.h.35.ln_1.weight": "pytorch_model-00003-of-00007.bin",
1045
+ "transformer.h.35.ln_2.bias": "pytorch_model-00003-of-00007.bin",
1046
+ "transformer.h.35.ln_2.weight": "pytorch_model-00003-of-00007.bin",
1047
+ "transformer.h.35.mlp.c_fc.bias": "pytorch_model-00003-of-00007.bin",
1048
+ "transformer.h.35.mlp.c_fc.weight": "pytorch_model-00003-of-00007.bin",
1049
+ "transformer.h.35.mlp.c_proj.bias": "pytorch_model-00003-of-00007.bin",
1050
+ "transformer.h.35.mlp.c_proj.weight": "pytorch_model-00003-of-00007.bin",
1051
+ "transformer.h.36.attn.c_attn.bias": "pytorch_model-00003-of-00007.bin",
1052
+ "transformer.h.36.attn.c_attn.weight": "pytorch_model-00003-of-00007.bin",
1053
+ "transformer.h.36.attn.c_proj.bias": "pytorch_model-00003-of-00007.bin",
1054
+ "transformer.h.36.attn.c_proj.weight": "pytorch_model-00003-of-00007.bin",
1055
+ "transformer.h.36.ln_1.bias": "pytorch_model-00003-of-00007.bin",
1056
+ "transformer.h.36.ln_1.weight": "pytorch_model-00003-of-00007.bin",
1057
+ "transformer.h.36.ln_2.bias": "pytorch_model-00003-of-00007.bin",
1058
+ "transformer.h.36.ln_2.weight": "pytorch_model-00003-of-00007.bin",
1059
+ "transformer.h.36.mlp.c_fc.bias": "pytorch_model-00003-of-00007.bin",
1060
+ "transformer.h.36.mlp.c_fc.weight": "pytorch_model-00003-of-00007.bin",
1061
+ "transformer.h.36.mlp.c_proj.bias": "pytorch_model-00003-of-00007.bin",
1062
+ "transformer.h.36.mlp.c_proj.weight": "pytorch_model-00003-of-00007.bin",
1063
+ "transformer.h.37.attn.c_attn.bias": "pytorch_model-00003-of-00007.bin",
1064
+ "transformer.h.37.attn.c_attn.weight": "pytorch_model-00003-of-00007.bin",
1065
+ "transformer.h.37.attn.c_proj.bias": "pytorch_model-00003-of-00007.bin",
1066
+ "transformer.h.37.attn.c_proj.weight": "pytorch_model-00003-of-00007.bin",
1067
+ "transformer.h.37.ln_1.bias": "pytorch_model-00003-of-00007.bin",
1068
+ "transformer.h.37.ln_1.weight": "pytorch_model-00003-of-00007.bin",
1069
+ "transformer.h.37.ln_2.bias": "pytorch_model-00003-of-00007.bin",
1070
+ "transformer.h.37.ln_2.weight": "pytorch_model-00003-of-00007.bin",
1071
+ "transformer.h.37.mlp.c_fc.bias": "pytorch_model-00003-of-00007.bin",
1072
+ "transformer.h.37.mlp.c_fc.weight": "pytorch_model-00003-of-00007.bin",
1073
+ "transformer.h.37.mlp.c_proj.bias": "pytorch_model-00003-of-00007.bin",
1074
+ "transformer.h.37.mlp.c_proj.weight": "pytorch_model-00003-of-00007.bin",
1075
+ "transformer.h.38.attn.c_attn.bias": "pytorch_model-00003-of-00007.bin",
1076
+ "transformer.h.38.attn.c_attn.weight": "pytorch_model-00003-of-00007.bin",
1077
+ "transformer.h.38.attn.c_proj.bias": "pytorch_model-00003-of-00007.bin",
1078
+ "transformer.h.38.attn.c_proj.weight": "pytorch_model-00003-of-00007.bin",
1079
+ "transformer.h.38.ln_1.bias": "pytorch_model-00003-of-00007.bin",
1080
+ "transformer.h.38.ln_1.weight": "pytorch_model-00003-of-00007.bin",
1081
+ "transformer.h.38.ln_2.bias": "pytorch_model-00003-of-00007.bin",
1082
+ "transformer.h.38.ln_2.weight": "pytorch_model-00003-of-00007.bin",
1083
+ "transformer.h.38.mlp.c_fc.bias": "pytorch_model-00004-of-00007.bin",
1084
+ "transformer.h.38.mlp.c_fc.weight": "pytorch_model-00004-of-00007.bin",
1085
+ "transformer.h.38.mlp.c_proj.bias": "pytorch_model-00004-of-00007.bin",
1086
+ "transformer.h.38.mlp.c_proj.weight": "pytorch_model-00004-of-00007.bin",
1087
