Upload ExaoneForCausalLM
Browse files- config.json +33 -0
- generation_config.json +7 -0
- model-00001-of-00004.safetensors +3 -0
- model-00002-of-00004.safetensors +3 -0
- model-00003-of-00004.safetensors +3 -0
- model-00004-of-00004.safetensors +3 -0
- model.safetensors.index.json +298 -0
config.json
ADDED
@@ -0,0 +1,33 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "LGAI-EXAONE/EXAONE-3.0-7.8B-Instruct",
|
3 |
+
"activation_function": "silu",
|
4 |
+
"architectures": [
|
5 |
+
"ExaoneForCausalLM"
|
6 |
+
],
|
7 |
+
"attention_dropout": 0.0,
|
8 |
+
"auto_map": {
|
9 |
+
"AutoConfig": "LGAI-EXAONE/EXAONE-3.0-7.8B-Instruct--configuration_exaone.ExaoneConfig",
|
10 |
+
"AutoModelForCausalLM": "LGAI-EXAONE/EXAONE-3.0-7.8B-Instruct--modeling_exaone.ExaoneForCausalLM",
|
11 |
+
"AutoModelForSequenceClassification": "LGAI-EXAONE/EXAONE-3.0-7.8B-Instruct--modeling_exaone.ExaoneForSequenceClassification"
|
12 |
+
},
|
13 |
+
"bos_token_id": 1,
|
14 |
+
"embed_dropout": 0.0,
|
15 |
+
"eos_token_id": 361,
|
16 |
+
"hidden_size": 4096,
|
17 |
+
"initializer_range": 0.02,
|
18 |
+
"intermediate_size": 14336,
|
19 |
+
"layer_norm_epsilon": 1e-05,
|
20 |
+
"max_position_embeddings": 4096,
|
21 |
+
"model_type": "exaone",
|
22 |
+
"num_attention_heads": 32,
|
23 |
+
"num_key_value_heads": 8,
|
24 |
+
"num_layers": 32,
|
25 |
+
"pad_token_id": 0,
|
26 |
+
"rope_scaling": null,
|
27 |
+
"rope_theta": 500000.0,
|
28 |
+
"tie_word_embeddings": false,
|
29 |
+
"torch_dtype": "float16",
|
30 |
+
"transformers_version": "4.46.2",
|
31 |
+
"use_cache": true,
|
32 |
+
"vocab_size": 102400
|
33 |
+
}
|
generation_config.json
ADDED
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_from_model_config": true,
|
3 |
+
"bos_token_id": 1,
|
4 |
+
"eos_token_id": 361,
|
5 |
+
"pad_token_id": 0,
|
6 |
+
"transformers_version": "4.46.2"
|
7 |
+
}
|
model-00001-of-00004.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f54165e07c670c590830dbcd9f7736c7d1966b7606040206311451b56e5e05eb
|
3 |
+
size 4966229888
|
model-00002-of-00004.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:0eb752078cf24a313e54192cd966080e3390c526fa06306bb1fd426a20f5365d
|
3 |
+
size 4915915984
|
model-00003-of-00004.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:7a115931d89038b854f93e6c4876ee9a881932a1d8c7247a52dbd4376080a959
|
3 |
+
size 4915924296
|
model-00004-of-00004.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:806f3e67b07fa035afb1f5f68dac3e4e150016596029be74c263e52d7a48ef82
|
3 |
+
size 838860928
|
model.safetensors.index.json
ADDED
@@ -0,0 +1,298 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"metadata": {
|
3 |
+
"total_size": 15636897792
|
4 |
+
},
|
5 |
+
"weight_map": {
|
6 |
+
"lm_head.weight": "model-00004-of-00004.safetensors",
|
7 |
+
"transformer.h.0.attn.attention.k_proj.weight": "model-00001-of-00004.safetensors",
|
8 |
+
"transformer.h.0.attn.attention.out_proj.weight": "model-00001-of-00004.safetensors",
|
9 |
+
"transformer.h.0.attn.attention.q_proj.weight": "model-00001-of-00004.safetensors",
|
10 |
+
"transformer.h.0.attn.attention.v_proj.weight": "model-00001-of-00004.safetensors",
|
11 |
+
"transformer.h.0.ln_1.weight": "model-00001-of-00004.safetensors",
|
12 |
+
"transformer.h.0.ln_2.weight": "model-00001-of-00004.safetensors",
|
13 |
+
"transformer.h.0.mlp.c_fc_0.weight": "model-00001-of-00004.safetensors",
|
14 |
+
"transformer.h.0.mlp.c_fc_1.weight": "model-00001-of-00004.safetensors",
|
15 |
+
"transformer.h.0.mlp.c_proj.weight": "model-00001-of-00004.safetensors",
