karthik commited on
Commit
7cf9ce7
·
1 Parent(s): 8fd12e8
config.json ADDED
@@ -0,0 +1,29 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "meta-llama/Llama-2-7b-chat-hf",
3
+ "architectures": [
4
+ "LLMForSequenceRegression"
5
+ ],
6
+ "attention_bias": false,
7
+ "attention_dropout": 0.0,
8
+ "bos_token_id": 1,
9
+ "eos_token_id": 2,
10
+ "hidden_act": "silu",
11
+ "hidden_size": 4096,
12
+ "initializer_range": 0.02,
13
+ "intermediate_size": 11008,
14
+ "max_position_embeddings": 4096,
15
+ "model_type": "llama",
16
+ "n_embd": 4096,
17
+ "num_attention_heads": 32,
18
+ "num_hidden_layers": 32,
19
+ "num_key_value_heads": 32,
20
+ "pretraining_tp": 1,
21
+ "rms_norm_eps": 1e-05,
22
+ "rope_scaling": null,
23
+ "rope_theta": 10000.0,
24
+ "tie_word_embeddings": false,
25
+ "torch_dtype": "float32",
26
+ "transformers_version": "4.37.2",
27
+ "use_cache": true,
28
+ "vocab_size": 32000
29
+ }
model-00001-of-00011.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2dac400809240982a42cfe5fd8ff3011bddb83cb68edcec937cc334b709e994d
3
+ size 4840396112
model-00002-of-00011.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a9d271d54e64357c5cda1b512329b7d83d09231d801d7c61b068928302763c8f
3
+ size 4857206528
model-00003-of-00011.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:20829829cbc11a0688e5c3ad12477b86edeed2a368168190a785c76e1652548e
3
+ size 4857206576
model-00004-of-00011.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9f9a8b3d5033b5ae8351ca0eb17ea08ff75d2c9aa71b912f9e8583450239da8a
3
+ size 4857206576
model-00005-of-00011.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:561fdf50dbe80945f61d46284c4a69772d4f2dfb87cdbd13fe49c6d38d745e73
3
+ size 4857206576
model-00006-of-00011.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:349b71736582c2dfe67529ade4998f3f2a38a54b5e6657f2427a5587c8456025
3
+ size 4932687352
model-00007-of-00011.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:aa9e20b57b5e893119a4aa65bcc8fb78c726e61b515a54447d3f88b2946af92a
3
+ size 4857207160
model-00008-of-00011.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:51a80303fcd9b7f17c34e1b6240ccc9ed05640ec41d086e60d4d28da6c555b30
3
+ size 4857207208
model-00009-of-00011.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:87d88e61709f77bffe50eda0d51a56e9e701785f3daa3afd335a20840dc00686
3
+ size 4857207216
model-00010-of-00011.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:70d4342b39dd198b364a030910ac15bdd50cc44a72af47b580efbcd3c1e27501
3
+ size 4857207216
model-00011-of-00011.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6c5cdeec3afbe1446dd4ca223888a9479150098cdaaffe99b3d376dea55d00ba
3
+ size 4228093840
model.safetensors.index.json ADDED
@@ -0,0 +1,588 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metadata": {
3
+ "total_size": 52858765312
4
+ },
5
+ "weight_map": {
6
+ "embed_tokens.weight": "model-00001-of-00011.safetensors",
7
+ "layers.0.input_layernorm.weight": "model-00001-of-00011.safetensors",
8
+ "layers.0.mlp.down_proj.weight": "model-00001-of-00011.safetensors",
9
+ "layers.0.mlp.gate_proj.weight": "model-00001-of-00011.safetensors",
10
+ "layers.0.mlp.up_proj.weight": "model-00001-of-00011.safetensors",
11
+ "layers.0.post_attention_layernorm.weight": "model-00001-of-00011.safetensors",
12
+ "layers.0.self_attn.k_proj.weight": "model-00001-of-00011.safetensors",
13
