Hugging Face's logo

mezes
/
xglm-1.7B-sharded

Text Generation
Transformers
PyTorch
xglm
Inference Endpoints
Model card Files Community
mezes commited on
Commit
5f14783
•
1 Parent(s): da1e00f

add model

Browse files
Files changed (8) hide show
  1. config.json +27 -0
  2. pytorch_model-00001-of-00006.bin +3 -0
  3. pytorch_model-00002-of-00006.bin +3 -0
  4. pytorch_model-00003-of-00006.bin +3 -0
  5. pytorch_model-00004-of-00006.bin +3 -0
  6. pytorch_model-00005-of-00006.bin +3 -0
  7. pytorch_model-00006-of-00006.bin +3 -0
  8. pytorch_model.bin.index.json +395 -0
config.json ADDED
@@ -0,0 +1,27 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "facebook/xglm-1.7B",
3
+ "activation_dropout": 0.0,
4
+ "activation_function": "gelu",
5
+ "architectures": [
6
+ "XGLMForCausalLM"
7
+ ],
8
+ "attention_dropout": 0.1,
9
+ "attention_heads": 16,
10
+ "bos_token_id": 0,
11
+ "d_model": 2048,
12
+ "decoder_start_token_id": 2,
13
+ "dropout": 0.1,
14
+ "eos_token_id": 2,
15
+ "ffn_dim": 8192,
16
+ "init_std": 0.02,
17
+ "layerdrop": 0.0,
18
+ "max_position_embeddings": 2048,
19
+ "model_type": "xglm",
20
+ "num_layers": 24,
21
+ "pad_token_id": 1,
22
+ "scale_embedding": true,
23
+ "torch_dtype": "float32",
24
+ "transformers_version": "4.21.1",
25
+ "use_cache": true,
26
+ "vocab_size": 256008
27
+ }
pytorch_model-00001-of-00006.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d030ad8db708280fcae77d87e973102039acd23a11bdecc3db8eb6c0ac940ee1
3
+ size 431
pytorch_model-00002-of-00006.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d4ed66c0bb590793db9b6aa499f9171da9635cff3d64704b7b2ea7ac7a538c1e
3
+ size 2097218347
pytorch_model-00003-of-00006.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:edde34f17af385479a5ff889ff6c3dcdf3965e05ba6ae22835b6c97ddfa92b3e
3
+ size 1947245001
pytorch_model-00004-of-00006.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:60102dc6c340a3fd0e139dfba5723c8f7b6f3f3075051c3b1477f52fe6a14161
3
+ size 1947237573
pytorch_model-00005-of-00006.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9d272cd3aafa944735122af03c368d4e50c31279b191386cce4270d9333658b2
3
+ size 940045261
pytorch_model-00006-of-00006.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:495638109628e742e0875d8870b68e7c576f0bbe0bd983a929671dd498c18221
3
+ size 2097218283
pytorch_model.bin.index.json ADDED
@@ -0,0 +1,395 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metadata": {
3
+ "total_size": 9028845568
4
+ },
5
+ "weight_map": {
6
+ "lm_head.weight": "pytorch_model-00006-of-00006.bin",
7
+ "model.embed_tokens.weight": "pytorch_model-00002-of-00006.bin",
8
+ "model.layer_norm.bias": "pytorch_model-00005-of-00006.bin",
9
+ "model.layer_norm.weight": "pytorch_model-00005-of-00006.bin",
10
+ "model.layers.0.fc1.bias": "pytorch_model-00003-of-00006.bin",
11
+ "model.layers.0.fc1.weight": "pytorch_model-00003-of-00006.bin",
12
+ "model.layers.0.fc2.bias": "pytorch_model-00003-of-00006.bin",
13
+ "model.layers.0.fc2.weight": "pytorch_model-00003-of-00006.bin",
14
+ "model.layers.0.final_layer_norm.bias": "pytorch_model-00003-of-00006.bin",
15
+ "model.layers.0.final_layer_norm.weight": "pytorch_model-00003-of-00006.bin",
16
+ "model.layers.0.self_attn.k_proj.bias": "pytorch_model-00003-of-00006.bin",
17
+ "model.layers.0.self_attn.k_proj.weight": "pytorch_model-00003-of-00006.bin",
18
+ "model.layers.0.self_attn.out_proj.bias": "pytorch_model-00003-of-00006.bin",
19
+ "model.layers.0.self_attn.out_proj.weight": "pytorch_model-00003-of-00006.bin",
20
+ "model.layers.0.self_attn.q_proj.bias": "pytorch_model-00003-of-00006.bin",
21
+ "model.layers.0.self_attn.q_proj.weight": "pytorch_model-00003-of-00006.bin",
22
+ "model.layers.0.self_attn.v_proj.bias": "pytorch_model-00003-of-00006.bin",
23
+ "model.layers.0.self_attn.v_proj.weight": "pytorch_model-00003-of-00006.bin",
24
+ "model.layers.0.self_attn_layer_norm.bias": "pytorch_model-00003-of-00006.bin",
25
+ "model.layers.0.self_attn_layer_norm.weight": "pytorch_model-00003-of-00006.bin",
26
+ "model.layers.1.fc1.bias": "pytorch_model-00003-of-00006.bin",
27
+ "model.layers.1.fc1.weight": "pytorch_model-00003-of-00006.bin",
28
+ "model.layers.1.fc2.bias": "pytorch_model-00003-of-00006.bin",
