Hugging Face's logo

naxautify
/
gpt2-8k-pile

Text Generation
Transformers
PyTorch
gpt2
text-generation-inference
Inference Endpoints
Model card Files Community
naxautify commited on
Commit
42fdf88
•
1 Parent(s): 19db37f

Upload model

Browse files
Files changed (11) hide show
  1. config.json +39 -0
  2. generation_config.json +6 -0
  3. pytorch_model-00001-of-00008.bin +3 -0
  4. pytorch_model-00002-of-00008.bin +3 -0
  5. pytorch_model-00003-of-00008.bin +3 -0
  6. pytorch_model-00004-of-00008.bin +3 -0
  7. pytorch_model-00005-of-00008.bin +3 -0
  8. pytorch_model-00006-of-00008.bin +3 -0
  9. pytorch_model-00007-of-00008.bin +3 -0
  10. pytorch_model-00008-of-00008.bin +3 -0
  11. pytorch_model.bin.index.json +180 -0
config.json ADDED
@@ -0,0 +1,39 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "gpt2",
3
+ "activation_function": "gelu_new",
4
+ "architectures": [
5
+ "GPT2LMHeadModel"
6
+ ],
7
+ "attn_pdrop": 0.1,
8
+ "bos_token_id": 50256,
9
+ "embd_pdrop": 0.1,
10
+ "eos_token_id": 50256,
11
+ "initializer_range": 0.02,
12
+ "layer_norm_epsilon": 1e-05,
13
+ "model_type": "gpt2",
14
+ "n_ctx": 8192,
15
+ "n_embd": 768,
16
+ "n_head": 12,
17
+ "n_inner": null,
18
+ "n_layer": 12,
19
+ "n_positions": 8192,
20
+ "reorder_and_upcast_attn": false,
21
+ "resid_pdrop": 0.1,
22
+ "scale_attn_by_inverse_layer_idx": false,
23
+ "scale_attn_weights": true,
24
+ "summary_activation": null,
25
+ "summary_first_dropout": 0.1,
26
+ "summary_proj_to_labels": true,
27
+ "summary_type": "cls_index",
28
+ "summary_use_proj": true,
29
+ "task_specific_params": {
30
+ "text-generation": {
31
+ "do_sample": true,
32
+ "max_length": 50
33
+ }
34
+ },
35
+ "torch_dtype": "float32",
36
+ "transformers_version": "4.27.1",
37
+ "use_cache": true,
38
+ "vocab_size": 50257
39
+ }
generation_config.json ADDED
@@ -0,0 +1,6 @@
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 50256,
4
+ "eos_token_id": 50256,
5
+ "transformers_version": "4.27.1"
6
+ }
pytorch_model-00001-of-00008.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:152b5902f740b80cdf4978d9b7e9d836913d898437de02b851919273c39ea9c7
3
+ size 537
pytorch_model-00002-of-00008.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:39c1b06d840981f1c0e58ab3000e848df290d067bd34aef834aeb8b35bd36674
3
+ size 154390442
pytorch_model-00003-of-00008.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bcfdd0f1e000d54bc6a895944a892dfaf25e8018876e16d6469385353bad5bf1
3
+ size 216103037
pytorch_model-00004-of-00008.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:724d4466e07355b390113c04a41f2159fcbf07687b4d662f8f6499b35b0e9e5d
3
+ size 267497303
pytorch_model-00005-of-00008.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:399e8085632f832eba0a49be42f527c4ea824a4690e5a0d6148ad160d31bed09
3
+ size 209828019
pytorch_model-00006-of-00008.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6dae9a8f3a9242098224cdebede998689d07a2aa37ccec504b3e3c1e1db4838e
3
+ size 267497303
pytorch_model-00007-of-00008.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ac93e4f23bef5a2dc6b93c15ac84c996dfa5e2437c820630e2db1980debeb8d0
3
+ size 209828019
pytorch_model-00008-of-00008.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a4a41ff3f28b9a8eb18af248f9249cdb20fafc37f6ac1f31d348ff656da45254
3
+ size 154390442