|
16 |
+
"transformer.h.1.attn.attention.k_proj.weight": "model-00001-of-00004.safetensors",
|
17 |
+
"transformer.h.1.attn.attention.out_proj.weight": "model-00001-of-00004.safetensors",
|
18 |
+
"transformer.h.1.attn.attention.q_proj.weight": "model-00001-of-00004.safetensors",
|
19 |
+
"transformer.h.1.attn.attention.v_proj.weight": "model-00001-of-00004.safetensors",
|
20 |
+
"transformer.h.1.ln_1.weight": "model-00001-of-00004.safetensors",
|
21 |
+
"transformer.h.1.ln_2.weight": "model-00001-of-00004.safetensors",
|
22 |
+
"transformer.h.1.mlp.c_fc_0.weight": "model-00001-of-00004.safetensors",
|
23 |
+
"transformer.h.1.mlp.c_fc_1.weight": "model-00001-of-00004.safetensors",
|
24 |
+
"transformer.h.1.mlp.c_proj.weight": "model-00001-of-00004.safetensors",
|
25 |
+
"transformer.h.10.attn.attention.k_proj.weight": "model-00002-of-00004.safetensors",
|
26 |
+
"transformer.h.10.attn.attention.out_proj.weight": "model-00002-of-00004.safetensors",
|
27 |
+
"transformer.h.10.attn.attention.q_proj.weight": "model-00002-of-00004.safetensors",
|
28 |
+
"transformer.h.10.attn.attention.v_proj.weight": "model-00002-of-00004.safetensors",
|
29 |
+
"transformer.h.10.ln_1.weight": "model-00002-of-00004.safetensors",
|
30 |
+
"transformer.h.10.ln_2.weight": "model-00002-of-00004.safetensors",
|
31 |
+
"transformer.h.10.mlp.c_fc_0.weight": "model-00002-of-00004.safetensors",
|
32 |
+
"transformer.h.10.mlp.c_fc_1.weight": "model-00002-of-00004.safetensors",
|
33 |
+
"transformer.h.10.mlp.c_proj.weight": "model-00002-of-00004.safetensors",
|
34 |
+
"transformer.h.11.attn.attention.k_proj.weight": "model-00002-of-00004.safetensors",
|
35 |
+
"transformer.h.11.attn.attention.out_proj.weight": "model-00002-of-00004.safetensors",
|
36 |
+
"transformer.h.11.attn.attention.q_proj.weight": "model-00002-of-00004.safetensors",
|
37 |
+
"transformer.h.11.attn.attention.v_proj.weight": "model-00002-of-00004.safetensors",
|
38 |
+
"transformer.h.11.ln_1.weight": "model-00002-of-00004.safetensors",
|
39 |
+
"transformer.h.11.ln_2.weight": "model-00002-of-00004.safetensors",
|
40 |
+
"transformer.h.11.mlp.c_fc_0.weight": "model-00002-of-00004.safetensors",
|
41 |
+
"transformer.h.11.mlp.c_fc_1.weight": "model-00002-of-00004.safetensors",
|
42 |
+
"transformer.h.11.mlp.c_proj.weight": "model-00002-of-00004.safetensors",
|
43 |
+
"transformer.h.12.attn.attention.k_proj.weight": "model-00002-of-00004.safetensors",
|
44 |
+
"transformer.h.12.attn.attention.out_proj.weight": "model-00002-of-00004.safetensors",
|
45 |
+
"transformer.h.12.attn.attention.q_proj.weight": "model-00002-of-00004.safetensors",
|
46 |
+
"transformer.h.12.attn.attention.v_proj.weight": "model-00002-of-00004.safetensors",
|
47 |
+
"transformer.h.12.ln_1.weight": "model-00002-of-00004.safetensors",
|
48 |
+
"transformer.h.12.ln_2.weight": "model-00002-of-00004.safetensors",
|
49 |
+
"transformer.h.12.mlp.c_fc_0.weight": "model-00002-of-00004.safetensors",
|
50 |
+
"transformer.h.12.mlp.c_fc_1.weight": "model-00002-of-00004.safetensors",
|
51 |
+
"transformer.h.12.mlp.c_proj.weight": "model-00002-of-00004.safetensors",
|
52 |
+
"transformer.h.13.attn.attention.k_proj.weight": "model-00002-of-00004.safetensors",
|
53 |
+
"transformer.h.13.attn.attention.out_proj.weight": "model-00002-of-00004.safetensors",
|
54 |
+
"transformer.h.13.attn.attention.q_proj.weight": "model-00002-of-00004.safetensors",
|
55 |
+
"transformer.h.13.attn.attention.v_proj.weight": "model-00002-of-00004.safetensors",