29
+ "model.layers.1.fc2.weight": "pytorch_model-00003-of-00006.bin",
30
+ "model.layers.1.final_layer_norm.bias": "pytorch_model-00003-of-00006.bin",
31
+ "model.layers.1.final_layer_norm.weight": "pytorch_model-00003-of-00006.bin",
32
+ "model.layers.1.self_attn.k_proj.bias": "pytorch_model-00003-of-00006.bin",
33
+ "model.layers.1.self_attn.k_proj.weight": "pytorch_model-00003-of-00006.bin",
34
+ "model.layers.1.self_attn.out_proj.bias": "pytorch_model-00003-of-00006.bin",
35
+ "model.layers.1.self_attn.out_proj.weight": "pytorch_model-00003-of-00006.bin",
36
+ "model.layers.1.self_attn.q_proj.bias": "pytorch_model-00003-of-00006.bin",
37
+ "model.layers.1.self_attn.q_proj.weight": "pytorch_model-00003-of-00006.bin",
38
+ "model.layers.1.self_attn.v_proj.bias": "pytorch_model-00003-of-00006.bin",
39
+ "model.layers.1.self_attn.v_proj.weight": "pytorch_model-00003-of-00006.bin",
40
+ "model.layers.1.self_attn_layer_norm.bias": "pytorch_model-00003-of-00006.bin",
41
+ "model.layers.1.self_attn_layer_norm.weight": "pytorch_model-00003-of-00006.bin",
42
+ "model.layers.10.fc1.bias": "pytorch_model-00004-of-00006.bin",
43
+ "model.layers.10.fc1.weight": "pytorch_model-00004-of-00006.bin",
44
+ "model.layers.10.fc2.bias": "pytorch_model-00004-of-00006.bin",
45
+ "model.layers.10.fc2.weight": "pytorch_model-00004-of-00006.bin",
46
+ "model.layers.10.final_layer_norm.bias": "pytorch_model-00004-of-00006.bin",
47
+ "model.layers.10.final_layer_norm.weight": "pytorch_model-00004-of-00006.bin",
48
+ "model.layers.10.self_attn.k_proj.bias": "pytorch_model-00004-of-00006.bin",
49
+ "model.layers.10.self_attn.k_proj.weight": "pytorch_model-00004-of-00006.bin",
50
+ "model.layers.10.self_attn.out_proj.bias": "pytorch_model-00004-of-00006.bin",
51
+ "model.layers.10.self_attn.out_proj.weight": "pytorch_model-00004-of-00006.bin",
52
+ "model.layers.10.self_attn.q_proj.bias": "pytorch_model-00004-of-00006.bin",
53
+ "model.layers.10.self_attn.q_proj.weight": "pytorch_model-00004-of-00006.bin",
54
+ "model.layers.10.self_attn.v_proj.bias": "pytorch_model-00004-of-00006.bin",
55
+ "model.layers.10.self_attn.v_proj.weight": "pytorch_model-00004-of-00006.bin",
56
+ "model.layers.10.self_attn_layer_norm.bias": "pytorch_model-00004-of-00006.bin",
57
+ "model.layers.10.self_attn_layer_norm.weight": "pytorch_model-00004-of-00006.bin",
58
+ "model.layers.11.fc1.bias": "pytorch_model-00004-of-00006.bin",
59
+ "model.layers.11.fc1.weight": "pytorch_model-00004-of-00006.bin",
60
+ "model.layers.11.fc2.bias": "pytorch_model-00004-of-00006.bin",
61
+ "model.layers.11.fc2.weight": "pytorch_model-00004-of-00006.bin",
62
+ "model.layers.11.final_layer_norm.bias": "pytorch_model-00004-of-00006.bin",
63
+ "model.layers.11.final_layer_norm.weight": "pytorch_model-00004-of-00006.bin",
64
+ "model.layers.11.self_attn.k_proj.bias": "pytorch_model-00004-of-00006.bin",
65
+ "model.layers.11.self_attn.k_proj.weight": "pytorch_model-00004-of-00006.bin",
66
+ "model.layers.11.self_attn.out_proj.bias": "pytorch_model-00004-of-00006.bin",
67
+ "model.layers.11.self_attn.out_proj.weight": "pytorch_model-00004-of-00006.bin",
68
+ "model.layers.11.self_attn.q_proj.bias": "pytorch_model-00004-of-00006.bin",
69
+ "model.layers.11.self_attn.q_proj.weight": "pytorch_model-00004-of-00006.bin",
70
+ "model.layers.11.self_attn.v_proj.bias": "pytorch_model-00004-of-00006.bin",
71
+ "model.layers.11.self_attn.v_proj.weight": "pytorch_model-00004-of-00006.bin",
72
+ "model.layers.11.self_attn_layer_norm.bias": "pytorch_model-00004-of-00006.bin",
73
+ "model.layers.11.self_attn_layer_norm.weight": "pytorch_model-00004-of-00006.bin",
74
+ "model.layers.12.fc1.bias": "pytorch_model-00004-of-00006.bin",
75
+ "model.layers.12.fc1.weight": "pytorch_model-00004-of-00006.bin",
76
+ "model.layers.12.fc2.bias": "pytorch_model-00004-of-00006.bin",
77
+ "model.layers.12.fc2.weight": "pytorch_model-00004-of-00006.bin",
78
+ "model.layers.12.final_layer_norm.bias": "pytorch_model-00004-of-00006.bin",
79
+ "model.layers.12.final_layer_norm.weight": "pytorch_model-00004-of-00006.bin",
80
+ "model.layers.12.self_attn.k_proj.bias": "pytorch_model-00004-of-00006.bin",
81
+ "model.layers.12.self_attn.k_proj.weight": "pytorch_model-00004-of-00006.bin",