pytorch_model.bin.index.json ADDED
@@ -0,0 +1,180 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metadata": {
3
+ "total_size": 774832176.0
4
+ },
5
+ "weight_map": {
6
+ "lm_head.weight": "pytorch_model-00008-of-00008.bin",
7
+ "transformer.h.0.attn.bias": "pytorch_model-00003-of-00008.bin",
8
+ "transformer.h.0.attn.c_attn.bias": "pytorch_model-00003-of-00008.bin",
9
+ "transformer.h.0.attn.c_attn.weight": "pytorch_model-00003-of-00008.bin",
10
+ "transformer.h.0.attn.c_proj.bias": "pytorch_model-00003-of-00008.bin",
11
+ "transformer.h.0.attn.c_proj.weight": "pytorch_model-00003-of-00008.bin",
12
+ "transformer.h.0.attn.masked_bias": "pytorch_model-00003-of-00008.bin",
13
+ "transformer.h.0.ln_1.bias": "pytorch_model-00003-of-00008.bin",
14
+ "transformer.h.0.ln_1.weight": "pytorch_model-00003-of-00008.bin",
15
+ "transformer.h.0.ln_2.bias": "pytorch_model-00003-of-00008.bin",
16
+ "transformer.h.0.ln_2.weight": "pytorch_model-00003-of-00008.bin",
17
+ "transformer.h.0.mlp.c_fc.bias": "pytorch_model-00003-of-00008.bin",
18
+ "transformer.h.0.mlp.c_fc.weight": "pytorch_model-00003-of-00008.bin",
19
+ "transformer.h.0.mlp.c_proj.bias": "pytorch_model-00003-of-00008.bin",
20
+ "transformer.h.0.mlp.c_proj.weight": "pytorch_model-00003-of-00008.bin",
21
+ "transformer.h.1.attn.bias": "pytorch_model-00003-of-00008.bin",
22
+ "transformer.h.1.attn.c_attn.bias": "pytorch_model-00003-of-00008.bin",
23
+ "transformer.h.1.attn.c_attn.weight": "pytorch_model-00003-of-00008.bin",
24
+ "transformer.h.1.attn.c_proj.bias": "pytorch_model-00003-of-00008.bin",
25
+ "transformer.h.1.attn.c_proj.weight": "pytorch_model-00003-of-00008.bin",
26
+ "transformer.h.1.attn.masked_bias": "pytorch_model-00003-of-00008.bin",
27
+ "transformer.h.1.ln_1.bias": "pytorch_model-00003-of-00008.bin",
28
+ "transformer.h.1.ln_1.weight": "pytorch_model-00003-of-00008.bin",
29
+ "transformer.h.1.ln_2.bias": "pytorch_model-00003-of-00008.bin",
30
+ "transformer.h.1.ln_2.weight": "pytorch_model-00003-of-00008.bin",
31
+ "transformer.h.1.mlp.c_fc.bias": "pytorch_model-00003-of-00008.bin",
32
+ "transformer.h.1.mlp.c_fc.weight": "pytorch_model-00003-of-00008.bin",
33
+ "transformer.h.1.mlp.c_proj.bias": "pytorch_model-00003-of-00008.bin",
34
+ "transformer.h.1.mlp.c_proj.weight": "pytorch_model-00003-of-00008.bin",
35
+ "transformer.h.10.attn.bias": "pytorch_model-00007-of-00008.bin",
36
+ "transformer.h.10.attn.c_attn.bias": "pytorch_model-00007-of-00008.bin",
37
+ "transformer.h.10.attn.c_attn.weight": "pytorch_model-00007-of-00008.bin",
38
+ "transformer.h.10.attn.c_proj.bias": "pytorch_model-00007-of-00008.bin",
39
+ "transformer.h.10.attn.c_proj.weight": "pytorch_model-00007-of-00008.bin",
40
+ "transformer.h.10.attn.masked_bias": "pytorch_model-00007-of-00008.bin",
41
+ "transformer.h.10.ln_1.bias": "pytorch_model-00007-of-00008.bin",
42
+ "transformer.h.10.ln_1.weight": "pytorch_model-00007-of-00008.bin",
43
+ "transformer.h.10.ln_2.bias": "pytorch_model-00007-of-00008.bin",
44
+ "transformer.h.10.ln_2.weight": "pytorch_model-00007-of-00008.bin",
45
+ "transformer.h.10.mlp.c_fc.bias": "pytorch_model-00007-of-00008.bin",
46
+ "transformer.h.10.mlp.c_fc.weight": "pytorch_model-00007-of-00008.bin",
47
+ "transformer.h.10.mlp.c_proj.bias": "pytorch_model-00007-of-00008.bin",
