Hugging Face's logo

normalcomputing
/
extended-mind-mpt-30b

Text Generation
Transformers
PyTorch
extended-mpt
custom_code
Model card Files Community
phoebeklett commited on
Commit
c8dde66
•
1 Parent(s): 75296bd

Upload ExtendedMptForCausalLM

Browse files
Files changed (16) hide show
  1. config.json +82 -0
  2. generation_config.json +5 -0
  3. pytorch_model-00001-of-00013.bin +3 -0
  4. pytorch_model-00002-of-00013.bin +3 -0
  5. pytorch_model-00003-of-00013.bin +3 -0
  6. pytorch_model-00004-of-00013.bin +3 -0
  7. pytorch_model-00005-of-00013.bin +3 -0
  8. pytorch_model-00006-of-00013.bin +3 -0
  9. pytorch_model-00007-of-00013.bin +3 -0
  10. pytorch_model-00008-of-00013.bin +3 -0
  11. pytorch_model-00009-of-00013.bin +3 -0
  12. pytorch_model-00010-of-00013.bin +3 -0
  13. pytorch_model-00011-of-00013.bin +3 -0
  14. pytorch_model-00012-of-00013.bin +3 -0
  15. pytorch_model-00013-of-00013.bin +3 -0
  16. pytorch_model.bin.index.json +298 -0
config.json ADDED
@@ -0,0 +1,82 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "mosaicml/mpt-30b",
3
+ "architectures": [
4
+ "ExtendedMptForCausalLM"
5
+ ],
6
+ "attn_config": {
7
+ "model_type": ""
8
+ },
9
+ "auto_map": {
10
+ "AutoConfig": "mosaicml/mpt-30b--configuration_mpt.MPTConfig",
11
+ "AutoModelForCausalLM": "mosaicml/mpt-30b--modeling_mpt.MPTForCausalLM"
12
+ },
13
+ "d_model": 7168,
14
+ "emb_pdrop": 0,
15
+ "embedding_fraction": 1.0,
16
+ "expansion_ratio": 4,
17
+ "init_config": {
18
+ "emb_init_std": null,
19
+ "emb_init_uniform_lim": null,
20
+ "fan_mode": "fan_in",
21
+ "init_div_is_residual": true,
22
+ "init_gain": 0.0,
23
+ "init_nonlinearity": "relu",
24
+ "init_std": null,
25
+ "name": "kaiming_normal_",
26
+ "verbose": 0
27
+ },
28
+ "init_device": "cpu",
29
+ "initializer_range": 0.02,
30
+ "layer_norm_epsilon": 1e-05,
31
+ "learned_pos_emb": true,
32
+ "logit_scale": null,
33
+ "max_seq_len": 8192,
34
+ "max_seq_len_train": 2048,
35
+ "model_type": "extended-mpt",
36
+ "n_heads": 64,
37
+ "n_layers": 48,
38
+ "no_bias": true,
39
+ "norm_type": "low_precision_layernorm",
40
+ "resid_pdrop": 0,
41
+ "tokenizer_name": "EleutherAI/gpt-neox-20b",
42
+ "torch_dtype": "float32",
43
+ "transformers_version": "4.33.0",
44
+ "use_cache": false,
45
+ "use_external_mind": true,
46
+ "use_external_mind_by_layer": [
47
+ true,
48
+ true,
49
+ true,
50
+ true,
51
+ true,
52
+ true,
53
+ true,
54
+ true,
55
+ true,
56
+ true,
57
+ true,
58
+ true,
59
+ true,
60
+ true,
61
+ true,
62
+ true,
63
+ true,
64
+ true,
65
+ true,
66
+ true,
67
+ true,
68
+ true,
69
+ true,
70
+ true,
71
+ true,
72
+ true,
73
+ true,
74
+ true,
75
+ true,
76
+ true,
77
+ true,
78
+ true
79
+ ],
80
+ "verbose": 0,
81
+ "vocab_size": 50432
82
+ }
generation_config.json ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "transformers_version": "4.33.0",
4
+ "use_cache": false
5
+ }
pytorch_model-00001-of-00013.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e75efb0af67d4e133d58485ac32399061d29f119fe7ca640a1a148a58946be8c
3
+ size 9667059549
pytorch_model-00002-of-00013.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6ba1cb975f1b3348894b7167a6110f8985118068966a709ad21dac9ffb7a30df
3
+ size 9865240711
pytorch_model-00003-of-00013.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f9c60cb629d2267148f2a85bc91f3e484a159cc8ba98aeca1312de3404399b43
3
+ size 9865240711
pytorch_model-00004-of-00013.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cba055b421bb5bc7bfa3781888cfbc7fca85c27922ee807f76ea71869b56287a
3
+ size 9865240711
pytorch_model-00005-of-00013.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7e14448673fe46a51e2a19d621207ed80f5e0bc80df01624051e852e4b118018
3
+ size 9865240711
