ybelkada commited on
Commit
3bb5c99
1 Parent(s): 5b987fd

Upload MptForCausalLM

Browse files
config.json ADDED
@@ -0,0 +1,118 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_commit_hash": "72e5f594ce36f9cabfa2a9fd8f58b491eb467ee7",
3
+ "_name_or_path": "mosaicml/mpt-7b",
4
+ "architectures": [
5
+ "MptForCausalLM"
6
+ ],
7
+ "attn_config": {
8
+ "_name_or_path": "",
9
+ "add_cross_attention": false,
10
+ "alibi": true,
11
+ "alibi_bias_max": 8,
12
+ "architectures": null,
13
+ "attn_impl": "torch",
14
+ "attn_pdrop": 0,
15
+ "attn_type": "multihead_attention",
16
+ "attn_uses_sequence_id": false,
17
+ "bad_words_ids": null,
18
+ "begin_suppress_tokens": null,
19
+ "bos_token_id": null,
20
+ "chunk_size_feed_forward": 0,
21
+ "clip_qkv": null,
22
+ "cross_attention_hidden_size": null,
23
+ "decoder_start_token_id": null,
24
+ "diversity_penalty": 0.0,
25
+ "do_sample": false,
26
+ "early_stopping": false,
27
+ "encoder_no_repeat_ngram_size": 0,
28
+ "eos_token_id": null,
29
+ "exponential_decay_length_penalty": null,
30
+ "finetuning_task": null,
31
+ "forced_bos_token_id": null,
32
+ "forced_eos_token_id": null,
33
+ "id2label": {
34
+ "0": "LABEL_0",
35
+ "1": "LABEL_1"
36
+ },
37
+ "is_decoder": false,
38
+ "is_encoder_decoder": false,
39
+ "label2id": {
40
+ "LABEL_0": 0,
41
+ "LABEL_1": 1
42
+ },
43
+ "length_penalty": 1.0,
44
+ "max_length": 20,
45
+ "min_length": 0,
46
+ "model_type": "",
47
+ "no_repeat_ngram_size": 0,
48
+ "num_beam_groups": 1,
49
+ "num_beams": 1,
50
+ "num_return_sequences": 1,
51
+ "output_attentions": false,
52
+ "output_hidden_states": false,
53
+ "output_scores": false,
54
+ "pad_token_id": null,
55
+ "prefix": null,
56
+ "prefix_lm": false,
57
+ "problem_type": null,
58
+ "pruned_heads": {},
59
+ "qk_ln": false,
60
+ "remove_invalid_values": false,
61
+ "repetition_penalty": 1.0,
62
+ "return_dict": true,
63
+ "return_dict_in_generate": false,
64
+ "sep_token_id": null,
65
+ "softmax_scale": null,
66
+ "suppress_tokens": null,
67
+ "task_specific_params": null,
68
+ "temperature": 1.0,
69
+ "tf_legacy_loss": false,
70
+ "tie_encoder_decoder": false,
71
+ "tie_word_embeddings": true,
72
+ "tokenizer_class": null,
73
+ "top_k": 50,
74
+ "top_p": 1.0,
75
+ "torch_dtype": null,
76
+ "torchscript": false,
77
+ "transformers_version": "4.32.0.dev0",
78
+ "typical_p": 1.0,
79
+ "use_bfloat16": false
80
+ },
81
+ "auto_map": {
82
+ "AutoConfig": "mosaicml/mpt-7b--configuration_mpt.MPTConfig",
83
+ "AutoModelForCausalLM": "mosaicml/mpt-7b--modeling_mpt.MPTForCausalLM"
84
+ },
85
+ "d_model": 4096,
86
+ "emb_pdrop": 0,
87
+ "embedding_fraction": 1.0,
88
+ "expansion_ratio": 4,
89
+ "init_config": {
90
+ "emb_init_std": null,
91
+ "emb_init_uniform_lim": null,
92
+ "fan_mode": "fan_in",
93
+ "init_div_is_residual": true,
94
+ "init_gain": 0,
95
+ "init_nonlinearity": "relu",
96
+ "init_std": 0.02,
97
+ "name": "kaiming_normal_",
98
+ "verbose": 0
99
+ },
100
+ "init_device": "cpu",
101
+ "initializer_range": 0.02,
102
+ "layer_norm_epsilon": 1e-05,
103