48
+ "transformer.h.10.mlp.c_proj.weight": "pytorch_model-00007-of-00008.bin",
49
+ "transformer.h.11.attn.bias": "pytorch_model-00007-of-00008.bin",
50
+ "transformer.h.11.attn.c_attn.bias": "pytorch_model-00007-of-00008.bin",
51
+ "transformer.h.11.attn.c_attn.weight": "pytorch_model-00007-of-00008.bin",
52
+ "transformer.h.11.attn.c_proj.bias": "pytorch_model-00007-of-00008.bin",
53
+ "transformer.h.11.attn.c_proj.weight": "pytorch_model-00007-of-00008.bin",
54
+ "transformer.h.11.attn.masked_bias": "pytorch_model-00007-of-00008.bin",
55
+ "transformer.h.11.ln_1.bias": "pytorch_model-00007-of-00008.bin",
56
+ "transformer.h.11.ln_1.weight": "pytorch_model-00007-of-00008.bin",
57
+ "transformer.h.11.ln_2.bias": "pytorch_model-00007-of-00008.bin",
58
+ "transformer.h.11.ln_2.weight": "pytorch_model-00007-of-00008.bin",
59
+ "transformer.h.11.mlp.c_fc.bias": "pytorch_model-00007-of-00008.bin",
60
+ "transformer.h.11.mlp.c_fc.weight": "pytorch_model-00007-of-00008.bin",
61
+ "transformer.h.11.mlp.c_proj.bias": "pytorch_model-00007-of-00008.bin",
62
+ "transformer.h.11.mlp.c_proj.weight": "pytorch_model-00007-of-00008.bin",
63
+ "transformer.h.2.attn.bias": "pytorch_model-00004-of-00008.bin",
64
+ "transformer.h.2.attn.c_attn.bias": "pytorch_model-00004-of-00008.bin",
65
+ "transformer.h.2.attn.c_attn.weight": "pytorch_model-00004-of-00008.bin",
66
+ "transformer.h.2.attn.c_proj.bias": "pytorch_model-00004-of-00008.bin",
67
+ "transformer.h.2.attn.c_proj.weight": "pytorch_model-00004-of-00008.bin",
68
+ "transformer.h.2.attn.masked_bias": "pytorch_model-00004-of-00008.bin",
69
+ "transformer.h.2.ln_1.bias": "pytorch_model-00003-of-00008.bin",
70
+ "transformer.h.2.ln_1.weight": "pytorch_model-00003-of-00008.bin",
71
+ "transformer.h.2.ln_2.bias": "pytorch_model-00004-of-00008.bin",
72
+ "transformer.h.2.ln_2.weight": "pytorch_model-00004-of-00008.bin",
73
+ "transformer.h.2.mlp.c_fc.bias": "pytorch_model-00004-of-00008.bin",
74
+ "transformer.h.2.mlp.c_fc.weight": "pytorch_model-00004-of-00008.bin",
75
+ "transformer.h.2.mlp.c_proj.bias": "pytorch_model-00004-of-00008.bin",
76
+ "transformer.h.2.mlp.c_proj.weight": "pytorch_model-00004-of-00008.bin",
77
+ "transformer.h.3.attn.bias": "pytorch_model-00004-of-00008.bin",
78
+ "transformer.h.3.attn.c_attn.bias": "pytorch_model-00004-of-00008.bin",
79
+ "transformer.h.3.attn.c_attn.weight": "pytorch_model-00004-of-00008.bin",
80
+ "transformer.h.3.attn.c_proj.bias": "pytorch_model-00004-of-00008.bin",
81
+ "transformer.h.3.attn.c_proj.weight": "pytorch_model-00004-of-00008.bin",
82
+ "transformer.h.3.attn.masked_bias": "pytorch_model-00004-of-00008.bin",
83
+ "transformer.h.3.ln_1.bias": "pytorch_model-00004-of-00008.bin",
84
+ "transformer.h.3.ln_1.weight": "pytorch_model-00004-of-00008.bin",
85
+ "transformer.h.3.ln_2.bias": "pytorch_model-00004-of-00008.bin",
86
+ "transformer.h.3.ln_2.weight": "pytorch_model-00004-of-00008.bin",
87
+ "transformer.h.3.mlp.c_fc.bias": "pytorch_model-00004-of-00008.bin",
88
+ "transformer.h.3.mlp.c_fc.weight": "pytorch_model-00004-of-00008.bin",
89
+ "transformer.h.3.mlp.c_proj.bias": "pytorch_model-00004-of-00008.bin",
90
+ "transformer.h.3.mlp.c_proj.weight": "pytorch_model-00004-of-00008.bin",
91
+ "transformer.h.4.attn.bias": "pytorch_model-00004-of-00008.bin",