pytorch_model-00006-of-00013.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4226ce5a36ca9378e0439f836c6fae6ad92d1330dd5b162c0b5588f6ba02c3c5
3
+ size 9865240711
pytorch_model-00007-of-00013.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ff6d1eb30614e1dde07e2a3544886579f3a97f6377940c4571fe15ea9c1b61c3
3
+ size 9865240711
pytorch_model-00008-of-00013.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:22e473f6b51fa3d6f0a1b054524bfcc8c3af7330b243eea49e8ae4ad6d9d9ec1
3
+ size 9865240711
pytorch_model-00009-of-00013.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:38b48a463d44930a33c9b8aee50edfe740206b9faf01522946896ace01656ed2
3
+ size 9865240711
pytorch_model-00010-of-00013.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fda6ca9ac2b6708978283919f89715c8582155ea9cf260a818d85ba606ea390b
3
+ size 9865240711
pytorch_model-00011-of-00013.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2a63dded0627410cfd3d36721653caa68cd6be2a75ceb5e663c83b88284be771
3
+ size 9865240711
pytorch_model-00012-of-00013.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e6dc38e8961a1be23d1659f094618a23dc7cb2b0d2d43bdd1b03c81318ec22f7
3
+ size 9865240711
pytorch_model-00013-of-00013.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d5548c40c61094663c5bcf66e9e50969848aa36b7be5375eff8c0397d5dace75
3
+ size 1644197388
pytorch_model.bin.index.json ADDED
@@ -0,0 +1,298 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metadata": {
3
+ "total_size": 119828803584
4
+ },
5
+ "weight_map": {
6
+ "lm_head.weight": "pytorch_model-00001-of-00013.bin",
7
+ "transformer.blocks.0.attn.Wqkv.weight": "pytorch_model-00001-of-00013.bin",
8
+ "transformer.blocks.0.attn.out_proj.weight": "pytorch_model-00001-of-00013.bin",
9
+ "transformer.blocks.0.ffn.down_proj.weight": "pytorch_model-00001-of-00013.bin",
10
+ "transformer.blocks.0.ffn.up_proj.weight": "pytorch_model-00001-of-00013.bin",
11
+ "transformer.blocks.0.norm_1.weight": "pytorch_model-00001-of-00013.bin",
12
+ "transformer.blocks.0.norm_2.weight": "pytorch_model-00001-of-00013.bin",
13
+ "transformer.blocks.1.attn.Wqkv.weight": "pytorch_model-00001-of-00013.bin",
14
+ "transformer.blocks.1.attn.out_proj.weight": "pytorch_model-00001-of-00013.bin",
15
+ "transformer.blocks.1.ffn.down_proj.weight": "pytorch_model-00001-of-00013.bin",
16
+ "transformer.blocks.1.ffn.up_proj.weight": "pytorch_model-00001-of-00013.bin",
17
+ "transformer.blocks.1.norm_1.weight": "pytorch_model-00001-of-00013.bin",
18
+ "transformer.blocks.1.norm_2.weight": "pytorch_model-00001-of-00013.bin",
19
+ "transformer.blocks.10.attn.Wqkv.weight": "pytorch_model-00003-of-00013.bin",
20
+ "transformer.blocks.10.attn.out_proj.weight": "pytorch_model-00003-of-00013.bin",
21
+ "transformer.blocks.10.ffn.down_proj.weight": "pytorch_model-00003-of-00013.bin",
22
+ "transformer.blocks.10.ffn.up_proj.weight": "pytorch_model-00003-of-00013.bin",
23
+ "transformer.blocks.10.norm_1.weight": "pytorch_model-00003-of-00013.bin",
24
+ "transformer.blocks.10.norm_2.weight": "pytorch_model-00003-of-00013.bin",
25
+ "transformer.blocks.11.attn.Wqkv.weight": "pytorch_model-00003-of-00013.bin",
26
+ "transformer.blocks.11.attn.out_proj.weight": "pytorch_model-00003-of-00013.bin",
27
+ "transformer.blocks.11.ffn.down_proj.weight": "pytorch_model-00004-of-00013.bin",
28
+ "transformer.blocks.11.ffn.up_proj.weight": "pytorch_model-00004-of-00013.bin",
29
+ "transformer.blocks.11.norm_1.weight": "pytorch_model-00003-of-00013.bin",
30
+ "transformer.blocks.11.norm_2.weight": "pytorch_model-00003-of-00013.bin",
31
+ "transformer.blocks.12.attn.Wqkv.weight": "pytorch_model-00004-of-00013.bin",
32
+ "transformer.blocks.12.attn.out_proj.weight": "pytorch_model-00004-of-00013.bin",