+ "layers.0.self_attn.o_proj.weight": "model-00001-of-00011.safetensors",
14
+ "layers.0.self_attn.q_proj.weight": "model-00001-of-00011.safetensors",
15
+ "layers.0.self_attn.v_proj.weight": "model-00001-of-00011.safetensors",
16
+ "layers.1.input_layernorm.weight": "model-00001-of-00011.safetensors",
17
+ "layers.1.mlp.down_proj.weight": "model-00001-of-00011.safetensors",
18
+ "layers.1.mlp.gate_proj.weight": "model-00001-of-00011.safetensors",
19
+ "layers.1.mlp.up_proj.weight": "model-00001-of-00011.safetensors",
20
+ "layers.1.post_attention_layernorm.weight": "model-00001-of-00011.safetensors",
21
+ "layers.1.self_attn.k_proj.weight": "model-00001-of-00011.safetensors",
22
+ "layers.1.self_attn.o_proj.weight": "model-00001-of-00011.safetensors",
23
+ "layers.1.self_attn.q_proj.weight": "model-00001-of-00011.safetensors",
24
+ "layers.1.self_attn.v_proj.weight": "model-00001-of-00011.safetensors",
25
+ "layers.10.input_layernorm.weight": "model-00002-of-00011.safetensors",
26
+ "layers.10.mlp.down_proj.weight": "model-00002-of-00011.safetensors",
27
+ "layers.10.mlp.gate_proj.weight": "model-00002-of-00011.safetensors",
28
+ "layers.10.mlp.up_proj.weight": "model-00002-of-00011.safetensors",
29
+ "layers.10.post_attention_layernorm.weight": "model-00002-of-00011.safetensors",
30
+ "layers.10.self_attn.k_proj.weight": "model-00002-of-00011.safetensors",
31
+ "layers.10.self_attn.o_proj.weight": "model-00002-of-00011.safetensors",
32
+ "layers.10.self_attn.q_proj.weight": "model-00002-of-00011.safetensors",
33
+ "layers.10.self_attn.v_proj.weight": "model-00002-of-00011.safetensors",
34
+ "layers.11.input_layernorm.weight": "model-00003-of-00011.safetensors",
35
+ "layers.11.mlp.down_proj.weight": "model-00003-of-00011.safetensors",
36
+ "layers.11.mlp.gate_proj.weight": "model-00003-of-00011.safetensors",
37
+ "layers.11.mlp.up_proj.weight": "model-00003-of-00011.safetensors",
38
+ "layers.11.post_attention_layernorm.weight": "model-00003-of-00011.safetensors",
39
+ "layers.11.self_attn.k_proj.weight": "model-00002-of-00011.safetensors",
40
+ "layers.11.self_attn.o_proj.weight": "model-00002-of-00011.safetensors",
41
+ "layers.11.self_attn.q_proj.weight": "model-00002-of-00011.safetensors",
42
+ "layers.11.self_attn.v_proj.weight": "model-00002-of-00011.safetensors",
43
+ "layers.12.input_layernorm.weight": "model-00003-of-00011.safetensors",
44
+ "layers.12.mlp.down_proj.weight": "model-00003-of-00011.safetensors",
45
+ "layers.12.mlp.gate_proj.weight": "model-00003-of-00011.safetensors",
46
+ "layers.12.mlp.up_proj.weight": "model-00003-of-00011.safetensors",
47
+ "layers.12.post_attention_layernorm.weight": "model-00003-of-00011.safetensors",
48
+ "layers.12.self_attn.k_proj.weight": "model-00003-of-00011.safetensors",
49
+ "layers.12.self_attn.o_proj.weight": "model-00003-of-00011.safetensors",
50
+ "layers.12.self_attn.q_proj.weight": "model-00003-of-00011.safetensors",
51
+ "layers.12.self_attn.v_proj.weight": "model-00003-of-00011.safetensors",
52
+ "layers.13.input_layernorm.weight": "model-00003-of-00011.safetensors",
53
+ "layers.13.mlp.down_proj.weight": "model-00003-of-00011.safetensors",
54
+ "layers.13.mlp.gate_proj.weight": "model-00003-of-00011.safetensors",
55
+ "layers.13.mlp.up_proj.weight": "model-00003-of-00011.safetensors",
56
+ "layers.13.post_attention_layernorm.weight": "model-00003-of-00011.safetensors",