+ "learned_pos_emb": true,
104
+ "logit_scale": null,
105
+ "max_seq_len": 2048,
106
+ "model_type": "mpt",
107
+ "n_heads": 32,
108
+ "n_layers": 32,
109
+ "no_bias": true,
110
+ "norm_type": "low_precision_layernorm",
111
+ "resid_pdrop": 0,
112
+ "tokenizer_name": "EleutherAI/gpt-neox-20b",
113
+ "torch_dtype": "bfloat16",
114
+ "transformers_version": "4.28.1",
115
+ "use_cache": false,
116
+ "verbose": 0,
117
+ "vocab_size": 50432
118
+ }
generation_config.json ADDED
@@ -0,0 +1,6 @@
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "eos_token_id": 0,
4
+ "transformers_version": "4.32.0.dev0",
5
+ "use_cache": false
6
+ }
pytorch_model-00001-of-00007.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8880f26aaa46cc2e06ff8b3377f25c5841b49298d4e88804af40fc96e87d3cd4
3
+ size 1889607935
pytorch_model-00002-of-00007.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:34ffe56e54aa3c9e0574fbe90840041a7860b3edbd41d04f73794456174c57b1
3
+ size 1879140121
pytorch_model-00003-of-00007.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0118c986b7fa1dc7f862fabe0c8404cbf226ef496b31e49d951d21acb678270a
3
+ size 1979794887
pytorch_model-00004-of-00007.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4688b0c69dd2714aef19bbbd9ac85f01d10535f64dc4275b68067a9472db3506
3
+ size 1912694553
pytorch_model-00005-of-00007.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bd0147e4ba0c8f02dd09b578d3f1c95018c99b20e117be1f59be1f53ba982f0c
3
+ size 1879131591
pytorch_model-00006-of-00007.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e260ae12dac8d6d2952eead92e24df96297a63bb64175e048391d223753cc495
3
+ size 1879140121
pytorch_model-00007-of-00007.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:131ca3760ef5e162998b00b6dc666dabf84d42276f184f3a5a1e4d1bf394e753
3
+ size 1879131253
pytorch_model.bin.index.json ADDED
@@ -0,0 +1,202 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metadata": {
3
+ "total_size": 13298573312
4
+ },
5
+ "weight_map": {
6
+ "lm_head.weight": "pytorch_model-00001-of-00007.bin",
7
+ "transformer.blocks.0.attn.Wqkv.weight": "pytorch_model-00001-of-00007.bin",
8
+ "transformer.blocks.0.attn.out_proj.weight": "pytorch_model-00001-of-00007.bin",
9
+ "transformer.blocks.0.ffn.down_proj.weight": "pytorch_model-00001-of-00007.bin",
10
+ "transformer.blocks.0.ffn.up_proj.weight": "pytorch_model-00001-of-00007.bin",
11
+ "transformer.blocks.0.norm_1.weight": "pytorch_model-00001-of-00007.bin",
12
+ "transformer.blocks.0.norm_2.weight": "pytorch_model-00001-of-00007.bin",
13
+ "transformer.blocks.1.attn.Wqkv.weight": "pytorch_model-00001-of-00007.bin",
14
+ "transformer.blocks.1.attn.out_proj.weight": "pytorch_model-00001-of-00007.bin",
15
+ "transformer.blocks.1.ffn.down_proj.weight": "pytorch_model-00001-of-00007.bin",
16
+ "transformer.blocks.1.ffn.up_proj.weight": "pytorch_model-00001-of-00007.bin",
17
+ "transformer.blocks.1.norm_1.weight": "pytorch_model-00001-of-00007.bin",
18
+ "transformer.blocks.1.norm_2.weight": "pytorch_model-00001-of-00007.bin",
19