92
+ "transformer.h.4.attn.c_attn.bias": "pytorch_model-00004-of-00008.bin",
93
+ "transformer.h.4.attn.c_attn.weight": "pytorch_model-00004-of-00008.bin",
94
+ "transformer.h.4.attn.c_proj.bias": "pytorch_model-00004-of-00008.bin",
95
+ "transformer.h.4.attn.c_proj.weight": "pytorch_model-00004-of-00008.bin",
96
+ "transformer.h.4.attn.masked_bias": "pytorch_model-00004-of-00008.bin",
97
+ "transformer.h.4.ln_1.bias": "pytorch_model-00004-of-00008.bin",
98
+ "transformer.h.4.ln_1.weight": "pytorch_model-00004-of-00008.bin",
99
+ "transformer.h.4.ln_2.bias": "pytorch_model-00004-of-00008.bin",
100
+ "transformer.h.4.ln_2.weight": "pytorch_model-00004-of-00008.bin",
101
+ "transformer.h.4.mlp.c_fc.bias": "pytorch_model-00005-of-00008.bin",
102
+ "transformer.h.4.mlp.c_fc.weight": "pytorch_model-00005-of-00008.bin",
103
+ "transformer.h.4.mlp.c_proj.bias": "pytorch_model-00005-of-00008.bin",
104
+ "transformer.h.4.mlp.c_proj.weight": "pytorch_model-00005-of-00008.bin",
105
+ "transformer.h.5.attn.bias": "pytorch_model-00005-of-00008.bin",
106
+ "transformer.h.5.attn.c_attn.bias": "pytorch_model-00005-of-00008.bin",
107
+ "transformer.h.5.attn.c_attn.weight": "pytorch_model-00005-of-00008.bin",
108
+ "transformer.h.5.attn.c_proj.bias": "pytorch_model-00005-of-00008.bin",
109
+ "transformer.h.5.attn.c_proj.weight": "pytorch_model-00005-of-00008.bin",
110
+ "transformer.h.5.attn.masked_bias": "pytorch_model-00005-of-00008.bin",
111
+ "transformer.h.5.ln_1.bias": "pytorch_model-00005-of-00008.bin",
112
+ "transformer.h.5.ln_1.weight": "pytorch_model-00005-of-00008.bin",
113
+ "transformer.h.5.ln_2.bias": "pytorch_model-00005-of-00008.bin",
114
+ "transformer.h.5.ln_2.weight": "pytorch_model-00005-of-00008.bin",
115
+ "transformer.h.5.mlp.c_fc.bias": "pytorch_model-00005-of-00008.bin",
116
+ "transformer.h.5.mlp.c_fc.weight": "pytorch_model-00005-of-00008.bin",
117
+ "transformer.h.5.mlp.c_proj.bias": "pytorch_model-00005-of-00008.bin",
118
+ "transformer.h.5.mlp.c_proj.weight": "pytorch_model-00005-of-00008.bin",
119
+ "transformer.h.6.attn.bias": "pytorch_model-00005-of-00008.bin",
120
+ "transformer.h.6.attn.c_attn.bias": "pytorch_model-00005-of-00008.bin",
121
+ "transformer.h.6.attn.c_attn.weight": "pytorch_model-00005-of-00008.bin",
122
+ "transformer.h.6.attn.c_proj.bias": "pytorch_model-00005-of-00008.bin",
123
+ "transformer.h.6.attn.c_proj.weight": "pytorch_model-00005-of-00008.bin",
124
+ "transformer.h.6.attn.masked_bias": "pytorch_model-00005-of-00008.bin",
125
+ "transformer.h.6.ln_1.bias": "pytorch_model-00005-of-00008.bin",
126
+ "transformer.h.6.ln_1.weight": "pytorch_model-00005-of-00008.bin",
127
+ "transformer.h.6.ln_2.bias": "pytorch_model-00005-of-00008.bin",
128
+ "transformer.h.6.ln_2.weight": "pytorch_model-00005-of-00008.bin",
129
+ "transformer.h.6.mlp.c_fc.bias": "pytorch_model-00005-of-00008.bin",
130
+ "transformer.h.6.mlp.c_fc.weight": "pytorch_model-00005-of-00008.bin",
131
+ "transformer.h.6.mlp.c_proj.bias": "pytorch_model-00005-of-00008.bin",
132
+ "transformer.h.6.mlp.c_proj.weight": "pytorch_model-00005-of-00008.bin",
133
+ "transformer.h.7.attn.bias": "pytorch_model-00006-of-00008.bin",
134
+ "transformer.h.7.attn.c_attn.bias": "pytorch_model-00006-of-00008.bin",
135