57
+ "layers.13.self_attn.k_proj.weight": "model-00003-of-00011.safetensors",
58
+ "layers.13.self_attn.o_proj.weight": "model-00003-of-00011.safetensors",
59
+ "layers.13.self_attn.q_proj.weight": "model-00003-of-00011.safetensors",
60
+ "layers.13.self_attn.v_proj.weight": "model-00003-of-00011.safetensors",
61
+ "layers.14.input_layernorm.weight": "model-00003-of-00011.safetensors",
62
+ "layers.14.mlp.down_proj.weight": "model-00003-of-00011.safetensors",
63
+ "layers.14.mlp.gate_proj.weight": "model-00003-of-00011.safetensors",
64
+ "layers.14.mlp.up_proj.weight": "model-00003-of-00011.safetensors",
65
+ "layers.14.post_attention_layernorm.weight": "model-00003-of-00011.safetensors",
66
+ "layers.14.self_attn.k_proj.weight": "model-00003-of-00011.safetensors",
67
+ "layers.14.self_attn.o_proj.weight": "model-00003-of-00011.safetensors",
68
+ "layers.14.self_attn.q_proj.weight": "model-00003-of-00011.safetensors",
69
+ "layers.14.self_attn.v_proj.weight": "model-00003-of-00011.safetensors",
70
+ "layers.15.input_layernorm.weight": "model-00003-of-00011.safetensors",
71
+ "layers.15.mlp.down_proj.weight": "model-00003-of-00011.safetensors",
72
+ "layers.15.mlp.gate_proj.weight": "model-00003-of-00011.safetensors",
73
+ "layers.15.mlp.up_proj.weight": "model-00003-of-00011.safetensors",
74
+ "layers.15.post_attention_layernorm.weight": "model-00003-of-00011.safetensors",
75
+ "layers.15.self_attn.k_proj.weight": "model-00003-of-00011.safetensors",
76
+ "layers.15.self_attn.o_proj.weight": "model-00003-of-00011.safetensors",
77
+ "layers.15.self_attn.q_proj.weight": "model-00003-of-00011.safetensors",
78
+ "layers.15.self_attn.v_proj.weight": "model-00003-of-00011.safetensors",
79
+ "layers.16.input_layernorm.weight": "model-00003-of-00011.safetensors",
80
+ "layers.16.mlp.down_proj.weight": "model-00003-of-00011.safetensors",
81
+ "layers.16.mlp.gate_proj.weight": "model-00003-of-00011.safetensors",
82
+ "layers.16.mlp.up_proj.weight": "model-00003-of-00011.safetensors",
83
+ "layers.16.post_attention_layernorm.weight": "model-00003-of-00011.safetensors",
84
+ "layers.16.self_attn.k_proj.weight": "model-00003-of-00011.safetensors",
85
+ "layers.16.self_attn.o_proj.weight": "model-00003-of-00011.safetensors",
86
+ "layers.16.self_attn.q_proj.weight": "model-00003-of-00011.safetensors",
87
+ "layers.16.self_attn.v_proj.weight": "model-00003-of-00011.safetensors",
88
+ "layers.17.input_layernorm.weight": "model-00004-of-00011.safetensors",
89
+ "layers.17.mlp.down_proj.weight": "model-00004-of-00011.safetensors",
90
+ "layers.17.mlp.gate_proj.weight": "model-00004-of-00011.safetensors",
91
+ "layers.17.mlp.up_proj.weight": "model-00004-of-00011.safetensors",
92
+ "layers.17.post_attention_layernorm.weight": "model-00004-of-00011.safetensors",
93
+ "layers.17.self_attn.k_proj.weight": "model-00003-of-00011.safetensors",
94
+ "layers.17.self_attn.o_proj.weight": "model-00003-of-00011.safetensors",
95
+ "layers.17.self_attn.q_proj.weight": "model-00003-of-00011.safetensors",
96
+ "layers.17.self_attn.v_proj.weight": "model-00003-of-00011.safetensors",
97
+ "layers.18.input_layernorm.weight": "model-00004-of-00011.safetensors",
98
+ "layers.18.mlp.down_proj.weight": "model-00004-of-00011.safetensors",
99
+ "layers.18.mlp.gate_proj.weight": "model-00004-of-00011.safetensors",
100
