debisoft commited on
Commit
a7da5dc
1 Parent(s): 6887de8

Upload MptForCausalLM

Browse files
config.json ADDED
@@ -0,0 +1,46 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "debisoft/mpt-7b-8k-peft-compatible",
3
+ "architectures": [
4
+ "MptForCausalLM"
5
+ ],
6
+ "attn_config": {
7
+ "model_type": ""
8
+ },
9
+ "auto_map": {
10
+ "AutoConfig": "debisoft/mpt-7b-8k-peft-compatible--configuration_mpt.MPTConfig",
11
+ "AutoModelForCausalLM": "debisoft/mpt-7b-8k-peft-compatible--modeling_mpt.MPTForCausalLM"
12
+ },
13
+ "d_model": 4096,
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
+ "model_type": "mpt",
35
+ "n_heads": 32,
36
+ "n_layers": 32,
37
+ "no_bias": true,
38
+ "norm_type": "low_precision_layernorm",
39
+ "resid_pdrop": 0,
40
+ "tokenizer_name": "EleutherAI/gpt-neox-20b",
41
+ "torch_dtype": "bfloat16",
42
+ "transformers_version": "4.35.2",
43
+ "use_cache": false,
44
+ "verbose": 0,
45
+ "vocab_size": 50432
46
+ }
generation_config.json ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "transformers_version": "4.35.2",
4
+ "use_cache": false
5
+ }
model-00001-of-00003.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:32c666b9c014641768dd9c9cb63066bbc4bd373d928f6aa22b9bd81502e2e1cc
3
+ size 4976746496
model-00002-of-00003.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6a88ca5bca34b267a70a6d94ad13cf25f4d1cef22a19168d696dabea422a6264
3
+ size 4966261064
model-00003-of-00003.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6da16131aca4765145c9045701bf6404c4d582bdcd8b7a641dbeb08c548ce583
3
+ size 3355588288
model.safetensors.index.json ADDED
@@ -0,0 +1,201 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metadata": {
3
+ "total_size": 13298573312
4
+ },
5
+ "weight_map": {
6
+ "transformer.blocks.0.attn.Wqkv.weight": "model-00001-of-00003.safetensors",
7
+ "transformer.blocks.0.attn.out_proj.weight": "model-00001-of-00003.safetensors",
8
+ "transformer.blocks.0.ffn.down_proj.weight": "model-00001-of-00003.safetensors",
9
+ "transformer.blocks.0.ffn.up_proj.weight": "model-00001-of-00003.safetensors",
10
+ "transformer.blocks.0.norm_1.weight": "model-00001-of-00003.safetensors",
11
+ "transformer.blocks.0.norm_2.weight": "model-00001-of-00003.safetensors",
12
+ "transformer.blocks.1.attn.Wqkv.weight": "model-00001-of-00003.safetensors",
13
+ "transformer.blocks.1.attn.out_proj.weight": "model-00001-of-00003.safetensors",
14
+ "transformer.blocks.1.ffn.down_proj.weight": "model-00001-of-00003.safetensors",
15
+ "transformer.blocks.1.ffn.up_proj.weight": "model-00001-of-00003.safetensors",
16
+ "transformer.blocks.1.norm_1.weight": "model-00001-of-00003.safetensors",
17
+ "transformer.blocks.1.norm_2.weight": "model-00001-of-00003.safetensors",
18
+ "transformer.blocks.10.attn.Wqkv.weight": "model-00001-of-00003.safetensors",
19
+ "transformer.blocks.10.attn.out_proj.weight": "model-00001-of-00003.safetensors",
20
+ "transformer.blocks.10.ffn.down_proj.weight": "model-00001-of-00003.safetensors",
21
+ "transformer.blocks.10.ffn.up_proj.weight": "model-00001-of-00003.safetensors",
22
+ "transformer.blocks.10.norm_1.weight": "model-00001-of-00003.safetensors",
23
+ "transformer.blocks.10.norm_2.weight": "model-00001-of-00003.safetensors",
24
+ "transformer.blocks.11.attn.Wqkv.weight": "model-00001-of-00003.safetensors",
25
+ "transformer.blocks.11.attn.out_proj.weight": "model-00001-of-00003.safetensors",
26
+ "transformer.blocks.11.ffn.down_proj.weight": "model-00002-of-00003.safetensors",
27
+ "transformer.blocks.11.ffn.up_proj.weight": "model-00002-of-00003.safetensors",
28