82
+ "model.layers.12.self_attn.out_proj.bias": "pytorch_model-00004-of-00006.bin",
83
+ "model.layers.12.self_attn.out_proj.weight": "pytorch_model-00004-of-00006.bin",
84
+ "model.layers.12.self_attn.q_proj.bias": "pytorch_model-00004-of-00006.bin",
85
+ "model.layers.12.self_attn.q_proj.weight": "pytorch_model-00004-of-00006.bin",
86
+ "model.layers.12.self_attn.v_proj.bias": "pytorch_model-00004-of-00006.bin",
87
+ "model.layers.12.self_attn.v_proj.weight": "pytorch_model-00004-of-00006.bin",
88
+ "model.layers.12.self_attn_layer_norm.bias": "pytorch_model-00004-of-00006.bin",
89
+ "model.layers.12.self_attn_layer_norm.weight": "pytorch_model-00004-of-00006.bin",
90
+ "model.layers.13.fc1.bias": "pytorch_model-00004-of-00006.bin",
91
+ "model.layers.13.fc1.weight": "pytorch_model-00004-of-00006.bin",
92
+ "model.layers.13.fc2.bias": "pytorch_model-00004-of-00006.bin",
93
+ "model.layers.13.fc2.weight": "pytorch_model-00004-of-00006.bin",
94
+ "model.layers.13.final_layer_norm.bias": "pytorch_model-00004-of-00006.bin",
95
+ "model.layers.13.final_layer_norm.weight": "pytorch_model-00004-of-00006.bin",
96
+ "model.layers.13.self_attn.k_proj.bias": "pytorch_model-00004-of-00006.bin",
97
+ "model.layers.13.self_attn.k_proj.weight": "pytorch_model-00004-of-00006.bin",
98
+ "model.layers.13.self_attn.out_proj.bias": "pytorch_model-00004-of-00006.bin",
99
+ "model.layers.13.self_attn.out_proj.weight": "pytorch_model-00004-of-00006.bin",
100
+ "model.layers.13.self_attn.q_proj.bias": "pytorch_model-00004-of-00006.bin",
101
+ "model.layers.13.self_attn.q_proj.weight": "pytorch_model-00004-of-00006.bin",
102
+ "model.layers.13.self_attn.v_proj.bias": "pytorch_model-00004-of-00006.bin",
103
+ "model.layers.13.self_attn.v_proj.weight": "pytorch_model-00004-of-00006.bin",
104
+ "model.layers.13.self_attn_layer_norm.bias": "pytorch_model-00004-of-00006.bin",
105
+ "model.layers.13.self_attn_layer_norm.weight": "pytorch_model-00004-of-00006.bin",
106
+ "model.layers.14.fc1.bias": "pytorch_model-00004-of-00006.bin",
107
+ "model.layers.14.fc1.weight": "pytorch_model-00004-of-00006.bin",
108
+ "model.layers.14.fc2.bias": "pytorch_model-00004-of-00006.bin",
109
+ "model.layers.14.fc2.weight": "pytorch_model-00004-of-00006.bin",
110
+ "model.layers.14.final_layer_norm.bias": "pytorch_model-00004-of-00006.bin",
111
+ "model.layers.14.final_layer_norm.weight": "pytorch_model-00004-of-00006.bin",
112
+ "model.layers.14.self_attn.k_proj.bias": "pytorch_model-00004-of-00006.bin",
113
+ "model.layers.14.self_attn.k_proj.weight": "pytorch_model-00004-of-00006.bin",
114
+ "model.layers.14.self_attn.out_proj.bias": "pytorch_model-00004-of-00006.bin",
115
+ "model.layers.14.self_attn.out_proj.weight": "pytorch_model-00004-of-00006.bin",
116
+ "model.layers.14.self_attn.q_proj.bias": "pytorch_model-00004-of-00006.bin",
117
+ "model.layers.14.self_attn.q_proj.weight": "pytorch_model-00004-of-00006.bin",
118
+ "model.layers.14.self_attn.v_proj.bias": "pytorch_model-00004-of-00006.bin",
119
+ "model.layers.14.self_attn.v_proj.weight": "pytorch_model-00004-of-00006.bin",
120
+ "model.layers.14.self_attn_layer_norm.bias": "pytorch_model-00004-of-00006.bin",
121
+ "model.layers.14.self_attn_layer_norm.weight": "pytorch_model-00004-of-00006.bin",
122
+ "model.layers.15.fc1.bias": "pytorch_model-00004-of-00006.bin",
123
+ "model.layers.15.fc1.weight": "pytorch_model-00004-of-00006.bin",
124
+ "model.layers.15.fc2.bias": "pytorch_model-00004-of-00006.bin",
125
+ "model.layers.15.fc2.weight": "pytorch_model-00004-of-00006.bin",
126
+ "model.layers.15.final_layer_norm.bias": "pytorch_model-00004-of-00006.bin",
127
+ "model.layers.15.final_layer_norm.weight": "pytorch_model-00004-of-00006.bin",
128
+ "model.layers.15.self_attn.k_proj.bias": "pytorch_model-00004-of-00006.bin",
129
+ "model.layers.15.self_attn.k_proj.weight": "pytorch_model-00004-of-00006.bin",
130
+ "model.layers.15.self_attn.out_proj.bias": "pytorch_model-00004-of-00006.bin",
131
+ "model.layers.15.self_attn.out_proj.weight": "pytorch_model-00004-of-00006.bin",
132
+ "model.layers.15.self_attn.q_proj.bias": "pytorch_model-00004-of-00006.bin",
133
+ "model.layers.15.self_attn.q_proj.weight": "pytorch_model-00004-of-00006.bin",