+ "transformer.h.7.attn.c_attn.weight": "pytorch_model-00006-of-00008.bin",
136
+ "transformer.h.7.attn.c_proj.bias": "pytorch_model-00006-of-00008.bin",
137
+ "transformer.h.7.attn.c_proj.weight": "pytorch_model-00006-of-00008.bin",
138
+ "transformer.h.7.attn.masked_bias": "pytorch_model-00006-of-00008.bin",
139
+ "transformer.h.7.ln_1.bias": "pytorch_model-00005-of-00008.bin",
140
+ "transformer.h.7.ln_1.weight": "pytorch_model-00005-of-00008.bin",
141
+ "transformer.h.7.ln_2.bias": "pytorch_model-00006-of-00008.bin",
142
+ "transformer.h.7.ln_2.weight": "pytorch_model-00006-of-00008.bin",
143
+ "transformer.h.7.mlp.c_fc.bias": "pytorch_model-00006-of-00008.bin",
144
+ "transformer.h.7.mlp.c_fc.weight": "pytorch_model-00006-of-00008.bin",
145
+ "transformer.h.7.mlp.c_proj.bias": "pytorch_model-00006-of-00008.bin",
146
+ "transformer.h.7.mlp.c_proj.weight": "pytorch_model-00006-of-00008.bin",
147
+ "transformer.h.8.attn.bias": "pytorch_model-00006-of-00008.bin",
148
+ "transformer.h.8.attn.c_attn.bias": "pytorch_model-00006-of-00008.bin",
149
+ "transformer.h.8.attn.c_attn.weight": "pytorch_model-00006-of-00008.bin",
150
+ "transformer.h.8.attn.c_proj.bias": "pytorch_model-00006-of-00008.bin",
151
+ "transformer.h.8.attn.c_proj.weight": "pytorch_model-00006-of-00008.bin",
152
+ "transformer.h.8.attn.masked_bias": "pytorch_model-00006-of-00008.bin",
153
+ "transformer.h.8.ln_1.bias": "pytorch_model-00006-of-00008.bin",
154
+ "transformer.h.8.ln_1.weight": "pytorch_model-00006-of-00008.bin",
155
+ "transformer.h.8.ln_2.bias": "pytorch_model-00006-of-00008.bin",
156
+ "transformer.h.8.ln_2.weight": "pytorch_model-00006-of-00008.bin",
157
+ "transformer.h.8.mlp.c_fc.bias": "pytorch_model-00006-of-00008.bin",
158
+ "transformer.h.8.mlp.c_fc.weight": "pytorch_model-00006-of-00008.bin",
159
+ "transformer.h.8.mlp.c_proj.bias": "pytorch_model-00006-of-00008.bin",
160
+ "transformer.h.8.mlp.c_proj.weight": "pytorch_model-00006-of-00008.bin",
161
+ "transformer.h.9.attn.bias": "pytorch_model-00006-of-00008.bin",
162
+ "transformer.h.9.attn.c_attn.bias": "pytorch_model-00006-of-00008.bin",
163
+ "transformer.h.9.attn.c_attn.weight": "pytorch_model-00006-of-00008.bin",
164
+ "transformer.h.9.attn.c_proj.bias": "pytorch_model-00006-of-00008.bin",
165
+ "transformer.h.9.attn.c_proj.weight": "pytorch_model-00006-of-00008.bin",
166
+ "transformer.h.9.attn.masked_bias": "pytorch_model-00006-of-00008.bin",
167
+ "transformer.h.9.ln_1.bias": "pytorch_model-00006-of-00008.bin",
168
+ "transformer.h.9.ln_1.weight": "pytorch_model-00006-of-00008.bin",
169
+ "transformer.h.9.ln_2.bias": "pytorch_model-00006-of-00008.bin",
170
+ "transformer.h.9.ln_2.weight": "pytorch_model-00006-of-00008.bin",
171
+ "transformer.h.9.mlp.c_fc.bias": "pytorch_model-00007-of-00008.bin",
172
+ "transformer.h.9.mlp.c_fc.weight": "pytorch_model-00007-of-00008.bin",
173
+ "transformer.h.9.mlp.c_proj.bias": "pytorch_model-00007-of-00008.bin",
174
+ "transformer.h.9.mlp.c_proj.weight": "pytorch_model-00007-of-00008.bin",
175
+ "transformer.ln_f.bias": "pytorch_model-00007-of-00008.bin",
176
+ "transformer.ln_f.weight": "pytorch_model-00007-of-00008.bin",
177
+ "transformer.wpe.weight": "pytorch_model-00003-of-00008.bin",
178
+ "transformer.wte.weight": "pytorch_model-00002-of-00008.bin"
179
+ }
180
+ }