+ "transformer.blocks.10.attn.Wqkv.weight": "pytorch_model-00003-of-00007.bin",
20
+ "transformer.blocks.10.attn.out_proj.weight": "pytorch_model-00003-of-00007.bin",
21
+ "transformer.blocks.10.ffn.down_proj.weight": "pytorch_model-00003-of-00007.bin",
22
+ "transformer.blocks.10.ffn.up_proj.weight": "pytorch_model-00003-of-00007.bin",
23
+ "transformer.blocks.10.norm_1.weight": "pytorch_model-00003-of-00007.bin",
24
+ "transformer.blocks.10.norm_2.weight": "pytorch_model-00003-of-00007.bin",
25
+ "transformer.blocks.11.attn.Wqkv.weight": "pytorch_model-00003-of-00007.bin",
26
+ "transformer.blocks.11.attn.out_proj.weight": "pytorch_model-00003-of-00007.bin",
27
+ "transformer.blocks.11.ffn.down_proj.weight": "pytorch_model-00003-of-00007.bin",
28
+ "transformer.blocks.11.ffn.up_proj.weight": "pytorch_model-00003-of-00007.bin",
29
+ "transformer.blocks.11.norm_1.weight": "pytorch_model-00003-of-00007.bin",
30
+ "transformer.blocks.11.norm_2.weight": "pytorch_model-00003-of-00007.bin",
31
+ "transformer.blocks.12.attn.Wqkv.weight": "pytorch_model-00003-of-00007.bin",
32
+ "transformer.blocks.12.attn.out_proj.weight": "pytorch_model-00003-of-00007.bin",
33
+ "transformer.blocks.12.ffn.down_proj.weight": "pytorch_model-00003-of-00007.bin",
34
+ "transformer.blocks.12.ffn.up_proj.weight": "pytorch_model-00003-of-00007.bin",
35
+ "transformer.blocks.12.norm_1.weight": "pytorch_model-00003-of-00007.bin",
36
+ "transformer.blocks.12.norm_2.weight": "pytorch_model-00003-of-00007.bin",
37
+ "transformer.blocks.13.attn.Wqkv.weight": "pytorch_model-00003-of-00007.bin",
38
+ "transformer.blocks.13.attn.out_proj.weight": "pytorch_model-00004-of-00007.bin",
39
+ "transformer.blocks.13.ffn.down_proj.weight": "pytorch_model-00004-of-00007.bin",
40
+ "transformer.blocks.13.ffn.up_proj.weight": "pytorch_model-00004-of-00007.bin",
41
+ "transformer.blocks.13.norm_1.weight": "pytorch_model-00003-of-00007.bin",
42
+ "transformer.blocks.13.norm_2.weight": "pytorch_model-00004-of-00007.bin",
43
+ "transformer.blocks.14.attn.Wqkv.weight": "pytorch_model-00004-of-00007.bin",
44
+ "transformer.blocks.14.attn.out_proj.weight": "pytorch_model-00004-of-00007.bin",
45
+ "transformer.blocks.14.ffn.down_proj.weight": "pytorch_model-00004-of-00007.bin",
46
+ "transformer.blocks.14.ffn.up_proj.weight": "pytorch_model-00004-of-00007.bin",
47
+ "transformer.blocks.14.norm_1.weight": "pytorch_model-00004-of-00007.bin",
48
+ "transformer.blocks.14.norm_2.weight": "pytorch_model-00004-of-00007.bin",
49
+ "transformer.blocks.15.attn.Wqkv.weight": "pytorch_model-00004-of-00007.bin",
50
+ "transformer.blocks.15.attn.out_proj.weight": "pytorch_model-00004-of-00007.bin",
51
+ "transformer.blocks.15.ffn.down_proj.weight": "pytorch_model-00004-of-00007.bin",
52
+ "transformer.blocks.15.ffn.up_proj.weight": "pytorch_model-00004-of-00007.bin",
53
+ "transformer.blocks.15.norm_1.weight": "pytorch_model-00004-of-00007.bin",
54
+ "transformer.blocks.15.norm_2.weight": "pytorch_model-00004-of-00007.bin",
55
+ "transformer.blocks.16.attn.Wqkv.weight": "pytorch_model-00004-of-00007.bin",