134
+ "model.layers.15.self_attn.v_proj.bias": "pytorch_model-00004-of-00006.bin",
135
+ "model.layers.15.self_attn.v_proj.weight": "pytorch_model-00004-of-00006.bin",
136
+ "model.layers.15.self_attn_layer_norm.bias": "pytorch_model-00004-of-00006.bin",
137
+ "model.layers.15.self_attn_layer_norm.weight": "pytorch_model-00004-of-00006.bin",
138
+ "model.layers.16.fc1.bias": "pytorch_model-00004-of-00006.bin",
139
+ "model.layers.16.fc1.weight": "pytorch_model-00004-of-00006.bin",
140
+ "model.layers.16.fc2.bias": "pytorch_model-00004-of-00006.bin",
141
+ "model.layers.16.fc2.weight": "pytorch_model-00004-of-00006.bin",
142
+ "model.layers.16.final_layer_norm.bias": "pytorch_model-00004-of-00006.bin",
143
+ "model.layers.16.final_layer_norm.weight": "pytorch_model-00004-of-00006.bin",
144
+ "model.layers.16.self_attn.k_proj.bias": "pytorch_model-00004-of-00006.bin",
145
+ "model.layers.16.self_attn.k_proj.weight": "pytorch_model-00004-of-00006.bin",
146
+ "model.layers.16.self_attn.out_proj.bias": "pytorch_model-00004-of-00006.bin",
147
+ "model.layers.16.self_attn.out_proj.weight": "pytorch_model-00004-of-00006.bin",
148
+ "model.layers.16.self_attn.q_proj.bias": "pytorch_model-00004-of-00006.bin",
149
+ "model.layers.16.self_attn.q_proj.weight": "pytorch_model-00004-of-00006.bin",
150
+ "model.layers.16.self_attn.v_proj.bias": "pytorch_model-00004-of-00006.bin",
151
+ "model.layers.16.self_attn.v_proj.weight": "pytorch_model-00004-of-00006.bin",
152
+ "model.layers.16.self_attn_layer_norm.bias": "pytorch_model-00004-of-00006.bin",
153
+ "model.layers.16.self_attn_layer_norm.weight": "pytorch_model-00004-of-00006.bin",
154
+ "model.layers.17.fc1.bias": "pytorch_model-00004-of-00006.bin",
155
+ "model.layers.17.fc1.weight": "pytorch_model-00004-of-00006.bin",
156
+ "model.layers.17.fc2.bias": "pytorch_model-00004-of-00006.bin",
157
+ "model.layers.17.fc2.weight": "pytorch_model-00004-of-00006.bin",
158
+ "model.layers.17.final_layer_norm.bias": "pytorch_model-00004-of-00006.bin",
159
+ "model.layers.17.final_layer_norm.weight": "pytorch_model-00004-of-00006.bin",
160
+ "model.layers.17.self_attn.k_proj.bias": "pytorch_model-00004-of-00006.bin",
161
+ "model.layers.17.self_attn.k_proj.weight": "pytorch_model-00004-of-00006.bin",
162
+ "model.layers.17.self_attn.out_proj.bias": "pytorch_model-00004-of-00006.bin",
163
+ "model.layers.17.self_attn.out_proj.weight": "pytorch_model-00004-of-00006.bin",
164
+ "model.layers.17.self_attn.q_proj.bias": "pytorch_model-00004-of-00006.bin",
165
+ "model.layers.17.self_attn.q_proj.weight": "pytorch_model-00004-of-00006.bin",
166
+ "model.layers.17.self_attn.v_proj.bias": "pytorch_model-00004-of-00006.bin",
167
+ "model.layers.17.self_attn.v_proj.weight": "pytorch_model-00004-of-00006.bin",
168
+ "model.layers.17.self_attn_layer_norm.bias": "pytorch_model-00004-of-00006.bin",
169
+ "model.layers.17.self_attn_layer_norm.weight": "pytorch_model-00004-of-00006.bin",
170
+ "model.layers.18.fc1.bias": "pytorch_model-00004-of-00006.bin",
171
+ "model.layers.18.fc1.weight": "pytorch_model-00004-of-00006.bin",
172
+ "model.layers.18.fc2.bias": "pytorch_model-00004-of-00006.bin",
173
+ "model.layers.18.fc2.weight": "pytorch_model-00004-of-00006.bin",
174
+ "model.layers.18.final_layer_norm.bias": "pytorch_model-00004-of-00006.bin",
175
+ "model.layers.18.final_layer_norm.weight": "pytorch_model-00004-of-00006.bin",
176
+ "model.layers.18.self_attn.k_proj.bias": "pytorch_model-00004-of-00006.bin",
177
+ "model.layers.18.self_attn.k_proj.weight": "pytorch_model-00004-of-00006.bin",
178
+ "model.layers.18.self_attn.out_proj.bias": "pytorch_model-00004-of-00006.bin",
179
+ "model.layers.18.self_attn.out_proj.weight": "pytorch_model-00004-of-00006.bin",
180
+ "model.layers.18.self_attn.q_proj.bias": "pytorch_model-00004-of-00006.bin",
181
+ "model.layers.18.self_attn.q_proj.weight": "pytorch_model-00004-of-00006.bin",
182
+ "model.layers.18.self_attn.v_proj.bias": "pytorch_model-00004-of-00006.bin",
183
+ "model.layers.18.self_attn.v_proj.weight": "pytorch_model-00004-of-00006.bin",
184
+ "model.layers.18.self_attn_layer_norm.bias": "pytorch_model-00004-of-00006.bin",
185