33
+ "transformer.blocks.12.ffn.down_proj.weight": "pytorch_model-00004-of-00013.bin",
34
+ "transformer.blocks.12.ffn.up_proj.weight": "pytorch_model-00004-of-00013.bin",
35
+ "transformer.blocks.12.norm_1.weight": "pytorch_model-00004-of-00013.bin",
36
+ "transformer.blocks.12.norm_2.weight": "pytorch_model-00004-of-00013.bin",
37
+ "transformer.blocks.13.attn.Wqkv.weight": "pytorch_model-00004-of-00013.bin",
38
+ "transformer.blocks.13.attn.out_proj.weight": "pytorch_model-00004-of-00013.bin",
39
+ "transformer.blocks.13.ffn.down_proj.weight": "pytorch_model-00004-of-00013.bin",
40
+ "transformer.blocks.13.ffn.up_proj.weight": "pytorch_model-00004-of-00013.bin",
41
+ "transformer.blocks.13.norm_1.weight": "pytorch_model-00004-of-00013.bin",
42
+ "transformer.blocks.13.norm_2.weight": "pytorch_model-00004-of-00013.bin",
43
+ "transformer.blocks.14.attn.Wqkv.weight": "pytorch_model-00004-of-00013.bin",
44
+ "transformer.blocks.14.attn.out_proj.weight": "pytorch_model-00004-of-00013.bin",
45
+ "transformer.blocks.14.ffn.down_proj.weight": "pytorch_model-00004-of-00013.bin",
46
+ "transformer.blocks.14.ffn.up_proj.weight": "pytorch_model-00004-of-00013.bin",
47
+ "transformer.blocks.14.norm_1.weight": "pytorch_model-00004-of-00013.bin",
48
+ "transformer.blocks.14.norm_2.weight": "pytorch_model-00004-of-00013.bin",
49
+ "transformer.blocks.15.attn.Wqkv.weight": "pytorch_model-00004-of-00013.bin",
50
+ "transformer.blocks.15.attn.out_proj.weight": "pytorch_model-00004-of-00013.bin",
51
+ "transformer.blocks.15.ffn.down_proj.weight": "pytorch_model-00005-of-00013.bin",
52
+ "transformer.blocks.15.ffn.up_proj.weight": "pytorch_model-00005-of-00013.bin",
53
+ "transformer.blocks.15.norm_1.weight": "pytorch_model-00004-of-00013.bin",
54
+ "transformer.blocks.15.norm_2.weight": "pytorch_model-00004-of-00013.bin",
55
+ "transformer.blocks.16.attn.Wqkv.weight": "pytorch_model-00005-of-00013.bin",
56
+ "transformer.blocks.16.attn.out_proj.weight": "pytorch_model-00005-of-00013.bin",
57
+ "transformer.blocks.16.ffn.down_proj.weight": "pytorch_model-00005-of-00013.bin",
58
+ "transformer.blocks.16.ffn.up_proj.weight": "pytorch_model-00005-of-00013.bin",
59
+ "transformer.blocks.16.norm_1.weight": "pytorch_model-00005-of-00013.bin",
60
+ "transformer.blocks.16.norm_2.weight": "pytorch_model-00005-of-00013.bin",
61
+ "transformer.blocks.17.attn.Wqkv.weight": "pytorch_model-00005-of-00013.bin",
62
+ "transformer.blocks.17.attn.out_proj.weight": "pytorch_model-00005-of-00013.bin",
63
+ "transformer.blocks.17.ffn.down_proj.weight": "pytorch_model-00005-of-00013.bin",
64
+ "transformer.blocks.17.ffn.up_proj.weight": "pytorch_model-00005-of-00013.bin",
65
+ "transformer.blocks.17.norm_1.weight": "pytorch_model-00005-of-00013.bin",
66
+ "transformer.blocks.17.norm_2.weight": "pytorch_model-00005-of-00013.bin",
67
+ "transformer.blocks.18.attn.Wqkv.weight": "pytorch_model-00005-of-00013.bin",
68
+ "transformer.blocks.18.attn.out_proj.weight": "pytorch_model-00005-of-00013.bin",
69
+ "transformer.blocks.18.ffn.down_proj.weight": "pytorch_model-00005-of-00013.bin",
70
+ "transformer.blocks.18.ffn.up_proj.weight": "pytorch_model-00005-of-00013.bin",
71
+ "transformer.blocks.18.norm_1.weight": "pytorch_model-00005-of-00013.bin",
72
+ "transformer.blocks.18.norm_2.weight": "pytorch_model-00005-of-00013.bin",
73
+ "transformer.blocks.19.attn.Wqkv.weight": "pytorch_model-00005-of-00013.bin",
74
+ "transformer.blocks.19.attn.out_proj.weight": "pytorch_model-00005-of-00013.bin",
75
+ "transformer.blocks.19.ffn.down_proj.weight": "pytorch_model-00006-of-00013.bin",
76
+ "transformer.blocks.19.ffn.up_proj.weight": "pytorch_model-00006-of-00013.bin",