56
+ "transformer.blocks.16.attn.out_proj.weight": "pytorch_model-00004-of-00007.bin",
57
+ "transformer.blocks.16.ffn.down_proj.weight": "pytorch_model-00004-of-00007.bin",
58
+ "transformer.blocks.16.ffn.up_proj.weight": "pytorch_model-00004-of-00007.bin",
59
+ "transformer.blocks.16.norm_1.weight": "pytorch_model-00004-of-00007.bin",
60
+ "transformer.blocks.16.norm_2.weight": "pytorch_model-00004-of-00007.bin",
61
+ "transformer.blocks.17.attn.Wqkv.weight": "pytorch_model-00004-of-00007.bin",
62
+ "transformer.blocks.17.attn.out_proj.weight": "pytorch_model-00004-of-00007.bin",
63
+ "transformer.blocks.17.ffn.down_proj.weight": "pytorch_model-00004-of-00007.bin",
64
+ "transformer.blocks.17.ffn.up_proj.weight": "pytorch_model-00004-of-00007.bin",
65
+ "transformer.blocks.17.norm_1.weight": "pytorch_model-00004-of-00007.bin",
66
+ "transformer.blocks.17.norm_2.weight": "pytorch_model-00004-of-00007.bin",
67
+ "transformer.blocks.18.attn.Wqkv.weight": "pytorch_model-00005-of-00007.bin",
68
+ "transformer.blocks.18.attn.out_proj.weight": "pytorch_model-00005-of-00007.bin",
69
+ "transformer.blocks.18.ffn.down_proj.weight": "pytorch_model-00005-of-00007.bin",
70
+ "transformer.blocks.18.ffn.up_proj.weight": "pytorch_model-00005-of-00007.bin",
71
+ "transformer.blocks.18.norm_1.weight": "pytorch_model-00004-of-00007.bin",
72
+ "transformer.blocks.18.norm_2.weight": "pytorch_model-00005-of-00007.bin",
73
+ "transformer.blocks.19.attn.Wqkv.weight": "pytorch_model-00005-of-00007.bin",
74
+ "transformer.blocks.19.attn.out_proj.weight": "pytorch_model-00005-of-00007.bin",
75
+ "transformer.blocks.19.ffn.down_proj.weight": "pytorch_model-00005-of-00007.bin",
76
+ "transformer.blocks.19.ffn.up_proj.weight": "pytorch_model-00005-of-00007.bin",
77
+ "transformer.blocks.19.norm_1.weight": "pytorch_model-00005-of-00007.bin",
78
+ "transformer.blocks.19.norm_2.weight": "pytorch_model-00005-of-00007.bin",
79
+ "transformer.blocks.2.attn.Wqkv.weight": "pytorch_model-00001-of-00007.bin",
80
+ "transformer.blocks.2.attn.out_proj.weight": "pytorch_model-00001-of-00007.bin",
81
+ "transformer.blocks.2.ffn.down_proj.weight": "pytorch_model-00001-of-00007.bin",
82
+ "transformer.blocks.2.ffn.up_proj.weight": "pytorch_model-00001-of-00007.bin",
83
+ "transformer.blocks.2.norm_1.weight": "pytorch_model-00001-of-00007.bin",
84
+ "transformer.blocks.2.norm_2.weight": "pytorch_model-00001-of-00007.bin",
85
+ "transformer.blocks.20.attn.Wqkv.weight": "pytorch_model-00005-of-00007.bin",
86
+ "transformer.blocks.20.attn.out_proj.weight": "pytorch_model-00005-of-00007.bin",
87
+ "transformer.blocks.20.ffn.down_proj.weight": "pytorch_model-00005-of-00007.bin",
88
+ "transformer.blocks.20.ffn.up_proj.weight": "pytorch_model-00005-of-00007.bin",
89
+ "transformer.blocks.20.norm_1.weight": "pytorch_model-00005-of-00007.bin",
90
+ "transformer.blocks.20.norm_2.weight": "pytorch_model-00005-of-00007.bin",
91
+ "transformer.blocks.21.attn.Wqkv.weight": "pytorch_model-00005-of-00007.bin",
92