+ "transformer.blocks.11.norm_1.weight": "model-00001-of-00003.safetensors",
29
+ "transformer.blocks.11.norm_2.weight": "model-00001-of-00003.safetensors",
30
+ "transformer.blocks.12.attn.Wqkv.weight": "model-00002-of-00003.safetensors",
31
+ "transformer.blocks.12.attn.out_proj.weight": "model-00002-of-00003.safetensors",
32
+ "transformer.blocks.12.ffn.down_proj.weight": "model-00002-of-00003.safetensors",
33
+ "transformer.blocks.12.ffn.up_proj.weight": "model-00002-of-00003.safetensors",
34
+ "transformer.blocks.12.norm_1.weight": "model-00002-of-00003.safetensors",
35
+ "transformer.blocks.12.norm_2.weight": "model-00002-of-00003.safetensors",
36
+ "transformer.blocks.13.attn.Wqkv.weight": "model-00002-of-00003.safetensors",
37
+ "transformer.blocks.13.attn.out_proj.weight": "model-00002-of-00003.safetensors",
38
+ "transformer.blocks.13.ffn.down_proj.weight": "model-00002-of-00003.safetensors",
39
+ "transformer.blocks.13.ffn.up_proj.weight": "model-00002-of-00003.safetensors",
40
+ "transformer.blocks.13.norm_1.weight": "model-00002-of-00003.safetensors",
41
+ "transformer.blocks.13.norm_2.weight": "model-00002-of-00003.safetensors",
42
+ "transformer.blocks.14.attn.Wqkv.weight": "model-00002-of-00003.safetensors",
43
+ "transformer.blocks.14.attn.out_proj.weight": "model-00002-of-00003.safetensors",
44
+ "transformer.blocks.14.ffn.down_proj.weight": "model-00002-of-00003.safetensors",
45
+ "transformer.blocks.14.ffn.up_proj.weight": "model-00002-of-00003.safetensors",
46
+ "transformer.blocks.14.norm_1.weight": "model-00002-of-00003.safetensors",
47
+ "transformer.blocks.14.norm_2.weight": "model-00002-of-00003.safetensors",
48
+ "transformer.blocks.15.attn.Wqkv.weight": "model-00002-of-00003.safetensors",
49
+ "transformer.blocks.15.attn.out_proj.weight": "model-00002-of-00003.safetensors",
50
+ "transformer.blocks.15.ffn.down_proj.weight": "model-00002-of-00003.safetensors",
51
+ "transformer.blocks.15.ffn.up_proj.weight": "model-00002-of-00003.safetensors",
52
+ "transformer.blocks.15.norm_1.weight": "model-00002-of-00003.safetensors",
53
+ "transformer.blocks.15.norm_2.weight": "model-00002-of-00003.safetensors",
54
+ "transformer.blocks.16.attn.Wqkv.weight": "model-00002-of-00003.safetensors",
55
+ "transformer.blocks.16.attn.out_proj.weight": "model-00002-of-00003.safetensors",
56
+ "transformer.blocks.16.ffn.down_proj.weight": "model-00002-of-00003.safetensors",
57
+ "transformer.blocks.16.ffn.up_proj.weight": "model-00002-of-00003.safetensors",
58
+ "transformer.blocks.16.norm_1.weight": "model-00002-of-00003.safetensors",
59
+ "transformer.blocks.16.norm_2.weight": "model-00002-of-00003.safetensors",
60
+ "transformer.blocks.17.attn.Wqkv.weight": "model-00002-of-00003.safetensors",
61
+ "transformer.blocks.17.attn.out_proj.weight": "model-00002-of-00003.safetensors",
62
+ "transformer.blocks.17.ffn.down_proj.weight": "model-00002-of-00003.safetensors",
63
+ "transformer.blocks.17.ffn.up_proj.weight": "model-00002-of-00003.safetensors",
64
+ "transformer.blocks.17.norm_1.weight": "model-00002-of-00003.safetensors",
65
+ "transformer.blocks.17.norm_2.weight": "model-00002-of-00003.safetensors",
66
+ "transformer.blocks.18.attn.Wqkv.weight": "model-00002-of-00003.safetensors",
67
+ "transformer.blocks.18.attn.out_proj.weight": "model-00002-of-00003.safetensors",
68
+ "transformer.blocks.18.ffn.down_proj.weight": "model-00002-of-00003.safetensors",
69
+ "transformer.blocks.18.ffn.up_proj.weight": "model-00002-of-00003.safetensors",
70
+ "transformer.blocks.18.norm_1.weight": "model-00002-of-00003.safetensors",
71