|
56 |
+
"transformer.h.13.ln_1.weight": "model-00002-of-00004.safetensors",
|
57 |
+
"transformer.h.13.ln_2.weight": "model-00002-of-00004.safetensors",
|
58 |
+
"transformer.h.13.mlp.c_fc_0.weight": "model-00002-of-00004.safetensors",
|
59 |
+
"transformer.h.13.mlp.c_fc_1.weight": "model-00002-of-00004.safetensors",
|
60 |
+
"transformer.h.13.mlp.c_proj.weight": "model-00002-of-00004.safetensors",
|
61 |
+
"transformer.h.14.attn.attention.k_proj.weight": "model-00002-of-00004.safetensors",
|
62 |
+
"transformer.h.14.attn.attention.out_proj.weight": "model-00002-of-00004.safetensors",
|
63 |
+
"transformer.h.14.attn.attention.q_proj.weight": "model-00002-of-00004.safetensors",
|
64 |
+
"transformer.h.14.attn.attention.v_proj.weight": "model-00002-of-00004.safetensors",
|
65 |
+
"transformer.h.14.ln_1.weight": "model-00002-of-00004.safetensors",
|
66 |
+
"transformer.h.14.ln_2.weight": "model-00002-of-00004.safetensors",
|
67 |
+
"transformer.h.14.mlp.c_fc_0.weight": "model-00002-of-00004.safetensors",
|
68 |
+
"transformer.h.14.mlp.c_fc_1.weight": "model-00002-of-00004.safetensors",
|
69 |
+
"transformer.h.14.mlp.c_proj.weight": "model-00002-of-00004.safetensors",
|
70 |
+
"transformer.h.15.attn.attention.k_proj.weight": "model-00002-of-00004.safetensors",
|
71 |
+
"transformer.h.15.attn.attention.out_proj.weight": "model-00002-of-00004.safetensors",
|
72 |
+
"transformer.h.15.attn.attention.q_proj.weight": "model-00002-of-00004.safetensors",
|
73 |
+
"transformer.h.15.attn.attention.v_proj.weight": "model-00002-of-00004.safetensors",
|
74 |
+
"transformer.h.15.ln_1.weight": "model-00002-of-00004.safetensors",
|
75 |
+
"transformer.h.15.ln_2.weight": "model-00002-of-00004.safetensors",
|
76 |
+
"transformer.h.15.mlp.c_fc_0.weight": "model-00002-of-00004.safetensors",
|
77 |
+
"transformer.h.15.mlp.c_fc_1.weight": "model-00002-of-00004.safetensors",
|
78 |
+
"transformer.h.15.mlp.c_proj.weight": "model-00002-of-00004.safetensors",
|
79 |
+
"transformer.h.16.attn.attention.k_proj.weight": "model-00002-of-00004.safetensors",
|
80 |
+
"transformer.h.16.attn.attention.out_proj.weight": "model-00002-of-00004.safetensors",
|
81 |
+
"transformer.h.16.attn.attention.q_proj.weight": "model-00002-of-00004.safetensors",
|
82 |
+
"transformer.h.16.attn.attention.v_proj.weight": "model-00002-of-00004.safetensors",
|
83 |
+
"transformer.h.16.ln_1.weight": "model-00002-of-00004.safetensors",
|
84 |
+
"transformer.h.16.ln_2.weight": "model-00002-of-00004.safetensors",
|
85 |
+
"transformer.h.16.mlp.c_fc_0.weight": "model-00002-of-00004.safetensors",
|
86 |
+
"transformer.h.16.mlp.c_fc_1.weight": "model-00002-of-00004.safetensors",
|
87 |
+
"transformer.h.16.mlp.c_proj.weight": "model-00002-of-00004.safetensors",
|
88 |
+
"transformer.h.17.attn.attention.k_proj.weight": "model-00002-of-00004.safetensors",
|
89 |
+
"transformer.h.17.attn.attention.out_proj.weight": "model-00002-of-00004.safetensors",
|
90 |
+
"transformer.h.17.attn.attention.q_proj.weight": "model-00002-of-00004.safetensors",
|
91 |
+
"transformer.h.17.attn.attention.v_proj.weight": "model-00002-of-00004.safetensors",
|
92 |
+
"transformer.h.17.ln_1.weight": "model-00002-of-00004.safetensors",
|
93 |
+
"transformer.h.17.ln_2.weight": "model-00002-of-00004.safetensors",
|
94 |
+
"transformer.h.17.mlp.c_fc_0.weight": "model-00002-of-00004.safetensors",
|
95 |
+
"transformer.h.17.mlp.c_fc_1.weight": "model-00002-of-00004.safetensors",
|
96 |
+