+ "model.layers.18.self_attn_layer_norm.weight": "pytorch_model-00004-of-00006.bin",
186
+ "model.layers.19.fc1.bias": "pytorch_model-00005-of-00006.bin",
187
+ "model.layers.19.fc1.weight": "pytorch_model-00005-of-00006.bin",
188
+ "model.layers.19.fc2.bias": "pytorch_model-00005-of-00006.bin",
189
+ "model.layers.19.fc2.weight": "pytorch_model-00005-of-00006.bin",
190
+ "model.layers.19.final_layer_norm.bias": "pytorch_model-00005-of-00006.bin",
191
+ "model.layers.19.final_layer_norm.weight": "pytorch_model-00005-of-00006.bin",
192
+ "model.layers.19.self_attn.k_proj.bias": "pytorch_model-00004-of-00006.bin",
193
+ "model.layers.19.self_attn.k_proj.weight": "pytorch_model-00004-of-00006.bin",
194
+ "model.layers.19.self_attn.out_proj.bias": "pytorch_model-00004-of-00006.bin",
195
+ "model.layers.19.self_attn.out_proj.weight": "pytorch_model-00004-of-00006.bin",
196
+ "model.layers.19.self_attn.q_proj.bias": "pytorch_model-00004-of-00006.bin",
197
+ "model.layers.19.self_attn.q_proj.weight": "pytorch_model-00004-of-00006.bin",
198
+ "model.layers.19.self_attn.v_proj.bias": "pytorch_model-00004-of-00006.bin",
199
+ "model.layers.19.self_attn.v_proj.weight": "pytorch_model-00004-of-00006.bin",
200
+ "model.layers.19.self_attn_layer_norm.bias": "pytorch_model-00004-of-00006.bin",
201
+ "model.layers.19.self_attn_layer_norm.weight": "pytorch_model-00004-of-00006.bin",
202
+ "model.layers.2.fc1.bias": "pytorch_model-00003-of-00006.bin",
203
+ "model.layers.2.fc1.weight": "pytorch_model-00003-of-00006.bin",
204
+ "model.layers.2.fc2.bias": "pytorch_model-00003-of-00006.bin",
205
+ "model.layers.2.fc2.weight": "pytorch_model-00003-of-00006.bin",
206
+ "model.layers.2.final_layer_norm.bias": "pytorch_model-00003-of-00006.bin",
207
+ "model.layers.2.final_layer_norm.weight": "pytorch_model-00003-of-00006.bin",
208
+ "model.layers.2.self_attn.k_proj.bias": "pytorch_model-00003-of-00006.bin",
209
+ "model.layers.2.self_attn.k_proj.weight": "pytorch_model-00003-of-00006.bin",
210
+ "model.layers.2.self_attn.out_proj.bias": "pytorch_model-00003-of-00006.bin",
211
+ "model.layers.2.self_attn.out_proj.weight": "pytorch_model-00003-of-00006.bin",
212
+ "model.layers.2.self_attn.q_proj.bias": "pytorch_model-00003-of-00006.bin",
213
+ "model.layers.2.self_attn.q_proj.weight": "pytorch_model-00003-of-00006.bin",
214
+ "model.layers.2.self_attn.v_proj.bias": "pytorch_model-00003-of-00006.bin",
215
+ "model.layers.2.self_attn.v_proj.weight": "pytorch_model-00003-of-00006.bin",
216
+ "model.layers.2.self_attn_layer_norm.bias": "pytorch_model-00003-of-00006.bin",
217
+ "model.layers.2.self_attn_layer_norm.weight": "pytorch_model-00003-of-00006.bin",
218
+ "model.layers.20.fc1.bias": "pytorch_model-00005-of-00006.bin",
219
+ "model.layers.20.fc1.weight": "pytorch_model-00005-of-00006.bin",
220
+ "model.layers.20.fc2.bias": "pytorch_model-00005-of-00006.bin",
221
+ "model.layers.20.fc2.weight": "pytorch_model-00005-of-00006.bin",
222
+ "model.layers.20.final_layer_norm.bias": "pytorch_model-00005-of-00006.bin",
223
+ "model.layers.20.final_layer_norm.weight": "pytorch_model-00005-of-00006.bin",
224
+ "model.layers.20.self_attn.k_proj.bias": "pytorch_model-00005-of-00006.bin",
225
+ "model.layers.20.self_attn.k_proj.weight": "pytorch_model-00005-of-00006.bin",
226
+ "model.layers.20.self_attn.out_proj.bias": "pytorch_model-00005-of-00006.bin",
227
+ "model.layers.20.self_attn.out_proj.weight": "pytorch_model-00005-of-00006.bin",
228
+ "model.layers.20.self_attn.q_proj.bias": "pytorch_model-00005-of-00006.bin",
229
+ "model.layers.20.self_attn.q_proj.weight": "pytorch_model-00005-of-00006.bin",
230
+ "model.layers.20.self_attn.v_proj.bias": "pytorch_model-00005-of-00006.bin",
231
+ "model.layers.20.self_attn.v_proj.weight": "pytorch_model-00005-of-00006.bin",
232
+ "model.layers.20.self_attn_layer_norm.bias": "pytorch_model-00005-of-00006.bin",
233
+ "model.layers.20.self_attn_layer_norm.weight": "pytorch_model-00005-of-00006.bin",
234
+ "model.layers.21.fc1.bias": "pytorch_model-00005-of-00006.bin",
235
+ "model.layers.21.fc1.weight": "pytorch_model-00005-of-00006.bin",
236
+ "model.layers.21.fc2.bias": "pytorch_model-00005-of-00006.bin",
237
+ "model.layers.21.fc2.weight": "pytorch_model-00005-of-00006.bin",