+ "transformer.blocks.18.norm_2.weight": "model-00002-of-00003.safetensors",
72
+ "transformer.blocks.19.attn.Wqkv.weight": "model-00002-of-00003.safetensors",
73
+ "transformer.blocks.19.attn.out_proj.weight": "model-00002-of-00003.safetensors",
74
+ "transformer.blocks.19.ffn.down_proj.weight": "model-00002-of-00003.safetensors",
75
+ "transformer.blocks.19.ffn.up_proj.weight": "model-00002-of-00003.safetensors",
76
+ "transformer.blocks.19.norm_1.weight": "model-00002-of-00003.safetensors",
77
+ "transformer.blocks.19.norm_2.weight": "model-00002-of-00003.safetensors",
78
+ "transformer.blocks.2.attn.Wqkv.weight": "model-00001-of-00003.safetensors",
79
+ "transformer.blocks.2.attn.out_proj.weight": "model-00001-of-00003.safetensors",
80
+ "transformer.blocks.2.ffn.down_proj.weight": "model-00001-of-00003.safetensors",
81
+ "transformer.blocks.2.ffn.up_proj.weight": "model-00001-of-00003.safetensors",
82
+ "transformer.blocks.2.norm_1.weight": "model-00001-of-00003.safetensors",
83
+ "transformer.blocks.2.norm_2.weight": "model-00001-of-00003.safetensors",
84
+ "transformer.blocks.20.attn.Wqkv.weight": "model-00002-of-00003.safetensors",
85
+ "transformer.blocks.20.attn.out_proj.weight": "model-00002-of-00003.safetensors",
86
+ "transformer.blocks.20.ffn.down_proj.weight": "model-00002-of-00003.safetensors",
87
+ "transformer.blocks.20.ffn.up_proj.weight": "model-00002-of-00003.safetensors",
88
+ "transformer.blocks.20.norm_1.weight": "model-00002-of-00003.safetensors",
89
+ "transformer.blocks.20.norm_2.weight": "model-00002-of-00003.safetensors",
90
+ "transformer.blocks.21.attn.Wqkv.weight": "model-00002-of-00003.safetensors",
91
+ "transformer.blocks.21.attn.out_proj.weight": "model-00002-of-00003.safetensors",
92
+ "transformer.blocks.21.ffn.down_proj.weight": "model-00002-of-00003.safetensors",
93
+ "transformer.blocks.21.ffn.up_proj.weight": "model-00002-of-00003.safetensors",
94
+ "transformer.blocks.21.norm_1.weight": "model-00002-of-00003.safetensors",
95
+ "transformer.blocks.21.norm_2.weight": "model-00002-of-00003.safetensors",
96
+ "transformer.blocks.22.attn.Wqkv.weight": "model-00002-of-00003.safetensors",
97
+ "transformer.blocks.22.attn.out_proj.weight": "model-00002-of-00003.safetensors",
98
+ "transformer.blocks.22.ffn.down_proj.weight": "model-00002-of-00003.safetensors",
99
+ "transformer.blocks.22.ffn.up_proj.weight": "model-00002-of-00003.safetensors",
100
+ "transformer.blocks.22.norm_1.weight": "model-00002-of-00003.safetensors",
101
+ "transformer.blocks.22.norm_2.weight": "model-00002-of-00003.safetensors",
102
+ "transformer.blocks.23.attn.Wqkv.weight": "model-00002-of-00003.safetensors",
103
+ "transformer.blocks.23.attn.out_proj.weight": "model-00002-of-00003.safetensors",
104
+ "transformer.blocks.23.ffn.down_proj.weight": "model-00003-of-00003.safetensors",
105
+ "transformer.blocks.23.ffn.up_proj.weight": "model-00002-of-00003.safetensors",
106
+ "transformer.blocks.23.norm_1.weight": "model-00002-of-00003.safetensors",
107
+ "transformer.blocks.23.norm_2.weight": "model-00002-of-00003.safetensors",
108
+ "transformer.blocks.24.attn.Wqkv.weight": "model-00003-of-00003.safetensors",
109
+ "transformer.blocks.24.attn.out_proj.weight": "model-00003-of-00003.safetensors",
110
+ "transformer.blocks.24.ffn.down_proj.weight": "model-00003-of-00003.safetensors",
111
+ "transformer.blocks.24.ffn.up_proj.weight": "model-00003-of-00003.safetensors",
112
+ "transformer.blocks.24.norm_1.weight": "model-00003-of-00003.safetensors",
113
+ "transformer.blocks.24.norm_2.weight": "model-00003-of-00003.safetensors",
114