"transformer.h.17.mlp.c_proj.weight": "model-00002-of-00004.safetensors",
|
97 |
+
"transformer.h.18.attn.attention.k_proj.weight": "model-00002-of-00004.safetensors",
|
98 |
+
"transformer.h.18.attn.attention.out_proj.weight": "model-00002-of-00004.safetensors",
|
99 |
+
"transformer.h.18.attn.attention.q_proj.weight": "model-00002-of-00004.safetensors",
|
100 |
+
"transformer.h.18.attn.attention.v_proj.weight": "model-00002-of-00004.safetensors",
|
101 |
+
"transformer.h.18.ln_1.weight": "model-00002-of-00004.safetensors",
|
102 |
+
"transformer.h.18.ln_2.weight": "model-00002-of-00004.safetensors",
|
103 |
+
"transformer.h.18.mlp.c_fc_0.weight": "model-00002-of-00004.safetensors",
|
104 |
+
"transformer.h.18.mlp.c_fc_1.weight": "model-00002-of-00004.safetensors",
|
105 |
+
"transformer.h.18.mlp.c_proj.weight": "model-00002-of-00004.safetensors",
|
106 |
+
"transformer.h.19.attn.attention.k_proj.weight": "model-00002-of-00004.safetensors",
|
107 |
+
"transformer.h.19.attn.attention.out_proj.weight": "model-00002-of-00004.safetensors",
|
108 |
+
"transformer.h.19.attn.attention.q_proj.weight": "model-00002-of-00004.safetensors",
|
109 |
+
"transformer.h.19.attn.attention.v_proj.weight": "model-00002-of-00004.safetensors",
|
110 |
+
"transformer.h.19.ln_1.weight": "model-00002-of-00004.safetensors",
|
111 |
+
"transformer.h.19.ln_2.weight": "model-00002-of-00004.safetensors",
|
112 |
+
"transformer.h.19.mlp.c_fc_0.weight": "model-00002-of-00004.safetensors",
|
113 |
+
"transformer.h.19.mlp.c_fc_1.weight": "model-00002-of-00004.safetensors",
|
114 |
+
"transformer.h.19.mlp.c_proj.weight": "model-00002-of-00004.safetensors",
|
115 |
+
"transformer.h.2.attn.attention.k_proj.weight": "model-00001-of-00004.safetensors",
|
116 |
+
"transformer.h.2.attn.attention.out_proj.weight": "model-00001-of-00004.safetensors",
|
117 |
+
"transformer.h.2.attn.attention.q_proj.weight": "model-00001-of-00004.safetensors",
|
118 |
+
"transformer.h.2.attn.attention.v_proj.weight": "model-00001-of-00004.safetensors",
|
119 |
+
"transformer.h.2.ln_1.weight": "model-00001-of-00004.safetensors",
|
120 |
+
"transformer.h.2.ln_2.weight": "model-00001-of-00004.safetensors",
|
121 |
+
"transformer.h.2.mlp.c_fc_0.weight": "model-00001-of-00004.safetensors",
|
122 |
+
"transformer.h.2.mlp.c_fc_1.weight": "model-00001-of-00004.safetensors",
|
123 |
+
"transformer.h.2.mlp.c_proj.weight": "model-00001-of-00004.safetensors",
|
124 |
+
"transformer.h.20.attn.attention.k_proj.weight": "model-00002-of-00004.safetensors",
|
125 |
+
"transformer.h.20.attn.attention.out_proj.weight": "model-00002-of-00004.safetensors",
|
126 |
+
"transformer.h.20.attn.attention.q_proj.weight": "model-00002-of-00004.safetensors",
|
127 |
+
"transformer.h.20.attn.attention.v_proj.weight": "model-00002-of-00004.safetensors",
|
128 |
+
"transformer.h.20.ln_1.weight": "model-00002-of-00004.safetensors",
|
129 |
+
"transformer.h.20.ln_2.weight": "model-00002-of-00004.safetensors",
|
130 |
+
"transformer.h.20.mlp.c_fc_0.weight": "model-00002-of-00004.safetensors",
|
131 |
+
"transformer.h.20.mlp.c_fc_1.weight": "model-00002-of-00004.safetensors",
|
132 |
+
"transformer.h.20.mlp.c_proj.weight": "model-00003-of-00004.safetensors",
|
133 |
+
"transformer.h.21.attn.attention.k_proj.weight": "model-00003-of-00004.safetensors",
|
134 |
+
"transformer.h.21.attn.attention.out_proj.weight": "model-00003-of-00004.safetensors",
|
135 |
+
"transformer.h.21.attn.attention.q_proj.weight": "model-00003-of-00004.safetensors",
|
136 |
+