+ "transformer.h.39.attn.c_attn.bias": "pytorch_model-00004-of-00007.bin",
1088
+ "transformer.h.39.attn.c_attn.weight": "pytorch_model-00004-of-00007.bin",
1089
+ "transformer.h.39.attn.c_proj.bias": "pytorch_model-00004-of-00007.bin",
1090
+ "transformer.h.39.attn.c_proj.weight": "pytorch_model-00004-of-00007.bin",
1091
+ "transformer.h.39.ln_1.bias": "pytorch_model-00004-of-00007.bin",
1092
+ "transformer.h.39.ln_1.weight": "pytorch_model-00004-of-00007.bin",
1093
+ "transformer.h.39.ln_2.bias": "pytorch_model-00004-of-00007.bin",
1094
+ "transformer.h.39.ln_2.weight": "pytorch_model-00004-of-00007.bin",
1095
+ "transformer.h.39.mlp.c_fc.bias": "pytorch_model-00004-of-00007.bin",
1096
+ "transformer.h.39.mlp.c_fc.weight": "pytorch_model-00004-of-00007.bin",
1097
+ "transformer.h.39.mlp.c_proj.bias": "pytorch_model-00004-of-00007.bin",
1098
+ "transformer.h.39.mlp.c_proj.weight": "pytorch_model-00004-of-00007.bin",
1099
+ "transformer.h.4.attn.c_attn.bias": "pytorch_model-00001-of-00007.bin",
1100
+ "transformer.h.4.attn.c_attn.weight": "pytorch_model-00001-of-00007.bin",
1101
+ "transformer.h.4.attn.c_proj.bias": "pytorch_model-00001-of-00007.bin",
1102
+ "transformer.h.4.attn.c_proj.weight": "pytorch_model-00001-of-00007.bin",
1103
+ "transformer.h.4.ln_1.bias": "pytorch_model-00001-of-00007.bin",
1104
+ "transformer.h.4.ln_1.weight": "pytorch_model-00001-of-00007.bin",
1105
+ "transformer.h.4.ln_2.bias": "pytorch_model-00001-of-00007.bin",
1106
+ "transformer.h.4.ln_2.weight": "pytorch_model-00001-of-00007.bin",
1107
+ "transformer.h.4.mlp.c_fc.bias": "pytorch_model-00001-of-00007.bin",
1108
+ "transformer.h.4.mlp.c_fc.weight": "pytorch_model-00001-of-00007.bin",
1109
+ "transformer.h.4.mlp.c_proj.bias": "pytorch_model-00001-of-00007.bin",
1110
+ "transformer.h.4.mlp.c_proj.weight": "pytorch_model-00001-of-00007.bin",
1111
+ "transformer.h.5.attn.c_attn.bias": "pytorch_model-00001-of-00007.bin",
1112
+ "transformer.h.5.attn.c_attn.weight": "pytorch_model-00001-of-00007.bin",
1113
+ "transformer.h.5.attn.c_proj.bias": "pytorch_model-00001-of-00007.bin",
1114
+ "transformer.h.5.attn.c_proj.weight": "pytorch_model-00001-of-00007.bin",
1115
+ "transformer.h.5.ln_1.bias": "pytorch_model-00001-of-00007.bin",
1116
+ "transformer.h.5.ln_1.weight": "pytorch_model-00001-of-00007.bin",
1117
+ "transformer.h.5.ln_2.bias": "pytorch_model-00001-of-00007.bin",
1118
+ "transformer.h.5.ln_2.weight": "pytorch_model-00001-of-00007.bin",
1119
+ "transformer.h.5.mlp.c_fc.bias": "pytorch_model-00001-of-00007.bin",
1120
+ "transformer.h.5.mlp.c_fc.weight": "pytorch_model-00001-of-00007.bin",
1121
+ "transformer.h.5.mlp.c_proj.bias": "pytorch_model-00001-of-00007.bin",
1122
+ "transformer.h.5.mlp.c_proj.weight": "pytorch_model-00001-of-00007.bin",
1123
+ "transformer.h.6.attn.c_attn.bias": "pytorch_model-00001-of-00007.bin",
1124
+ "transformer.h.6.attn.c_attn.weight": "pytorch_model-00001-of-00007.bin",
1125
+ "transformer.h.6.attn.c_proj.bias": "pytorch_model-00001-of-00007.bin",
1126
+ "transformer.h.6.attn.c_proj.weight": "pytorch_model-00001-of-00007.bin",
1127
+ "transformer.h.6.ln_1.bias": "pytorch_model-00001-of-00007.bin",
1128
+ "transformer.h.6.ln_1.weight": "pytorch_model-00001-of-00007.bin",
1129
+ "transformer.h.6.ln_2.bias": "pytorch_model-00001-of-00007.bin",
1130
+ "transformer.h.6.ln_2.weight": "pytorch_model-00001-of-00007.bin",
1131