+ "layers.18.mlp.up_proj.weight": "model-00004-of-00011.safetensors",
101
+ "layers.18.post_attention_layernorm.weight": "model-00004-of-00011.safetensors",
102
+ "layers.18.self_attn.k_proj.weight": "model-00004-of-00011.safetensors",
103
+ "layers.18.self_attn.o_proj.weight": "model-00004-of-00011.safetensors",
104
+ "layers.18.self_attn.q_proj.weight": "model-00004-of-00011.safetensors",
105
+ "layers.18.self_attn.v_proj.weight": "model-00004-of-00011.safetensors",
106
+ "layers.19.input_layernorm.weight": "model-00004-of-00011.safetensors",
107
+ "layers.19.mlp.down_proj.weight": "model-00004-of-00011.safetensors",
108
+ "layers.19.mlp.gate_proj.weight": "model-00004-of-00011.safetensors",
109
+ "layers.19.mlp.up_proj.weight": "model-00004-of-00011.safetensors",
110
+ "layers.19.post_attention_layernorm.weight": "model-00004-of-00011.safetensors",
111
+ "layers.19.self_attn.k_proj.weight": "model-00004-of-00011.safetensors",
112
+ "layers.19.self_attn.o_proj.weight": "model-00004-of-00011.safetensors",
113
+ "layers.19.self_attn.q_proj.weight": "model-00004-of-00011.safetensors",
114
+ "layers.19.self_attn.v_proj.weight": "model-00004-of-00011.safetensors",
115
+ "layers.2.input_layernorm.weight": "model-00001-of-00011.safetensors",
116
+ "layers.2.mlp.down_proj.weight": "model-00001-of-00011.safetensors",
117
+ "layers.2.mlp.gate_proj.weight": "model-00001-of-00011.safetensors",
118
+ "layers.2.mlp.up_proj.weight": "model-00001-of-00011.safetensors",
119
+ "layers.2.post_attention_layernorm.weight": "model-00001-of-00011.safetensors",
120
+ "layers.2.self_attn.k_proj.weight": "model-00001-of-00011.safetensors",
121
+ "layers.2.self_attn.o_proj.weight": "model-00001-of-00011.safetensors",
122
+ "layers.2.self_attn.q_proj.weight": "model-00001-of-00011.safetensors",
123
+ "layers.2.self_attn.v_proj.weight": "model-00001-of-00011.safetensors",
124
+ "layers.20.input_layernorm.weight": "model-00004-of-00011.safetensors",
125
+ "layers.20.mlp.down_proj.weight": "model-00004-of-00011.safetensors",
126
+ "layers.20.mlp.gate_proj.weight": "model-00004-of-00011.safetensors",
127
+ "layers.20.mlp.up_proj.weight": "model-00004-of-00011.safetensors",
128
+ "layers.20.post_attention_layernorm.weight": "model-00004-of-00011.safetensors",
129
+ "layers.20.self_attn.k_proj.weight": "model-00004-of-00011.safetensors",
130
+ "layers.20.self_attn.o_proj.weight": "model-00004-of-00011.safetensors",
131
+ "layers.20.self_attn.q_proj.weight": "model-00004-of-00011.safetensors",
132
+ "layers.20.self_attn.v_proj.weight": "model-00004-of-00011.safetensors",
133
+ "layers.21.input_layernorm.weight": "model-00004-of-00011.safetensors",
134
+ "layers.21.mlp.down_proj.weight": "model-00004-of-00011.safetensors",
135
+ "layers.21.mlp.gate_proj.weight": "model-00004-of-00011.safetensors",
136
+ "layers.21.mlp.up_proj.weight": "model-00004-of-00011.safetensors",
137
+ "layers.21.post_attention_layernorm.weight": "model-00004-of-00011.safetensors",
138
+ "layers.21.self_attn.k_proj.weight": "model-00004-of-00011.safetensors",
139
+ "layers.21.self_attn.o_proj.weight": "model-00004-of-00011.safetensors",
140
+ "layers.21.self_attn.q_proj.weight": "model-00004-of-00011.safetensors",
141
+ "layers.21.self_attn.v_proj.weight": "model-00004-of-00011.safetensors",
142
+ "layers.22.input_layernorm.weight": "model-00004-of-00011.safetensors",
143