"transformer.h.21.attn.attention.v_proj.weight": "model-00003-of-00004.safetensors",
|
137 |
+
"transformer.h.21.ln_1.weight": "model-00003-of-00004.safetensors",
|
138 |
+
"transformer.h.21.ln_2.weight": "model-00003-of-00004.safetensors",
|
139 |
+
"transformer.h.21.mlp.c_fc_0.weight": "model-00003-of-00004.safetensors",
|
140 |
+
"transformer.h.21.mlp.c_fc_1.weight": "model-00003-of-00004.safetensors",
|
141 |
+
"transformer.h.21.mlp.c_proj.weight": "model-00003-of-00004.safetensors",
|
142 |
+
"transformer.h.22.attn.attention.k_proj.weight": "model-00003-of-00004.safetensors",
|
143 |
+
"transformer.h.22.attn.attention.out_proj.weight": "model-00003-of-00004.safetensors",
|
144 |
+
"transformer.h.22.attn.attention.q_proj.weight": "model-00003-of-00004.safetensors",
|
145 |
+
"transformer.h.22.attn.attention.v_proj.weight": "model-00003-of-00004.safetensors",
|
146 |
+
"transformer.h.22.ln_1.weight": "model-00003-of-00004.safetensors",
|
147 |
+
"transformer.h.22.ln_2.weight": "model-00003-of-00004.safetensors",
|
148 |
+
"transformer.h.22.mlp.c_fc_0.weight": "model-00003-of-00004.safetensors",
|
149 |
+
"transformer.h.22.mlp.c_fc_1.weight": "model-00003-of-00004.safetensors",
|
150 |
+
"transformer.h.22.mlp.c_proj.weight": "model-00003-of-00004.safetensors",
|
151 |
+
"transformer.h.23.attn.attention.k_proj.weight": "model-00003-of-00004.safetensors",
|
152 |
+
"transformer.h.23.attn.attention.out_proj.weight": "model-00003-of-00004.safetensors",
|
153 |
+
"transformer.h.23.attn.attention.q_proj.weight": "model-00003-of-00004.safetensors",
|
154 |
+
"transformer.h.23.attn.attention.v_proj.weight": "model-00003-of-00004.safetensors",
|
155 |
+
"transformer.h.23.ln_1.weight": "model-00003-of-00004.safetensors",
|
156 |
+
"transformer.h.23.ln_2.weight": "model-00003-of-00004.safetensors",
|
157 |
+
"transformer.h.23.mlp.c_fc_0.weight": "model-00003-of-00004.safetensors",
|
158 |
+
"transformer.h.23.mlp.c_fc_1.weight": "model-00003-of-00004.safetensors",
|
159 |
+
"transformer.h.23.mlp.c_proj.weight": "model-00003-of-00004.safetensors",
|
160 |
+
"transformer.h.24.attn.attention.k_proj.weight": "model-00003-of-00004.safetensors",
|
161 |
+
"transformer.h.24.attn.attention.out_proj.weight": "model-00003-of-00004.safetensors",
|
162 |
+
"transformer.h.24.attn.attention.q_proj.weight": "model-00003-of-00004.safetensors",
|
163 |
+
"transformer.h.24.attn.attention.v_proj.weight": "model-00003-of-00004.safetensors",
|
164 |
+
"transformer.h.24.ln_1.weight": "model-00003-of-00004.safetensors",
|
165 |
+
"transformer.h.24.ln_2.weight": "model-00003-of-00004.safetensors",
|
166 |
+
"transformer.h.24.mlp.c_fc_0.weight": "model-00003-of-00004.safetensors",
|
167 |
+
"transformer.h.24.mlp.c_fc_1.weight": "model-00003-of-00004.safetensors",
|
168 |
+
"transformer.h.24.mlp.c_proj.weight": "model-00003-of-00004.safetensors",
|
169 |
+
"transformer.h.25.attn.attention.k_proj.weight": "model-00003-of-00004.safetensors",
|
170 |
+
"transformer.h.25.attn.attention.out_proj.weight": "model-00003-of-00004.safetensors",
|
171 |
+
"transformer.h.25.attn.attention.q_proj.weight": "model-00003-of-00004.safetensors",
|
172 |
+
"transformer.h.25.attn.attention.v_proj.weight": "model-00003-of-00004.safetensors",
|
173 |
+
"transformer.h.25.ln_1.weight": "model-00003-of-00004.safetensors",
|
174 |
+
"transformer.h.25.ln_2.weight": "model-00003-of-00004.safetensors",
|
175 |
+
"transformer.h.25.mlp.c_fc_0.weight": "model-00003-of-00004.safetensors",
|
176 |
+