238
+ "model.layers.21.final_layer_norm.bias": "pytorch_model-00005-of-00006.bin",
239
+ "model.layers.21.final_layer_norm.weight": "pytorch_model-00005-of-00006.bin",
240
+ "model.layers.21.self_attn.k_proj.bias": "pytorch_model-00005-of-00006.bin",
241
+ "model.layers.21.self_attn.k_proj.weight": "pytorch_model-00005-of-00006.bin",
242
+ "model.layers.21.self_attn.out_proj.bias": "pytorch_model-00005-of-00006.bin",
243
+ "model.layers.21.self_attn.out_proj.weight": "pytorch_model-00005-of-00006.bin",
244
+ "model.layers.21.self_attn.q_proj.bias": "pytorch_model-00005-of-00006.bin",
245
+ "model.layers.21.self_attn.q_proj.weight": "pytorch_model-00005-of-00006.bin",
246
+ "model.layers.21.self_attn.v_proj.bias": "pytorch_model-00005-of-00006.bin",
247
+ "model.layers.21.self_attn.v_proj.weight": "pytorch_model-00005-of-00006.bin",
248
+ "model.layers.21.self_attn_layer_norm.bias": "pytorch_model-00005-of-00006.bin",
249
+ "model.layers.21.self_attn_layer_norm.weight": "pytorch_model-00005-of-00006.bin",
250
+ "model.layers.22.fc1.bias": "pytorch_model-00005-of-00006.bin",
251
+ "model.layers.22.fc1.weight": "pytorch_model-00005-of-00006.bin",
252
+ "model.layers.22.fc2.bias": "pytorch_model-00005-of-00006.bin",
253
+ "model.layers.22.fc2.weight": "pytorch_model-00005-of-00006.bin",
254
+ "model.layers.22.final_layer_norm.bias": "pytorch_model-00005-of-00006.bin",
255
+ "model.layers.22.final_layer_norm.weight": "pytorch_model-00005-of-00006.bin",
256
+ "model.layers.22.self_attn.k_proj.bias": "pytorch_model-00005-of-00006.bin",
257
+ "model.layers.22.self_attn.k_proj.weight": "pytorch_model-00005-of-00006.bin",
258
+ "model.layers.22.self_attn.out_proj.bias": "pytorch_model-00005-of-00006.bin",
259
+ "model.layers.22.self_attn.out_proj.weight": "pytorch_model-00005-of-00006.bin",
260
+ "model.layers.22.self_attn.q_proj.bias": "pytorch_model-00005-of-00006.bin",
261
+ "model.layers.22.self_attn.q_proj.weight": "pytorch_model-00005-of-00006.bin",
262
+ "model.layers.22.self_attn.v_proj.bias": "pytorch_model-00005-of-00006.bin",
263
+ "model.layers.22.self_attn.v_proj.weight": "pytorch_model-00005-of-00006.bin",
264
+ "model.layers.22.self_attn_layer_norm.bias": "pytorch_model-00005-of-00006.bin",
265
+ "model.layers.22.self_attn_layer_norm.weight": "pytorch_model-00005-of-00006.bin",
266
+ "model.layers.23.fc1.bias": "pytorch_model-00005-of-00006.bin",
267
+ "model.layers.23.fc1.weight": "pytorch_model-00005-of-00006.bin",
268
+ "model.layers.23.fc2.bias": "pytorch_model-00005-of-00006.bin",
269
+ "model.layers.23.fc2.weight": "pytorch_model-00005-of-00006.bin",
270
+ "model.layers.23.final_layer_norm.bias": "pytorch_model-00005-of-00006.bin",
271
+ "model.layers.23.final_layer_norm.weight": "pytorch_model-00005-of-00006.bin",
272
+ "model.layers.23.self_attn.k_proj.bias": "pytorch_model-00005-of-00006.bin",
273
+ "model.layers.23.self_attn.k_proj.weight": "pytorch_model-00005-of-00006.bin",
274
+ "model.layers.23.self_attn.out_proj.bias": "pytorch_model-00005-of-00006.bin",
275
+ "model.layers.23.self_attn.out_proj.weight": "pytorch_model-00005-of-00006.bin",
276
+ "model.layers.23.self_attn.q_proj.bias": "pytorch_model-00005-of-00006.bin",
277
+ "model.layers.23.self_attn.q_proj.weight": "pytorch_model-00005-of-00006.bin",
278
+ "model.layers.23.self_attn.v_proj.bias": "pytorch_model-00005-of-00006.bin",
279
+ "model.layers.23.self_attn.v_proj.weight": "pytorch_model-00005-of-00006.bin",
280
+ "model.layers.23.self_attn_layer_norm.bias": "pytorch_model-00005-of-00006.bin",
281
+ "model.layers.23.self_attn_layer_norm.weight": "pytorch_model-00005-of-00006.bin",
282
+ "model.layers.3.fc1.bias": "pytorch_model-00003-of-00006.bin",
283
+ "model.layers.3.fc1.weight": "pytorch_model-00003-of-00006.bin",
284
+ "model.layers.3.fc2.bias": "pytorch_model-00003-of-00006.bin",
285
+ "model.layers.3.fc2.weight": "pytorch_model-00003-of-00006.bin",
286
+ "model.layers.3.final_layer_norm.bias": "pytorch_model-00003-of-00006.bin",
287
+ "model.layers.3.final_layer_norm.weight": "pytorch_model-00003-of-00006.bin",
288
+ "model.layers.3.self_attn.k_proj.bias": "pytorch_model-00003-of-00006.bin",
289
+ "model.layers.3.self_attn.k_proj.weight": "pytorch_model-00003-of-00006.bin",