77
+ "transformer.blocks.19.norm_1.weight": "pytorch_model-00005-of-00013.bin",
78
+ "transformer.blocks.19.norm_2.weight": "pytorch_model-00005-of-00013.bin",
79
+ "transformer.blocks.2.attn.Wqkv.weight": "pytorch_model-00001-of-00013.bin",
80
+ "transformer.blocks.2.attn.out_proj.weight": "pytorch_model-00001-of-00013.bin",
81
+ "transformer.blocks.2.ffn.down_proj.weight": "pytorch_model-00001-of-00013.bin",
82
+ "transformer.blocks.2.ffn.up_proj.weight": "pytorch_model-00001-of-00013.bin",
83
+ "transformer.blocks.2.norm_1.weight": "pytorch_model-00001-of-00013.bin",
84
+ "transformer.blocks.2.norm_2.weight": "pytorch_model-00001-of-00013.bin",
85
+ "transformer.blocks.20.attn.Wqkv.weight": "pytorch_model-00006-of-00013.bin",
86
+ "transformer.blocks.20.attn.out_proj.weight": "pytorch_model-00006-of-00013.bin",
87
+ "transformer.blocks.20.ffn.down_proj.weight": "pytorch_model-00006-of-00013.bin",
88
+ "transformer.blocks.20.ffn.up_proj.weight": "pytorch_model-00006-of-00013.bin",
89
+ "transformer.blocks.20.norm_1.weight": "pytorch_model-00006-of-00013.bin",
90
+ "transformer.blocks.20.norm_2.weight": "pytorch_model-00006-of-00013.bin",
91
+ "transformer.blocks.21.attn.Wqkv.weight": "pytorch_model-00006-of-00013.bin",
92
+ "transformer.blocks.21.attn.out_proj.weight": "pytorch_model-00006-of-00013.bin",
93
+ "transformer.blocks.21.ffn.down_proj.weight": "pytorch_model-00006-of-00013.bin",
94
+ "transformer.blocks.21.ffn.up_proj.weight": "pytorch_model-00006-of-00013.bin",
95
+ "transformer.blocks.21.norm_1.weight": "pytorch_model-00006-of-00013.bin",
96
+ "transformer.blocks.21.norm_2.weight": "pytorch_model-00006-of-00013.bin",
97
+ "transformer.blocks.22.attn.Wqkv.weight": "pytorch_model-00006-of-00013.bin",
98
+ "transformer.blocks.22.attn.out_proj.weight": "pytorch_model-00006-of-00013.bin",
99
+ "transformer.blocks.22.ffn.down_proj.weight": "pytorch_model-00006-of-00013.bin",
100
+ "transformer.blocks.22.ffn.up_proj.weight": "pytorch_model-00006-of-00013.bin",
101
+ "transformer.blocks.22.norm_1.weight": "pytorch_model-00006-of-00013.bin",
102
+ "transformer.blocks.22.norm_2.weight": "pytorch_model-00006-of-00013.bin",
103
+ "transformer.blocks.23.attn.Wqkv.weight": "pytorch_model-00006-of-00013.bin",
104
+ "transformer.blocks.23.attn.out_proj.weight": "pytorch_model-00006-of-00013.bin",
105
+ "transformer.blocks.23.ffn.down_proj.weight": "pytorch_model-00007-of-00013.bin",
106
+ "transformer.blocks.23.ffn.up_proj.weight": "pytorch_model-00007-of-00013.bin",
107
+ "transformer.blocks.23.norm_1.weight": "pytorch_model-00006-of-00013.bin",
108
+ "transformer.blocks.23.norm_2.weight": "pytorch_model-00006-of-00013.bin",
109
+ "transformer.blocks.24.attn.Wqkv.weight": "pytorch_model-00007-of-00013.bin",
110
+ "transformer.blocks.24.attn.out_proj.weight": "pytorch_model-00007-of-00013.bin",
111
+ "transformer.blocks.24.ffn.down_proj.weight": "pytorch_model-00007-of-00013.bin",
112
+ "transformer.blocks.24.ffn.up_proj.weight": "pytorch_model-00007-of-00013.bin",
113
+ "transformer.blocks.24.norm_1.weight": "pytorch_model-00007-of-00013.bin",
114
+ "transformer.blocks.24.norm_2.weight": "pytorch_model-00007-of-00013.bin",
115
+ "transformer.blocks.25.attn.Wqkv.weight": "pytorch_model-00007-of-00013.bin",
116
+ "transformer.blocks.25.attn.out_proj.weight": "pytorch_model-00007-of-00013.bin",
117
+ "transformer.blocks.25.ffn.down_proj.weight": "pytorch_model-00007-of-00013.bin",
118
+ "transformer.blocks.25.ffn.up_proj.weight": "pytorch_model-00007-of-00013.bin",
119
+ "transformer.blocks.25.norm_1.weight": "pytorch_model-00007-of-00013.bin",
120
+ "transformer.blocks.25.norm_2.weight": "pytorch_model-00007-of-00013.bin",
121