"transformer.h.25.mlp.c_fc_1.weight": "model-00003-of-00004.safetensors",
|
177 |
+
"transformer.h.25.mlp.c_proj.weight": "model-00003-of-00004.safetensors",
|
178 |
+
"transformer.h.26.attn.attention.k_proj.weight": "model-00003-of-00004.safetensors",
|
179 |
+
"transformer.h.26.attn.attention.out_proj.weight": "model-00003-of-00004.safetensors",
|
180 |
+
"transformer.h.26.attn.attention.q_proj.weight": "model-00003-of-00004.safetensors",
|
181 |
+
"transformer.h.26.attn.attention.v_proj.weight": "model-00003-of-00004.safetensors",
|
182 |
+
"transformer.h.26.ln_1.weight": "model-00003-of-00004.safetensors",
|
183 |
+
"transformer.h.26.ln_2.weight": "model-00003-of-00004.safetensors",
|
184 |
+
"transformer.h.26.mlp.c_fc_0.weight": "model-00003-of-00004.safetensors",
|
185 |
+
"transformer.h.26.mlp.c_fc_1.weight": "model-00003-of-00004.safetensors",
|
186 |
+
"transformer.h.26.mlp.c_proj.weight": "model-00003-of-00004.safetensors",
|
187 |
+
"transformer.h.27.attn.attention.k_proj.weight": "model-00003-of-00004.safetensors",
|
188 |
+
"transformer.h.27.attn.attention.out_proj.weight": "model-00003-of-00004.safetensors",
|
189 |
+
"transformer.h.27.attn.attention.q_proj.weight": "model-00003-of-00004.safetensors",
|
190 |
+
"transformer.h.27.attn.attention.v_proj.weight": "model-00003-of-00004.safetensors",
|
191 |
+
"transformer.h.27.ln_1.weight": "model-00003-of-00004.safetensors",
|
192 |
+
"transformer.h.27.ln_2.weight": "model-00003-of-00004.safetensors",
|
193 |
+
"transformer.h.27.mlp.c_fc_0.weight": "model-00003-of-00004.safetensors",
|
194 |
+
"transformer.h.27.mlp.c_fc_1.weight": "model-00003-of-00004.safetensors",
|
195 |
+
"transformer.h.27.mlp.c_proj.weight": "model-00003-of-00004.safetensors",
|
196 |
+
"transformer.h.28.attn.attention.k_proj.weight": "model-00003-of-00004.safetensors",
|
197 |
+
"transformer.h.28.attn.attention.out_proj.weight": "model-00003-of-00004.safetensors",
|
198 |
+
"transformer.h.28.attn.attention.q_proj.weight": "model-00003-of-00004.safetensors",
|
199 |
+
"transformer.h.28.attn.attention.v_proj.weight": "model-00003-of-00004.safetensors",
|
200 |
+
"transformer.h.28.ln_1.weight": "model-00003-of-00004.safetensors",
|
201 |
+
"transformer.h.28.ln_2.weight": "model-00003-of-00004.safetensors",
|
202 |
+
"transformer.h.28.mlp.c_fc_0.weight": "model-00003-of-00004.safetensors",
|
203 |
+
"transformer.h.28.mlp.c_fc_1.weight": "model-00003-of-00004.safetensors",
|
204 |
+
"transformer.h.28.mlp.c_proj.weight": "model-00003-of-00004.safetensors",
|
205 |
+
"transformer.h.29.attn.attention.k_proj.weight": "model-00003-of-00004.safetensors",
|
206 |
+
"transformer.h.29.attn.attention.out_proj.weight": "model-00003-of-00004.safetensors",
|
207 |
+
"transformer.h.29.attn.attention.q_proj.weight": "model-00003-of-00004.safetensors",
|
208 |
+
"transformer.h.29.attn.attention.v_proj.weight": "model-00003-of-00004.safetensors",
|
209 |
+
"transformer.h.29.ln_1.weight": "model-00003-of-00004.safetensors",
|
210 |
+
"transformer.h.29.ln_2.weight": "model-00003-of-00004.safetensors",
|
211 |
+
"transformer.h.29.mlp.c_fc_0.weight": "model-00003-of-00004.safetensors",
|
212 |
+
"transformer.h.29.mlp.c_fc_1.weight": "model-00003-of-00004.safetensors",
|
213 |
+
"transformer.h.29.mlp.c_proj.weight": "model-00003-of-00004.safetensors",
|
214 |
+
"transformer.h.3.attn.attention.k_proj.weight": "model-00001-of-00004.safetensors",
|
215 |
+
"transformer.h.3.attn.attention.out_proj.weight": "model-00001-of-00004.safetensors",
|
216 |
+