+ "transformer.blocks.25.attn.Wqkv.weight": "model-00003-of-00003.safetensors",
115
+ "transformer.blocks.25.attn.out_proj.weight": "model-00003-of-00003.safetensors",
116
+ "transformer.blocks.25.ffn.down_proj.weight": "model-00003-of-00003.safetensors",
117
+ "transformer.blocks.25.ffn.up_proj.weight": "model-00003-of-00003.safetensors",
118
+ "transformer.blocks.25.norm_1.weight": "model-00003-of-00003.safetensors",
119
+ "transformer.blocks.25.norm_2.weight": "model-00003-of-00003.safetensors",
120
+ "transformer.blocks.26.attn.Wqkv.weight": "model-00003-of-00003.safetensors",
121
+ "transformer.blocks.26.attn.out_proj.weight": "model-00003-of-00003.safetensors",
122
+ "transformer.blocks.26.ffn.down_proj.weight": "model-00003-of-00003.safetensors",
123
+ "transformer.blocks.26.ffn.up_proj.weight": "model-00003-of-00003.safetensors",
124
+ "transformer.blocks.26.norm_1.weight": "model-00003-of-00003.safetensors",
125
+ "transformer.blocks.26.norm_2.weight": "model-00003-of-00003.safetensors",
126
+ "transformer.blocks.27.attn.Wqkv.weight": "model-00003-of-00003.safetensors",
127
+ "transformer.blocks.27.attn.out_proj.weight": "model-00003-of-00003.safetensors",
128
+ "transformer.blocks.27.ffn.down_proj.weight": "model-00003-of-00003.safetensors",
129
+ "transformer.blocks.27.ffn.up_proj.weight": "model-00003-of-00003.safetensors",
130
+ "transformer.blocks.27.norm_1.weight": "model-00003-of-00003.safetensors",
131
+ "transformer.blocks.27.norm_2.weight": "model-00003-of-00003.safetensors",
132
+ "transformer.blocks.28.attn.Wqkv.weight": "model-00003-of-00003.safetensors",
133
+ "transformer.blocks.28.attn.out_proj.weight": "model-00003-of-00003.safetensors",
134
+ "transformer.blocks.28.ffn.down_proj.weight": "model-00003-of-00003.safetensors",
135
+ "transformer.blocks.28.ffn.up_proj.weight": "model-00003-of-00003.safetensors",
136
+ "transformer.blocks.28.norm_1.weight": "model-00003-of-00003.safetensors",
137
+ "transformer.blocks.28.norm_2.weight": "model-00003-of-00003.safetensors",
138
+ "transformer.blocks.29.attn.Wqkv.weight": "model-00003-of-00003.safetensors",
139
+ "transformer.blocks.29.attn.out_proj.weight": "model-00003-of-00003.safetensors",
140
+ "transformer.blocks.29.ffn.down_proj.weight": "model-00003-of-00003.safetensors",
141
+ "transformer.blocks.29.ffn.up_proj.weight": "model-00003-of-00003.safetensors",
142
+ "transformer.blocks.29.norm_1.weight": "model-00003-of-00003.safetensors",
143
+ "transformer.blocks.29.norm_2.weight": "model-00003-of-00003.safetensors",
144
+ "transformer.blocks.3.attn.Wqkv.weight": "model-00001-of-00003.safetensors",
145
+ "transformer.blocks.3.attn.out_proj.weight": "model-00001-of-00003.safetensors",
146
+ "transformer.blocks.3.ffn.down_proj.weight": "model-00001-of-00003.safetensors",
147
+ "transformer.blocks.3.ffn.up_proj.weight": "model-00001-of-00003.safetensors",
148
+ "transformer.blocks.3.norm_1.weight": "model-00001-of-00003.safetensors",
149
+ "transformer.blocks.3.norm_2.weight": "model-00001-of-00003.safetensors",
150
+ "transformer.blocks.30.attn.Wqkv.weight": "model-00003-of-00003.safetensors",
151
+ "transformer.blocks.30.attn.out_proj.weight": "model-00003-of-00003.safetensors",
152
+ "transformer.blocks.30.ffn.down_proj.weight": "model-00003-of-00003.safetensors",
153
+ "transformer.blocks.30.ffn.up_proj.weight": "model-00003-of-00003.safetensors",
154
+ "transformer.blocks.30.norm_1.weight": "model-00003-of-00003.safetensors",
155
+ "transformer.blocks.30.norm_2.weight": "model-00003-of-00003.safetensors",
156
+ "transformer.blocks.31.attn.Wqkv.weight": "model-00003-of-00003.safetensors",