+ "transformer.blocks.26.attn.Wqkv.weight": "pytorch_model-00007-of-00013.bin",
122
+ "transformer.blocks.26.attn.out_proj.weight": "pytorch_model-00007-of-00013.bin",
123
+ "transformer.blocks.26.ffn.down_proj.weight": "pytorch_model-00007-of-00013.bin",
124
+ "transformer.blocks.26.ffn.up_proj.weight": "pytorch_model-00007-of-00013.bin",
125
+ "transformer.blocks.26.norm_1.weight": "pytorch_model-00007-of-00013.bin",
126
+ "transformer.blocks.26.norm_2.weight": "pytorch_model-00007-of-00013.bin",
127
+ "transformer.blocks.27.attn.Wqkv.weight": "pytorch_model-00007-of-00013.bin",
128
+ "transformer.blocks.27.attn.out_proj.weight": "pytorch_model-00007-of-00013.bin",
129
+ "transformer.blocks.27.ffn.down_proj.weight": "pytorch_model-00008-of-00013.bin",
130
+ "transformer.blocks.27.ffn.up_proj.weight": "pytorch_model-00008-of-00013.bin",
131
+ "transformer.blocks.27.norm_1.weight": "pytorch_model-00007-of-00013.bin",
132
+ "transformer.blocks.27.norm_2.weight": "pytorch_model-00007-of-00013.bin",
133
+ "transformer.blocks.28.attn.Wqkv.weight": "pytorch_model-00008-of-00013.bin",
134
+ "transformer.blocks.28.attn.out_proj.weight": "pytorch_model-00008-of-00013.bin",
135
+ "transformer.blocks.28.ffn.down_proj.weight": "pytorch_model-00008-of-00013.bin",
136
+ "transformer.blocks.28.ffn.up_proj.weight": "pytorch_model-00008-of-00013.bin",
137
+ "transformer.blocks.28.norm_1.weight": "pytorch_model-00008-of-00013.bin",
138
+ "transformer.blocks.28.norm_2.weight": "pytorch_model-00008-of-00013.bin",
139
+ "transformer.blocks.29.attn.Wqkv.weight": "pytorch_model-00008-of-00013.bin",
140
+ "transformer.blocks.29.attn.out_proj.weight": "pytorch_model-00008-of-00013.bin",
141
+ "transformer.blocks.29.ffn.down_proj.weight": "pytorch_model-00008-of-00013.bin",
142
+ "transformer.blocks.29.ffn.up_proj.weight": "pytorch_model-00008-of-00013.bin",
143
+ "transformer.blocks.29.norm_1.weight": "pytorch_model-00008-of-00013.bin",
144
+ "transformer.blocks.29.norm_2.weight": "pytorch_model-00008-of-00013.bin",
145
+ "transformer.blocks.3.attn.Wqkv.weight": "pytorch_model-00001-of-00013.bin",
146
+ "transformer.blocks.3.attn.out_proj.weight": "pytorch_model-00001-of-00013.bin",
147
+ "transformer.blocks.3.ffn.down_proj.weight": "pytorch_model-00002-of-00013.bin",
148
+ "transformer.blocks.3.ffn.up_proj.weight": "pytorch_model-00002-of-00013.bin",
149
+ "transformer.blocks.3.norm_1.weight": "pytorch_model-00001-of-00013.bin",
150
+ "transformer.blocks.3.norm_2.weight": "pytorch_model-00001-of-00013.bin",
151
+ "transformer.blocks.30.attn.Wqkv.weight": "pytorch_model-00008-of-00013.bin",
152
+ "transformer.blocks.30.attn.out_proj.weight": "pytorch_model-00008-of-00013.bin",
153
+ "transformer.blocks.30.ffn.down_proj.weight": "pytorch_model-00008-of-00013.bin",
154
+ "transformer.blocks.30.ffn.up_proj.weight": "pytorch_model-00008-of-00013.bin",
155
+ "transformer.blocks.30.norm_1.weight": "pytorch_model-00008-of-00013.bin",
156
+ "transformer.blocks.30.norm_2.weight": "pytorch_model-00008-of-00013.bin",
157
+ "transformer.blocks.31.attn.Wqkv.weight": "pytorch_model-00008-of-00013.bin",
158
+ "transformer.blocks.31.attn.out_proj.weight": "pytorch_model-00008-of-00013.bin",
159
+ "transformer.blocks.31.ffn.down_proj.weight": "pytorch_model-00009-of-00013.bin",
160
+ "transformer.blocks.31.ffn.up_proj.weight": "pytorch_model-00009-of-00013.bin",
161
+ "transformer.blocks.31.norm_1.weight": "pytorch_model-00008-of-00013.bin",
162
+ "transformer.blocks.31.norm_2.weight": "pytorch_model-00008-of-00013.bin",
163
+ "transformer.blocks.32.attn.Wqkv.weight": "pytorch_model-00009-of-00013.bin",
164
+ "transformer.blocks.32.attn.out_proj.weight": "pytorch_model-00009-of-00013.bin",