"transformer.h.3.attn.attention.q_proj.weight": "model-00001-of-00004.safetensors",
|
217 |
+
"transformer.h.3.attn.attention.v_proj.weight": "model-00001-of-00004.safetensors",
|
218 |
+
"transformer.h.3.ln_1.weight": "model-00001-of-00004.safetensors",
|
219 |
+
"transformer.h.3.ln_2.weight": "model-00001-of-00004.safetensors",
|
220 |
+
"transformer.h.3.mlp.c_fc_0.weight": "model-00001-of-00004.safetensors",
|
221 |
+
"transformer.h.3.mlp.c_fc_1.weight": "model-00001-of-00004.safetensors",
|
222 |
+
"transformer.h.3.mlp.c_proj.weight": "model-00001-of-00004.safetensors",
|
223 |
+
"transformer.h.30.attn.attention.k_proj.weight": "model-00003-of-00004.safetensors",
|
224 |
+
"transformer.h.30.attn.attention.out_proj.weight": "model-00003-of-00004.safetensors",
|
225 |
+
"transformer.h.30.attn.attention.q_proj.weight": "model-00003-of-00004.safetensors",
|
226 |
+
"transformer.h.30.attn.attention.v_proj.weight": "model-00003-of-00004.safetensors",
|
227 |
+
"transformer.h.30.ln_1.weight": "model-00003-of-00004.safetensors",
|
228 |
+
"transformer.h.30.ln_2.weight": "model-00003-of-00004.safetensors",
|
229 |
+
"transformer.h.30.mlp.c_fc_0.weight": "model-00003-of-00004.safetensors",
|
230 |
+
"transformer.h.30.mlp.c_fc_1.weight": "model-00003-of-00004.safetensors",
|
231 |
+
"transformer.h.30.mlp.c_proj.weight": "model-00003-of-00004.safetensors",
|
232 |
+
"transformer.h.31.attn.attention.k_proj.weight": "model-00003-of-00004.safetensors",
|
233 |
+
"transformer.h.31.attn.attention.out_proj.weight": "model-00003-of-00004.safetensors",
|
234 |
+
"transformer.h.31.attn.attention.q_proj.weight": "model-00003-of-00004.safetensors",
|
235 |
+
"transformer.h.31.attn.attention.v_proj.weight": "model-00003-of-00004.safetensors",
|
236 |
+
"transformer.h.31.ln_1.weight": "model-00003-of-00004.safetensors",
|
237 |
+
"transformer.h.31.ln_2.weight": "model-00003-of-00004.safetensors",
|
238 |
+
"transformer.h.31.mlp.c_fc_0.weight": "model-00003-of-00004.safetensors",
|
239 |
+
"transformer.h.31.mlp.c_fc_1.weight": "model-00003-of-00004.safetensors",
|
240 |
+
"transformer.h.31.mlp.c_proj.weight": "model-00003-of-00004.safetensors",
|
241 |
+
"transformer.h.4.attn.attention.k_proj.weight": "model-00001-of-00004.safetensors",
|
242 |
+
"transformer.h.4.attn.attention.out_proj.weight": "model-00001-of-00004.safetensors",
|
243 |
+
"transformer.h.4.attn.attention.q_proj.weight": "model-00001-of-00004.safetensors",
|
244 |
+
"transformer.h.4.attn.attention.v_proj.weight": "model-00001-of-00004.safetensors",
|
245 |
+
"transformer.h.4.ln_1.weight": "model-00001-of-00004.safetensors",
|
246 |
+
"transformer.h.4.ln_2.weight": "model-00001-of-00004.safetensors",
|
247 |
+
"transformer.h.4.mlp.c_fc_0.weight": "model-00001-of-00004.safetensors",
|
248 |
+
"transformer.h.4.mlp.c_fc_1.weight": "model-00001-of-00004.safetensors",
|
249 |
+
"transformer.h.4.mlp.c_proj.weight": "model-00001-of-00004.safetensors",
|
250 |
+
"transformer.h.5.attn.attention.k_proj.weight": "model-00001-of-00004.safetensors",
|
251 |
+
"transformer.h.5.attn.attention.out_proj.weight": "model-00001-of-00004.safetensors",
|
252 |
+
"transformer.h.5.attn.attention.q_proj.weight": "model-00001-of-00004.safetensors",
|
253 |
+
"transformer.h.5.attn.attention.v_proj.weight": "model-00001-of-00004.safetensors",
|
254 |
+
"transformer.h.5.ln_1.weight": "model-00001-of-00004.safetensors",
|
255 |
+
"transformer.h.5.ln_2.weight": "model-00001-of-00004.safetensors",
|
256 |
+