290
+ "model.layers.3.self_attn.out_proj.bias": "pytorch_model-00003-of-00006.bin",
291
+ "model.layers.3.self_attn.out_proj.weight": "pytorch_model-00003-of-00006.bin",
292
+ "model.layers.3.self_attn.q_proj.bias": "pytorch_model-00003-of-00006.bin",
293
+ "model.layers.3.self_attn.q_proj.weight": "pytorch_model-00003-of-00006.bin",
294
+ "model.layers.3.self_attn.v_proj.bias": "pytorch_model-00003-of-00006.bin",
295
+ "model.layers.3.self_attn.v_proj.weight": "pytorch_model-00003-of-00006.bin",
296
+ "model.layers.3.self_attn_layer_norm.bias": "pytorch_model-00003-of-00006.bin",
297
+ "model.layers.3.self_attn_layer_norm.weight": "pytorch_model-00003-of-00006.bin",
298
+ "model.layers.4.fc1.bias": "pytorch_model-00003-of-00006.bin",
299
+ "model.layers.4.fc1.weight": "pytorch_model-00003-of-00006.bin",
300
+ "model.layers.4.fc2.bias": "pytorch_model-00003-of-00006.bin",
301
+ "model.layers.4.fc2.weight": "pytorch_model-00003-of-00006.bin",
302
+ "model.layers.4.final_layer_norm.bias": "pytorch_model-00003-of-00006.bin",
303
+ "model.layers.4.final_layer_norm.weight": "pytorch_model-00003-of-00006.bin",
304
+ "model.layers.4.self_attn.k_proj.bias": "pytorch_model-00003-of-00006.bin",
305
+ "model.layers.4.self_attn.k_proj.weight": "pytorch_model-00003-of-00006.bin",
306
+ "model.layers.4.self_attn.out_proj.bias": "pytorch_model-00003-of-00006.bin",
307
+ "model.layers.4.self_attn.out_proj.weight": "pytorch_model-00003-of-00006.bin",
308
+ "model.layers.4.self_attn.q_proj.bias": "pytorch_model-00003-of-00006.bin",
309
+ "model.layers.4.self_attn.q_proj.weight": "pytorch_model-00003-of-00006.bin",
310
+ "model.layers.4.self_attn.v_proj.bias": "pytorch_model-00003-of-00006.bin",
311
+ "model.layers.4.self_attn.v_proj.weight": "pytorch_model-00003-of-00006.bin",
312
+ "model.layers.4.self_attn_layer_norm.bias": "pytorch_model-00003-of-00006.bin",
313
+ "model.layers.4.self_attn_layer_norm.weight": "pytorch_model-00003-of-00006.bin",
314
+ "model.layers.5.fc1.bias": "pytorch_model-00003-of-00006.bin",
315
+ "model.layers.5.fc1.weight": "pytorch_model-00003-of-00006.bin",
316
+ "model.layers.5.fc2.bias": "pytorch_model-00003-of-00006.bin",
317
+ "model.layers.5.fc2.weight": "pytorch_model-00003-of-00006.bin",
318
+ "model.layers.5.final_layer_norm.bias": "pytorch_model-00003-of-00006.bin",
319
+ "model.layers.5.final_layer_norm.weight": "pytorch_model-00003-of-00006.bin",
320
+ "model.layers.5.self_attn.k_proj.bias": "pytorch_model-00003-of-00006.bin",
321
+ "model.layers.5.self_attn.k_proj.weight": "pytorch_model-00003-of-00006.bin",
322
+ "model.layers.5.self_attn.out_proj.bias": "pytorch_model-00003-of-00006.bin",
323
+ "model.layers.5.self_attn.out_proj.weight": "pytorch_model-00003-of-00006.bin",
324
+ "model.layers.5.self_attn.q_proj.bias": "pytorch_model-00003-of-00006.bin",
325
+ "model.layers.5.self_attn.q_proj.weight": "pytorch_model-00003-of-00006.bin",
326
+ "model.layers.5.self_attn.v_proj.bias": "pytorch_model-00003-of-00006.bin",
327
+ "model.layers.5.self_attn.v_proj.weight": "pytorch_model-00003-of-00006.bin",
328
+ "model.layers.5.self_attn_layer_norm.bias": "pytorch_model-00003-of-00006.bin",
329
+ "model.layers.5.self_attn_layer_norm.weight": "pytorch_model-00003-of-00006.bin",
330
+ "model.layers.6.fc1.bias": "pytorch_model-00003-of-00006.bin",
331
+ "model.layers.6.fc1.weight": "pytorch_model-00003-of-00006.bin",
332
+ "model.layers.6.fc2.bias": "pytorch_model-00003-of-00006.bin",
333
+ "model.layers.6.fc2.weight": "pytorch_model-00003-of-00006.bin",
334
+ "model.layers.6.final_layer_norm.bias": "pytorch_model-00003-of-00006.bin",
335
+ "model.layers.6.final_layer_norm.weight": "pytorch_model-00003-of-00006.bin",
336
+ "model.layers.6.self_attn.k_proj.bias": "pytorch_model-00003-of-00006.bin",
337
+ "model.layers.6.self_attn.k_proj.weight": "pytorch_model-00003-of-00006.bin",
338
+ "model.layers.6.self_attn.out_proj.bias": "pytorch_model-00003-of-00006.bin",
339
+ "model.layers.6.self_attn.out_proj.weight": "pytorch_model-00003-of-00006.bin",
340
+ "model.layers.6.self_attn.q_proj.bias": "pytorch_model-00003-of-00006.bin",
341
+ "model.layers.6.self_attn.q_proj.weight": "pytorch_model-00003-of-00006.bin",