165
+ "transformer.blocks.32.ffn.down_proj.weight": "pytorch_model-00009-of-00013.bin",
166
+ "transformer.blocks.32.ffn.up_proj.weight": "pytorch_model-00009-of-00013.bin",
167
+ "transformer.blocks.32.norm_1.weight": "pytorch_model-00009-of-00013.bin",
168
+ "transformer.blocks.32.norm_2.weight": "pytorch_model-00009-of-00013.bin",
169
+ "transformer.blocks.33.attn.Wqkv.weight": "pytorch_model-00009-of-00013.bin",
170
+ "transformer.blocks.33.attn.out_proj.weight": "pytorch_model-00009-of-00013.bin",
171
+ "transformer.blocks.33.ffn.down_proj.weight": "pytorch_model-00009-of-00013.bin",
172
+ "transformer.blocks.33.ffn.up_proj.weight": "pytorch_model-00009-of-00013.bin",
173
+ "transformer.blocks.33.norm_1.weight": "pytorch_model-00009-of-00013.bin",
174
+ "transformer.blocks.33.norm_2.weight": "pytorch_model-00009-of-00013.bin",
175
+ "transformer.blocks.34.attn.Wqkv.weight": "pytorch_model-00009-of-00013.bin",
176
+ "transformer.blocks.34.attn.out_proj.weight": "pytorch_model-00009-of-00013.bin",
177
+ "transformer.blocks.34.ffn.down_proj.weight": "pytorch_model-00009-of-00013.bin",
178
+ "transformer.blocks.34.ffn.up_proj.weight": "pytorch_model-00009-of-00013.bin",
179
+ "transformer.blocks.34.norm_1.weight": "pytorch_model-00009-of-00013.bin",
180
+ "transformer.blocks.34.norm_2.weight": "pytorch_model-00009-of-00013.bin",
181
+ "transformer.blocks.35.attn.Wqkv.weight": "pytorch_model-00009-of-00013.bin",
182
+ "transformer.blocks.35.attn.out_proj.weight": "pytorch_model-00009-of-00013.bin",
183
+ "transformer.blocks.35.ffn.down_proj.weight": "pytorch_model-00010-of-00013.bin",
184
+ "transformer.blocks.35.ffn.up_proj.weight": "pytorch_model-00010-of-00013.bin",
185
+ "transformer.blocks.35.norm_1.weight": "pytorch_model-00009-of-00013.bin",
186
+ "transformer.blocks.35.norm_2.weight": "pytorch_model-00009-of-00013.bin",
187
+ "transformer.blocks.36.attn.Wqkv.weight": "pytorch_model-00010-of-00013.bin",
188
+ "transformer.blocks.36.attn.out_proj.weight": "pytorch_model-00010-of-00013.bin",
189
+ "transformer.blocks.36.ffn.down_proj.weight": "pytorch_model-00010-of-00013.bin",
190
+ "transformer.blocks.36.ffn.up_proj.weight": "pytorch_model-00010-of-00013.bin",
191
+ "transformer.blocks.36.norm_1.weight": "pytorch_model-00010-of-00013.bin",
192
+ "transformer.blocks.36.norm_2.weight": "pytorch_model-00010-of-00013.bin",
193
+ "transformer.blocks.37.attn.Wqkv.weight": "pytorch_model-00010-of-00013.bin",
194
+ "transformer.blocks.37.attn.out_proj.weight": "pytorch_model-00010-of-00013.bin",
195
+ "transformer.blocks.37.ffn.down_proj.weight": "pytorch_model-00010-of-00013.bin",
196
+ "transformer.blocks.37.ffn.up_proj.weight": "pytorch_model-00010-of-00013.bin",
197
+ "transformer.blocks.37.norm_1.weight": "pytorch_model-00010-of-00013.bin",
198
+ "transformer.blocks.37.norm_2.weight": "pytorch_model-00010-of-00013.bin",
199
+ "transformer.blocks.38.attn.Wqkv.weight": "pytorch_model-00010-of-00013.bin",
200
+ "transformer.blocks.38.attn.out_proj.weight": "pytorch_model-00010-of-00013.bin",
201
+ "transformer.blocks.38.ffn.down_proj.weight": "pytorch_model-00010-of-00013.bin",
202
+ "transformer.blocks.38.ffn.up_proj.weight": "pytorch_model-00010-of-00013.bin",
203
+ "transformer.blocks.38.norm_1.weight": "pytorch_model-00010-of-00013.bin",
204
+ "transformer.blocks.38.norm_2.weight": "pytorch_model-00010-of-00013.bin",
205
+ "transformer.blocks.39.attn.Wqkv.weight": "pytorch_model-00010-of-00013.bin",
206
+ "transformer.blocks.39.attn.out_proj.weight": "pytorch_model-00010-of-00013.bin",
207
+ "transformer.blocks.39.ffn.down_proj.weight": "pytorch_model-00011-of-00013.bin",
208
+ "transformer.blocks.39.ffn.up_proj.weight": "pytorch_model-00011-of-00013.bin",