+ "transformer.blocks.21.attn.out_proj.weight": "pytorch_model-00005-of-00007.bin",
93
+ "transformer.blocks.21.ffn.down_proj.weight": "pytorch_model-00005-of-00007.bin",
94
+ "transformer.blocks.21.ffn.up_proj.weight": "pytorch_model-00005-of-00007.bin",
95
+ "transformer.blocks.21.norm_1.weight": "pytorch_model-00005-of-00007.bin",
96
+ "transformer.blocks.21.norm_2.weight": "pytorch_model-00005-of-00007.bin",
97
+ "transformer.blocks.22.attn.Wqkv.weight": "pytorch_model-00005-of-00007.bin",
98
+ "transformer.blocks.22.attn.out_proj.weight": "pytorch_model-00005-of-00007.bin",
99
+ "transformer.blocks.22.ffn.down_proj.weight": "pytorch_model-00006-of-00007.bin",
100
+ "transformer.blocks.22.ffn.up_proj.weight": "pytorch_model-00005-of-00007.bin",
101
+ "transformer.blocks.22.norm_1.weight": "pytorch_model-00005-of-00007.bin",
102
+ "transformer.blocks.22.norm_2.weight": "pytorch_model-00005-of-00007.bin",
103
+ "transformer.blocks.23.attn.Wqkv.weight": "pytorch_model-00006-of-00007.bin",
104
+ "transformer.blocks.23.attn.out_proj.weight": "pytorch_model-00006-of-00007.bin",
105
+ "transformer.blocks.23.ffn.down_proj.weight": "pytorch_model-00006-of-00007.bin",
106
+ "transformer.blocks.23.ffn.up_proj.weight": "pytorch_model-00006-of-00007.bin",
107
+ "transformer.blocks.23.norm_1.weight": "pytorch_model-00006-of-00007.bin",
108
+ "transformer.blocks.23.norm_2.weight": "pytorch_model-00006-of-00007.bin",
109
+ "transformer.blocks.24.attn.Wqkv.weight": "pytorch_model-00006-of-00007.bin",
110
+ "transformer.blocks.24.attn.out_proj.weight": "pytorch_model-00006-of-00007.bin",
111
+ "transformer.blocks.24.ffn.down_proj.weight": "pytorch_model-00006-of-00007.bin",
112
+ "transformer.blocks.24.ffn.up_proj.weight": "pytorch_model-00006-of-00007.bin",
113
+ "transformer.blocks.24.norm_1.weight": "pytorch_model-00006-of-00007.bin",
114
+ "transformer.blocks.24.norm_2.weight": "pytorch_model-00006-of-00007.bin",
115
+ "transformer.blocks.25.attn.Wqkv.weight": "pytorch_model-00006-of-00007.bin",
116
+ "transformer.blocks.25.attn.out_proj.weight": "pytorch_model-00006-of-00007.bin",
117
+ "transformer.blocks.25.ffn.down_proj.weight": "pytorch_model-00006-of-00007.bin",
118
+ "transformer.blocks.25.ffn.up_proj.weight": "pytorch_model-00006-of-00007.bin",
119
+ "transformer.blocks.25.norm_1.weight": "pytorch_model-00006-of-00007.bin",
120
+ "transformer.blocks.25.norm_2.weight": "pytorch_model-00006-of-00007.bin",
121
+ "transformer.blocks.26.attn.Wqkv.weight": "pytorch_model-00006-of-00007.bin",
122
+ "transformer.blocks.26.attn.out_proj.weight": "pytorch_model-00006-of-00007.bin",
123
+ "transformer.blocks.26.ffn.down_proj.weight": "pytorch_model-00006-of-00007.bin",
124
+ "transformer.blocks.26.ffn.up_proj.weight": "pytorch_model-00006-of-00007.bin",
125
+ "transformer.blocks.26.norm_1.weight": "pytorch_model-00006-of-00007.bin",
126
+ "transformer.blocks.26.norm_2.weight": "pytorch_model-00006-of-00007.bin",
127
+ "transformer.blocks.27.attn.Wqkv.weight": "pytorch_model-00006-of-00007.bin",
128