"transformer.h.5.mlp.c_fc_0.weight": "model-00001-of-00004.safetensors",
|
257 |
+
"transformer.h.5.mlp.c_fc_1.weight": "model-00001-of-00004.safetensors",
|
258 |
+
"transformer.h.5.mlp.c_proj.weight": "model-00001-of-00004.safetensors",
|
259 |
+
"transformer.h.6.attn.attention.k_proj.weight": "model-00001-of-00004.safetensors",
|
260 |
+
"transformer.h.6.attn.attention.out_proj.weight": "model-00001-of-00004.safetensors",
|
261 |
+
"transformer.h.6.attn.attention.q_proj.weight": "model-00001-of-00004.safetensors",
|
262 |
+
"transformer.h.6.attn.attention.v_proj.weight": "model-00001-of-00004.safetensors",
|
263 |
+
"transformer.h.6.ln_1.weight": "model-00001-of-00004.safetensors",
|
264 |
+
"transformer.h.6.ln_2.weight": "model-00001-of-00004.safetensors",
|
265 |
+
"transformer.h.6.mlp.c_fc_0.weight": "model-00001-of-00004.safetensors",
|
266 |
+
"transformer.h.6.mlp.c_fc_1.weight": "model-00001-of-00004.safetensors",
|
267 |
+
"transformer.h.6.mlp.c_proj.weight": "model-00001-of-00004.safetensors",
|
268 |
+
"transformer.h.7.attn.attention.k_proj.weight": "model-00001-of-00004.safetensors",
|
269 |
+
"transformer.h.7.attn.attention.out_proj.weight": "model-00001-of-00004.safetensors",
|
270 |
+
"transformer.h.7.attn.attention.q_proj.weight": "model-00001-of-00004.safetensors",
|
271 |
+
"transformer.h.7.attn.attention.v_proj.weight": "model-00001-of-00004.safetensors",
|
272 |
+
"transformer.h.7.ln_1.weight": "model-00001-of-00004.safetensors",
|
273 |
+
"transformer.h.7.ln_2.weight": "model-00001-of-00004.safetensors",
|
274 |
+
"transformer.h.7.mlp.c_fc_0.weight": "model-00001-of-00004.safetensors",
|
275 |
+
"transformer.h.7.mlp.c_fc_1.weight": "model-00001-of-00004.safetensors",
|
276 |
+
"transformer.h.7.mlp.c_proj.weight": "model-00001-of-00004.safetensors",
|
277 |
+
"transformer.h.8.attn.attention.k_proj.weight": "model-00001-of-00004.safetensors",
|
278 |
+
"transformer.h.8.attn.attention.out_proj.weight": "model-00001-of-00004.safetensors",
|
279 |
+
"transformer.h.8.attn.attention.q_proj.weight": "model-00001-of-00004.safetensors",
|
280 |
+
"transformer.h.8.attn.attention.v_proj.weight": "model-00001-of-00004.safetensors",
|
281 |
+
"transformer.h.8.ln_1.weight": "model-00001-of-00004.safetensors",
|
282 |
+
"transformer.h.8.ln_2.weight": "model-00001-of-00004.safetensors",
|
283 |
+
"transformer.h.8.mlp.c_fc_0.weight": "model-00001-of-00004.safetensors",
|
284 |
+
"transformer.h.8.mlp.c_fc_1.weight": "model-00001-of-00004.safetensors",
|
285 |
+
"transformer.h.8.mlp.c_proj.weight": "model-00001-of-00004.safetensors",
|
286 |
+
"transformer.h.9.attn.attention.k_proj.weight": "model-00001-of-00004.safetensors",
|
287 |
+
"transformer.h.9.attn.attention.out_proj.weight": "model-00001-of-00004.safetensors",
|
288 |
+
"transformer.h.9.attn.attention.q_proj.weight": "model-00001-of-00004.safetensors",
|
289 |
+
"transformer.h.9.attn.attention.v_proj.weight": "model-00001-of-00004.safetensors",
|
290 |
+
"transformer.h.9.ln_1.weight": "model-00001-of-00004.safetensors",
|
291 |
+
"transformer.h.9.ln_2.weight": "model-00001-of-00004.safetensors",
|
292 |
+
"transformer.h.9.mlp.c_fc_0.weight": "model-00001-of-00004.safetensors",
|
293 |
+
"transformer.h.9.mlp.c_fc_1.weight": "model-00002-of-00004.safetensors",
|
294 |
+
"transformer.h.9.mlp.c_proj.weight": "model-00002-of-00004.safetensors",
|
295 |
+
"transformer.ln_f.weight": "model-00003-of-00004.safetensors",
|
296 |
+
"transformer.wte.weight": "model-00001-of-00004.safetensors"
|
297 |
+
}
|
298 |
+
}
|