209
+ "transformer.blocks.39.norm_1.weight": "pytorch_model-00010-of-00013.bin",
210
+ "transformer.blocks.39.norm_2.weight": "pytorch_model-00010-of-00013.bin",
211
+ "transformer.blocks.4.attn.Wqkv.weight": "pytorch_model-00002-of-00013.bin",
212
+ "transformer.blocks.4.attn.out_proj.weight": "pytorch_model-00002-of-00013.bin",
213
+ "transformer.blocks.4.ffn.down_proj.weight": "pytorch_model-00002-of-00013.bin",
214
+ "transformer.blocks.4.ffn.up_proj.weight": "pytorch_model-00002-of-00013.bin",
215
+ "transformer.blocks.4.norm_1.weight": "pytorch_model-00002-of-00013.bin",
216
+ "transformer.blocks.4.norm_2.weight": "pytorch_model-00002-of-00013.bin",
217
+ "transformer.blocks.40.attn.Wqkv.weight": "pytorch_model-00011-of-00013.bin",
218
+ "transformer.blocks.40.attn.out_proj.weight": "pytorch_model-00011-of-00013.bin",
219
+ "transformer.blocks.40.ffn.down_proj.weight": "pytorch_model-00011-of-00013.bin",
220
+ "transformer.blocks.40.ffn.up_proj.weight": "pytorch_model-00011-of-00013.bin",
221
+ "transformer.blocks.40.norm_1.weight": "pytorch_model-00011-of-00013.bin",
222
+ "transformer.blocks.40.norm_2.weight": "pytorch_model-00011-of-00013.bin",
223
+ "transformer.blocks.41.attn.Wqkv.weight": "pytorch_model-00011-of-00013.bin",
224
+ "transformer.blocks.41.attn.out_proj.weight": "pytorch_model-00011-of-00013.bin",
225
+ "transformer.blocks.41.ffn.down_proj.weight": "pytorch_model-00011-of-00013.bin",
226
+ "transformer.blocks.41.ffn.up_proj.weight": "pytorch_model-00011-of-00013.bin",
227
+ "transformer.blocks.41.norm_1.weight": "pytorch_model-00011-of-00013.bin",
228
+ "transformer.blocks.41.norm_2.weight": "pytorch_model-00011-of-00013.bin",
229
+ "transformer.blocks.42.attn.Wqkv.weight": "pytorch_model-00011-of-00013.bin",
230
+ "transformer.blocks.42.attn.out_proj.weight": "pytorch_model-00011-of-00013.bin",
231
+ "transformer.blocks.42.ffn.down_proj.weight": "pytorch_model-00011-of-00013.bin",
232
+ "transformer.blocks.42.ffn.up_proj.weight": "pytorch_model-00011-of-00013.bin",
233
+ "transformer.blocks.42.norm_1.weight": "pytorch_model-00011-of-00013.bin",
234
+ "transformer.blocks.42.norm_2.weight": "pytorch_model-00011-of-00013.bin",
235
+ "transformer.blocks.43.attn.Wqkv.weight": "pytorch_model-00011-of-00013.bin",
236
+ "transformer.blocks.43.attn.out_proj.weight": "pytorch_model-00011-of-00013.bin",
237
+ "transformer.blocks.43.ffn.down_proj.weight": "pytorch_model-00012-of-00013.bin",
238
+ "transformer.blocks.43.ffn.up_proj.weight": "pytorch_model-00012-of-00013.bin",
239
+ "transformer.blocks.43.norm_1.weight": "pytorch_model-00011-of-00013.bin",
240
+ "transformer.blocks.43.norm_2.weight": "pytorch_model-00011-of-00013.bin",
241
+ "transformer.blocks.44.attn.Wqkv.weight": "pytorch_model-00012-of-00013.bin",
242
+ "transformer.blocks.44.attn.out_proj.weight": "pytorch_model-00012-of-00013.bin",
243
+ "transformer.blocks.44.ffn.down_proj.weight": "pytorch_model-00012-of-00013.bin",
244
+ "transformer.blocks.44.ffn.up_proj.weight": "pytorch_model-00012-of-00013.bin",
245
+ "transformer.blocks.44.norm_1.weight": "pytorch_model-00012-of-00013.bin",
246
+ "transformer.blocks.44.norm_2.weight": "pytorch_model-00012-of-00013.bin",
247
+ "transformer.blocks.45.attn.Wqkv.weight": "pytorch_model-00012-of-00013.bin",
248
+ "transformer.blocks.45.attn.out_proj.weight": "pytorch_model-00012-of-00013.bin",
249
+ "transformer.blocks.45.ffn.down_proj.weight": "pytorch_model-00012-of-00013.bin",
250
+ "transformer.blocks.45.ffn.up_proj.weight": "pytorch_model-00012-of-00013.bin",
251
+ "transformer.blocks.45.norm_1.weight": "pytorch_model-00012-of-00013.bin",
252
+ "transformer.blocks.45.norm_2.weight": "pytorch_model-00012-of-00013.bin",