+ "transformer.blocks.27.attn.out_proj.weight": "pytorch_model-00006-of-00007.bin",
129
+ "transformer.blocks.27.ffn.down_proj.weight": "pytorch_model-00007-of-00007.bin",
130
+ "transformer.blocks.27.ffn.up_proj.weight": "pytorch_model-00007-of-00007.bin",
131
+ "transformer.blocks.27.norm_1.weight": "pytorch_model-00006-of-00007.bin",
132
+ "transformer.blocks.27.norm_2.weight": "pytorch_model-00006-of-00007.bin",
133
+ "transformer.blocks.28.attn.Wqkv.weight": "pytorch_model-00007-of-00007.bin",
134
+ "transformer.blocks.28.attn.out_proj.weight": "pytorch_model-00007-of-00007.bin",
135
+ "transformer.blocks.28.ffn.down_proj.weight": "pytorch_model-00007-of-00007.bin",
136
+ "transformer.blocks.28.ffn.up_proj.weight": "pytorch_model-00007-of-00007.bin",
137
+ "transformer.blocks.28.norm_1.weight": "pytorch_model-00007-of-00007.bin",
138
+ "transformer.blocks.28.norm_2.weight": "pytorch_model-00007-of-00007.bin",
139
+ "transformer.blocks.29.attn.Wqkv.weight": "pytorch_model-00007-of-00007.bin",
140
+ "transformer.blocks.29.attn.out_proj.weight": "pytorch_model-00007-of-00007.bin",
141
+ "transformer.blocks.29.ffn.down_proj.weight": "pytorch_model-00007-of-00007.bin",
142
+ "transformer.blocks.29.ffn.up_proj.weight": "pytorch_model-00007-of-00007.bin",
143
+ "transformer.blocks.29.norm_1.weight": "pytorch_model-00007-of-00007.bin",
144
+ "transformer.blocks.29.norm_2.weight": "pytorch_model-00007-of-00007.bin",
145
+ "transformer.blocks.3.attn.Wqkv.weight": "pytorch_model-00001-of-00007.bin",
146
+ "transformer.blocks.3.attn.out_proj.weight": "pytorch_model-00001-of-00007.bin",
147
+ "transformer.blocks.3.ffn.down_proj.weight": "pytorch_model-00002-of-00007.bin",
148
+ "transformer.blocks.3.ffn.up_proj.weight": "pytorch_model-00001-of-00007.bin",
149
+ "transformer.blocks.3.norm_1.weight": "pytorch_model-00001-of-00007.bin",
150
+ "transformer.blocks.3.norm_2.weight": "pytorch_model-00001-of-00007.bin",
151
+ "transformer.blocks.30.attn.Wqkv.weight": "pytorch_model-00007-of-00007.bin",
152
+ "transformer.blocks.30.attn.out_proj.weight": "pytorch_model-00007-of-00007.bin",
153
+ "transformer.blocks.30.ffn.down_proj.weight": "pytorch_model-00007-of-00007.bin",
154
+ "transformer.blocks.30.ffn.up_proj.weight": "pytorch_model-00007-of-00007.bin",
155
+ "transformer.blocks.30.norm_1.weight": "pytorch_model-00007-of-00007.bin",
156
+ "transformer.blocks.30.norm_2.weight": "pytorch_model-00007-of-00007.bin",
157
+ "transformer.blocks.31.attn.Wqkv.weight": "pytorch_model-00007-of-00007.bin",
158
+ "transformer.blocks.31.attn.out_proj.weight": "pytorch_model-00007-of-00007.bin",
159
+ "transformer.blocks.31.ffn.down_proj.weight": "pytorch_model-00007-of-00007.bin",
160
+ "transformer.blocks.31.ffn.up_proj.weight": "pytorch_model-00007-of-00007.bin",
161
+ "transformer.blocks.31.norm_1.weight": "pytorch_model-00007-of-00007.bin",
162
+ "transformer.blocks.31.norm_2.weight": "pytorch_model-00007-of-00007.bin",
163
+ "transformer.blocks.4.attn.Wqkv.weight": "pytorch_model-00002-of-00007.bin",
164