157
+ "transformer.blocks.31.attn.out_proj.weight": "model-00003-of-00003.safetensors",
158
+ "transformer.blocks.31.ffn.down_proj.weight": "model-00003-of-00003.safetensors",
159
+ "transformer.blocks.31.ffn.up_proj.weight": "model-00003-of-00003.safetensors",
160
+ "transformer.blocks.31.norm_1.weight": "model-00003-of-00003.safetensors",
161
+ "transformer.blocks.31.norm_2.weight": "model-00003-of-00003.safetensors",
162
+ "transformer.blocks.4.attn.Wqkv.weight": "model-00001-of-00003.safetensors",
163
+ "transformer.blocks.4.attn.out_proj.weight": "model-00001-of-00003.safetensors",
164
+ "transformer.blocks.4.ffn.down_proj.weight": "model-00001-of-00003.safetensors",
165
+ "transformer.blocks.4.ffn.up_proj.weight": "model-00001-of-00003.safetensors",
166
+ "transformer.blocks.4.norm_1.weight": "model-00001-of-00003.safetensors",
167
+ "transformer.blocks.4.norm_2.weight": "model-00001-of-00003.safetensors",
168
+ "transformer.blocks.5.attn.Wqkv.weight": "model-00001-of-00003.safetensors",
169
+ "transformer.blocks.5.attn.out_proj.weight": "model-00001-of-00003.safetensors",
170
+ "transformer.blocks.5.ffn.down_proj.weight": "model-00001-of-00003.safetensors",
171
+ "transformer.blocks.5.ffn.up_proj.weight": "model-00001-of-00003.safetensors",
172
+ "transformer.blocks.5.norm_1.weight": "model-00001-of-00003.safetensors",
173
+ "transformer.blocks.5.norm_2.weight": "model-00001-of-00003.safetensors",
174
+ "transformer.blocks.6.attn.Wqkv.weight": "model-00001-of-00003.safetensors",
175
+ "transformer.blocks.6.attn.out_proj.weight": "model-00001-of-00003.safetensors",
176
+ "transformer.blocks.6.ffn.down_proj.weight": "model-00001-of-00003.safetensors",
177
+ "transformer.blocks.6.ffn.up_proj.weight": "model-00001-of-00003.safetensors",
178
+ "transformer.blocks.6.norm_1.weight": "model-00001-of-00003.safetensors",
179
+ "transformer.blocks.6.norm_2.weight": "model-00001-of-00003.safetensors",
180
+ "transformer.blocks.7.attn.Wqkv.weight": "model-00001-of-00003.safetensors",
181
+ "transformer.blocks.7.attn.out_proj.weight": "model-00001-of-00003.safetensors",
182
+ "transformer.blocks.7.ffn.down_proj.weight": "model-00001-of-00003.safetensors",
183
+ "transformer.blocks.7.ffn.up_proj.weight": "model-00001-of-00003.safetensors",
184
+ "transformer.blocks.7.norm_1.weight": "model-00001-of-00003.safetensors",
185
+ "transformer.blocks.7.norm_2.weight": "model-00001-of-00003.safetensors",
186
+ "transformer.blocks.8.attn.Wqkv.weight": "model-00001-of-00003.safetensors",
187
+ "transformer.blocks.8.attn.out_proj.weight": "model-00001-of-00003.safetensors",
188
+ "transformer.blocks.8.ffn.down_proj.weight": "model-00001-of-00003.safetensors",
189
+ "transformer.blocks.8.ffn.up_proj.weight": "model-00001-of-00003.safetensors",
190
+ "transformer.blocks.8.norm_1.weight": "model-00001-of-00003.safetensors",
191
+ "transformer.blocks.8.norm_2.weight": "model-00001-of-00003.safetensors",
192
+ "transformer.blocks.9.attn.Wqkv.weight": "model-00001-of-00003.safetensors",
193
+ "transformer.blocks.9.attn.out_proj.weight": "model-00001-of-00003.safetensors",
194
+ "transformer.blocks.9.ffn.down_proj.weight": "model-00001-of-00003.safetensors",
195
+ "transformer.blocks.9.ffn.up_proj.weight": "model-00001-of-00003.safetensors",
196
+ "transformer.blocks.9.norm_1.weight": "model-00001-of-00003.safetensors",
197
+ "transformer.blocks.9.norm_2.weight": "model-00001-of-00003.safetensors",
198
+ "transformer.norm_f.weight": "model-00003-of-00003.safetensors",
199
+ "transformer.wte.weight": "model-00001-of-00003.safetensors"
200
+ }
201
+ }