342
+ "model.layers.6.self_attn.v_proj.bias": "pytorch_model-00003-of-00006.bin",
343
+ "model.layers.6.self_attn.v_proj.weight": "pytorch_model-00003-of-00006.bin",
344
+ "model.layers.6.self_attn_layer_norm.bias": "pytorch_model-00003-of-00006.bin",
345
+ "model.layers.6.self_attn_layer_norm.weight": "pytorch_model-00003-of-00006.bin",
346
+ "model.layers.7.fc1.bias": "pytorch_model-00003-of-00006.bin",
347
+ "model.layers.7.fc1.weight": "pytorch_model-00003-of-00006.bin",
348
+ "model.layers.7.fc2.bias": "pytorch_model-00003-of-00006.bin",
349
+ "model.layers.7.fc2.weight": "pytorch_model-00003-of-00006.bin",
350
+ "model.layers.7.final_layer_norm.bias": "pytorch_model-00003-of-00006.bin",
351
+ "model.layers.7.final_layer_norm.weight": "pytorch_model-00003-of-00006.bin",
352
+ "model.layers.7.self_attn.k_proj.bias": "pytorch_model-00003-of-00006.bin",
353
+ "model.layers.7.self_attn.k_proj.weight": "pytorch_model-00003-of-00006.bin",
354
+ "model.layers.7.self_attn.out_proj.bias": "pytorch_model-00003-of-00006.bin",
355
+ "model.layers.7.self_attn.out_proj.weight": "pytorch_model-00003-of-00006.bin",
356
+ "model.layers.7.self_attn.q_proj.bias": "pytorch_model-00003-of-00006.bin",
357
+ "model.layers.7.self_attn.q_proj.weight": "pytorch_model-00003-of-00006.bin",
358
+ "model.layers.7.self_attn.v_proj.bias": "pytorch_model-00003-of-00006.bin",
359
+ "model.layers.7.self_attn.v_proj.weight": "pytorch_model-00003-of-00006.bin",
360
+ "model.layers.7.self_attn_layer_norm.bias": "pytorch_model-00003-of-00006.bin",
361
+ "model.layers.7.self_attn_layer_norm.weight": "pytorch_model-00003-of-00006.bin",
362
+ "model.layers.8.fc1.bias": "pytorch_model-00003-of-00006.bin",
363
+ "model.layers.8.fc1.weight": "pytorch_model-00003-of-00006.bin",
364
+ "model.layers.8.fc2.bias": "pytorch_model-00003-of-00006.bin",
365
+ "model.layers.8.fc2.weight": "pytorch_model-00003-of-00006.bin",
366
+ "model.layers.8.final_layer_norm.bias": "pytorch_model-00003-of-00006.bin",
367
+ "model.layers.8.final_layer_norm.weight": "pytorch_model-00003-of-00006.bin",
368
+ "model.layers.8.self_attn.k_proj.bias": "pytorch_model-00003-of-00006.bin",
369
+ "model.layers.8.self_attn.k_proj.weight": "pytorch_model-00003-of-00006.bin",
370
+ "model.layers.8.self_attn.out_proj.bias": "pytorch_model-00003-of-00006.bin",
371
+ "model.layers.8.self_attn.out_proj.weight": "pytorch_model-00003-of-00006.bin",
372
+ "model.layers.8.self_attn.q_proj.bias": "pytorch_model-00003-of-00006.bin",
373
+ "model.layers.8.self_attn.q_proj.weight": "pytorch_model-00003-of-00006.bin",
374
+ "model.layers.8.self_attn.v_proj.bias": "pytorch_model-00003-of-00006.bin",
375
+ "model.layers.8.self_attn.v_proj.weight": "pytorch_model-00003-of-00006.bin",
376
+ "model.layers.8.self_attn_layer_norm.bias": "pytorch_model-00003-of-00006.bin",
377
+ "model.layers.8.self_attn_layer_norm.weight": "pytorch_model-00003-of-00006.bin",
378
+ "model.layers.9.fc1.bias": "pytorch_model-00003-of-00006.bin",
379
+ "model.layers.9.fc1.weight": "pytorch_model-00003-of-00006.bin",
380
+ "model.layers.9.fc2.bias": "pytorch_model-00004-of-00006.bin",
381
+ "model.layers.9.fc2.weight": "pytorch_model-00004-of-00006.bin",
382
+ "model.layers.9.final_layer_norm.bias": "pytorch_model-00004-of-00006.bin",
383
+ "model.layers.9.final_layer_norm.weight": "pytorch_model-00004-of-00006.bin",
384
+ "model.layers.9.self_attn.k_proj.bias": "pytorch_model-00003-of-00006.bin",
385
+ "model.layers.9.self_attn.k_proj.weight": "pytorch_model-00003-of-00006.bin",
386
+ "model.layers.9.self_attn.out_proj.bias": "pytorch_model-00003-of-00006.bin",
387
+ "model.layers.9.self_attn.out_proj.weight": "pytorch_model-00003-of-00006.bin",
388
+ "model.layers.9.self_attn.q_proj.bias": "pytorch_model-00003-of-00006.bin",
389
+ "model.layers.9.self_attn.q_proj.weight": "pytorch_model-00003-of-00006.bin",
390
+ "model.layers.9.self_attn.v_proj.bias": "pytorch_model-00003-of-00006.bin",
391
+ "model.layers.9.self_attn.v_proj.weight": "pytorch_model-00003-of-00006.bin",
392
+ "model.layers.9.self_attn_layer_norm.bias": "pytorch_model-00003-of-00006.bin",
393
+ "model.layers.9.self_attn_layer_norm.weight": "pytorch_model-00003-of-00006.bin"
394
+ }
395
+ }