+ "transformer.blocks.4.attn.out_proj.weight": "pytorch_model-00002-of-00007.bin",
165
+ "transformer.blocks.4.ffn.down_proj.weight": "pytorch_model-00002-of-00007.bin",
166
+ "transformer.blocks.4.ffn.up_proj.weight": "pytorch_model-00002-of-00007.bin",
167
+ "transformer.blocks.4.norm_1.weight": "pytorch_model-00002-of-00007.bin",
168
+ "transformer.blocks.4.norm_2.weight": "pytorch_model-00002-of-00007.bin",
169
+ "transformer.blocks.5.attn.Wqkv.weight": "pytorch_model-00002-of-00007.bin",
170
+ "transformer.blocks.5.attn.out_proj.weight": "pytorch_model-00002-of-00007.bin",
171
+ "transformer.blocks.5.ffn.down_proj.weight": "pytorch_model-00002-of-00007.bin",
172
+ "transformer.blocks.5.ffn.up_proj.weight": "pytorch_model-00002-of-00007.bin",
173
+ "transformer.blocks.5.norm_1.weight": "pytorch_model-00002-of-00007.bin",
174
+ "transformer.blocks.5.norm_2.weight": "pytorch_model-00002-of-00007.bin",
175
+ "transformer.blocks.6.attn.Wqkv.weight": "pytorch_model-00002-of-00007.bin",
176
+ "transformer.blocks.6.attn.out_proj.weight": "pytorch_model-00002-of-00007.bin",
177
+ "transformer.blocks.6.ffn.down_proj.weight": "pytorch_model-00002-of-00007.bin",
178
+ "transformer.blocks.6.ffn.up_proj.weight": "pytorch_model-00002-of-00007.bin",
179
+ "transformer.blocks.6.norm_1.weight": "pytorch_model-00002-of-00007.bin",
180
+ "transformer.blocks.6.norm_2.weight": "pytorch_model-00002-of-00007.bin",
181
+ "transformer.blocks.7.attn.Wqkv.weight": "pytorch_model-00002-of-00007.bin",
182
+ "transformer.blocks.7.attn.out_proj.weight": "pytorch_model-00002-of-00007.bin",
183
+ "transformer.blocks.7.ffn.down_proj.weight": "pytorch_model-00002-of-00007.bin",
184
+ "transformer.blocks.7.ffn.up_proj.weight": "pytorch_model-00002-of-00007.bin",
185
+ "transformer.blocks.7.norm_1.weight": "pytorch_model-00002-of-00007.bin",
186
+ "transformer.blocks.7.norm_2.weight": "pytorch_model-00002-of-00007.bin",
187
+ "transformer.blocks.8.attn.Wqkv.weight": "pytorch_model-00002-of-00007.bin",
188
+ "transformer.blocks.8.attn.out_proj.weight": "pytorch_model-00002-of-00007.bin",
189
+ "transformer.blocks.8.ffn.down_proj.weight": "pytorch_model-00003-of-00007.bin",
190
+ "transformer.blocks.8.ffn.up_proj.weight": "pytorch_model-00003-of-00007.bin",
191
+ "transformer.blocks.8.norm_1.weight": "pytorch_model-00002-of-00007.bin",
192
+ "transformer.blocks.8.norm_2.weight": "pytorch_model-00002-of-00007.bin",
193
+ "transformer.blocks.9.attn.Wqkv.weight": "pytorch_model-00003-of-00007.bin",
194
+ "transformer.blocks.9.attn.out_proj.weight": "pytorch_model-00003-of-00007.bin",
195
+ "transformer.blocks.9.ffn.down_proj.weight": "pytorch_model-00003-of-00007.bin",
196
+ "transformer.blocks.9.ffn.up_proj.weight": "pytorch_model-00003-of-00007.bin",
197
+ "transformer.blocks.9.norm_1.weight": "pytorch_model-00003-of-00007.bin",
198
+ "transformer.blocks.9.norm_2.weight": "pytorch_model-00003-of-00007.bin",
199
+ "transformer.norm_f.weight": "pytorch_model-00007-of-00007.bin",
200
+ "transformer.wte.weight": "pytorch_model-00001-of-00007.bin"
201
+ }
202
+ }