253
+ "transformer.blocks.46.attn.Wqkv.weight": "pytorch_model-00012-of-00013.bin",
254
+ "transformer.blocks.46.attn.out_proj.weight": "pytorch_model-00012-of-00013.bin",
255
+ "transformer.blocks.46.ffn.down_proj.weight": "pytorch_model-00012-of-00013.bin",
256
+ "transformer.blocks.46.ffn.up_proj.weight": "pytorch_model-00012-of-00013.bin",
257
+ "transformer.blocks.46.norm_1.weight": "pytorch_model-00012-of-00013.bin",
258
+ "transformer.blocks.46.norm_2.weight": "pytorch_model-00012-of-00013.bin",
259
+ "transformer.blocks.47.attn.Wqkv.weight": "pytorch_model-00012-of-00013.bin",
260
+ "transformer.blocks.47.attn.out_proj.weight": "pytorch_model-00012-of-00013.bin",
261
+ "transformer.blocks.47.ffn.down_proj.weight": "pytorch_model-00013-of-00013.bin",
262
+ "transformer.blocks.47.ffn.up_proj.weight": "pytorch_model-00013-of-00013.bin",
263
+ "transformer.blocks.47.norm_1.weight": "pytorch_model-00012-of-00013.bin",
264
+ "transformer.blocks.47.norm_2.weight": "pytorch_model-00012-of-00013.bin",
265
+ "transformer.blocks.5.attn.Wqkv.weight": "pytorch_model-00002-of-00013.bin",
266
+ "transformer.blocks.5.attn.out_proj.weight": "pytorch_model-00002-of-00013.bin",
267
+ "transformer.blocks.5.ffn.down_proj.weight": "pytorch_model-00002-of-00013.bin",
268
+ "transformer.blocks.5.ffn.up_proj.weight": "pytorch_model-00002-of-00013.bin",
269
+ "transformer.blocks.5.norm_1.weight": "pytorch_model-00002-of-00013.bin",
270
+ "transformer.blocks.5.norm_2.weight": "pytorch_model-00002-of-00013.bin",
271
+ "transformer.blocks.6.attn.Wqkv.weight": "pytorch_model-00002-of-00013.bin",
272
+ "transformer.blocks.6.attn.out_proj.weight": "pytorch_model-00002-of-00013.bin",
273
+ "transformer.blocks.6.ffn.down_proj.weight": "pytorch_model-00002-of-00013.bin",
274
+ "transformer.blocks.6.ffn.up_proj.weight": "pytorch_model-00002-of-00013.bin",
275
+ "transformer.blocks.6.norm_1.weight": "pytorch_model-00002-of-00013.bin",
276
+ "transformer.blocks.6.norm_2.weight": "pytorch_model-00002-of-00013.bin",
277
+ "transformer.blocks.7.attn.Wqkv.weight": "pytorch_model-00002-of-00013.bin",
278
+ "transformer.blocks.7.attn.out_proj.weight": "pytorch_model-00002-of-00013.bin",
279
+ "transformer.blocks.7.ffn.down_proj.weight": "pytorch_model-00003-of-00013.bin",
280
+ "transformer.blocks.7.ffn.up_proj.weight": "pytorch_model-00003-of-00013.bin",
281
+ "transformer.blocks.7.norm_1.weight": "pytorch_model-00002-of-00013.bin",
282
+ "transformer.blocks.7.norm_2.weight": "pytorch_model-00002-of-00013.bin",
283
+ "transformer.blocks.8.attn.Wqkv.weight": "pytorch_model-00003-of-00013.bin",
284
+ "transformer.blocks.8.attn.out_proj.weight": "pytorch_model-00003-of-00013.bin",
285
+ "transformer.blocks.8.ffn.down_proj.weight": "pytorch_model-00003-of-00013.bin",
286
+ "transformer.blocks.8.ffn.up_proj.weight": "pytorch_model-00003-of-00013.bin",
287
+ "transformer.blocks.8.norm_1.weight": "pytorch_model-00003-of-00013.bin",
288
+ "transformer.blocks.8.norm_2.weight": "pytorch_model-00003-of-00013.bin",
289
+ "transformer.blocks.9.attn.Wqkv.weight": "pytorch_model-00003-of-00013.bin",
290
+ "transformer.blocks.9.attn.out_proj.weight": "pytorch_model-00003-of-00013.bin",
291
+ "transformer.blocks.9.ffn.down_proj.weight": "pytorch_model-00003-of-00013.bin",
292
+ "transformer.blocks.9.ffn.up_proj.weight": "pytorch_model-00003-of-00013.bin",
293
+ "transformer.blocks.9.norm_1.weight": "pytorch_model-00003-of-00013.bin",
294
+ "transformer.blocks.9.norm_2.weight": "pytorch_model-00003-of-00013.bin",
295
+ "transformer.norm_f.weight": "pytorch_model-00013-of-00013.bin",
296
+ "transformer.wte.weight": "pytorch_model-00001-of-00013.bin"
297
+ }
298
+ }