+ "transformer.h.6.mlp.c_fc.bias": "pytorch_model-00001-of-00007.bin",
1132
+ "transformer.h.6.mlp.c_fc.weight": "pytorch_model-00001-of-00007.bin",
1133
+ "transformer.h.6.mlp.c_proj.bias": "pytorch_model-00001-of-00007.bin",
1134
+ "transformer.h.6.mlp.c_proj.weight": "pytorch_model-00001-of-00007.bin",
1135
+ "transformer.h.7.attn.c_attn.bias": "pytorch_model-00001-of-00007.bin",
1136
+ "transformer.h.7.attn.c_attn.weight": "pytorch_model-00001-of-00007.bin",
1137
+ "transformer.h.7.attn.c_proj.bias": "pytorch_model-00001-of-00007.bin",
1138
+ "transformer.h.7.attn.c_proj.weight": "pytorch_model-00001-of-00007.bin",
1139
+ "transformer.h.7.ln_1.bias": "pytorch_model-00001-of-00007.bin",
1140
+ "transformer.h.7.ln_1.weight": "pytorch_model-00001-of-00007.bin",
1141
+ "transformer.h.7.ln_2.bias": "pytorch_model-00001-of-00007.bin",
1142
+ "transformer.h.7.ln_2.weight": "pytorch_model-00001-of-00007.bin",
1143
+ "transformer.h.7.mlp.c_fc.bias": "pytorch_model-00001-of-00007.bin",
1144
+ "transformer.h.7.mlp.c_fc.weight": "pytorch_model-00001-of-00007.bin",
1145
+ "transformer.h.7.mlp.c_proj.bias": "pytorch_model-00001-of-00007.bin",
1146
+ "transformer.h.7.mlp.c_proj.weight": "pytorch_model-00001-of-00007.bin",
1147
+ "transformer.h.8.attn.c_attn.bias": "pytorch_model-00001-of-00007.bin",
1148
+ "transformer.h.8.attn.c_attn.weight": "pytorch_model-00001-of-00007.bin",
1149
+ "transformer.h.8.attn.c_proj.bias": "pytorch_model-00001-of-00007.bin",
1150
+ "transformer.h.8.attn.c_proj.weight": "pytorch_model-00001-of-00007.bin",
1151
+ "transformer.h.8.ln_1.bias": "pytorch_model-00001-of-00007.bin",
1152
+ "transformer.h.8.ln_1.weight": "pytorch_model-00001-of-00007.bin",
1153
+ "transformer.h.8.ln_2.bias": "pytorch_model-00001-of-00007.bin",
1154
+ "transformer.h.8.ln_2.weight": "pytorch_model-00001-of-00007.bin",
1155
+ "transformer.h.8.mlp.c_fc.bias": "pytorch_model-00001-of-00007.bin",
1156
+ "transformer.h.8.mlp.c_fc.weight": "pytorch_model-00001-of-00007.bin",
1157
+ "transformer.h.8.mlp.c_proj.bias": "pytorch_model-00001-of-00007.bin",
1158
+ "transformer.h.8.mlp.c_proj.weight": "pytorch_model-00001-of-00007.bin",
1159
+ "transformer.h.9.attn.c_attn.bias": "pytorch_model-00001-of-00007.bin",
1160
+ "transformer.h.9.attn.c_attn.weight": "pytorch_model-00001-of-00007.bin",
1161
+ "transformer.h.9.attn.c_proj.bias": "pytorch_model-00001-of-00007.bin",
1162
+ "transformer.h.9.attn.c_proj.weight": "pytorch_model-00001-of-00007.bin",
1163
+ "transformer.h.9.ln_1.bias": "pytorch_model-00001-of-00007.bin",
1164
+ "transformer.h.9.ln_1.weight": "pytorch_model-00001-of-00007.bin",
1165
+ "transformer.h.9.ln_2.bias": "pytorch_model-00001-of-00007.bin",
1166
+ "transformer.h.9.ln_2.weight": "pytorch_model-00001-of-00007.bin",
1167
+ "transformer.h.9.mlp.c_fc.bias": "pytorch_model-00001-of-00007.bin",
1168
+ "transformer.h.9.mlp.c_fc.weight": "pytorch_model-00001-of-00007.bin",
1169
+ "transformer.h.9.mlp.c_proj.bias": "pytorch_model-00001-of-00007.bin",
1170
+ "transformer.h.9.mlp.c_proj.weight": "pytorch_model-00001-of-00007.bin",
1171
+ "transformer.ln_f.bias": "pytorch_model-00004-of-00007.bin",
1172
+ "transformer.ln_f.weight": "pytorch_model-00004-of-00007.bin",
1173
+ "transformer.wpe.weight": "pytorch_model-00001-of-00007.bin",
1174
+ "transformer.wte.weight": "pytorch_model-00001-of-00007.bin"
1175
+ }
1176
+ }