+ "layers.22.mlp.down_proj.weight": "model-00004-of-00011.safetensors",
144
+ "layers.22.mlp.gate_proj.weight": "model-00004-of-00011.safetensors",
145
+ "layers.22.mlp.up_proj.weight": "model-00004-of-00011.safetensors",
146
+ "layers.22.post_attention_layernorm.weight": "model-00004-of-00011.safetensors",
147
+ "layers.22.self_attn.k_proj.weight": "model-00004-of-00011.safetensors",
148
+ "layers.22.self_attn.o_proj.weight": "model-00004-of-00011.safetensors",
149
+ "layers.22.self_attn.q_proj.weight": "model-00004-of-00011.safetensors",
150
+ "layers.22.self_attn.v_proj.weight": "model-00004-of-00011.safetensors",
151
+ "layers.23.input_layernorm.weight": "model-00005-of-00011.safetensors",
152
+ "layers.23.mlp.down_proj.weight": "model-00005-of-00011.safetensors",
153
+ "layers.23.mlp.gate_proj.weight": "model-00005-of-00011.safetensors",
154
+ "layers.23.mlp.up_proj.weight": "model-00005-of-00011.safetensors",
155
+ "layers.23.post_attention_layernorm.weight": "model-00005-of-00011.safetensors",
156
+ "layers.23.self_attn.k_proj.weight": "model-00004-of-00011.safetensors",
157
+ "layers.23.self_attn.o_proj.weight": "model-00004-of-00011.safetensors",
158
+ "layers.23.self_attn.q_proj.weight": "model-00004-of-00011.safetensors",
159
+ "layers.23.self_attn.v_proj.weight": "model-00004-of-00011.safetensors",
160
+ "layers.24.input_layernorm.weight": "model-00005-of-00011.safetensors",
161
+ "layers.24.mlp.down_proj.weight": "model-00005-of-00011.safetensors",
162
+ "layers.24.mlp.gate_proj.weight": "model-00005-of-00011.safetensors",
163
+ "layers.24.mlp.up_proj.weight": "model-00005-of-00011.safetensors",
164
+ "layers.24.post_attention_layernorm.weight": "model-00005-of-00011.safetensors",
165
+ "layers.24.self_attn.k_proj.weight": "model-00005-of-00011.safetensors",
166
+ "layers.24.self_attn.o_proj.weight": "model-00005-of-00011.safetensors",
167
+ "layers.24.self_attn.q_proj.weight": "model-00005-of-00011.safetensors",
168
+ "layers.24.self_attn.v_proj.weight": "model-00005-of-00011.safetensors",
169
+ "layers.25.input_layernorm.weight": "model-00005-of-00011.safetensors",
170
+ "layers.25.mlp.down_proj.weight": "model-00005-of-00011.safetensors",
171
+ "layers.25.mlp.gate_proj.weight": "model-00005-of-00011.safetensors",
172
+ "layers.25.mlp.up_proj.weight": "model-00005-of-00011.safetensors",
173
+ "layers.25.post_attention_layernorm.weight": "model-00005-of-00011.safetensors",
174
+ "layers.25.self_attn.k_proj.weight": "model-00005-of-00011.safetensors",
175
+ "layers.25.self_attn.o_proj.weight": "model-00005-of-00011.safetensors",
176
+ "layers.25.self_attn.q_proj.weight": "model-00005-of-00011.safetensors",
177
+ "layers.25.self_attn.v_proj.weight": "model-00005-of-00011.safetensors",
178
+ "layers.26.input_layernorm.weight": "model-00005-of-00011.safetensors",
179
+ "layers.26.mlp.down_proj.weight": "model-00005-of-00011.safetensors",
180
+ "layers.26.mlp.gate_proj.weight": "model-00005-of-00011.safetensors",
181
+ "layers.26.mlp.up_proj.weight": "model-00005-of-00011.safetensors",
182
+ "layers.26.post_attention_layernorm.weight": "model-00005-of-00011.safetensors",
183
+ "layers.26.self_attn.k_proj.weight": "model-00005-of-00011.safetensors",
184
+ "layers.26.self_attn.o_proj.weight": "model-00005-of-00011.safetensors",
185
+ "layers.26.self_attn.q_proj.weight": "model-00005-of-00011.safetensors",
186
+ "layers.26.self_attn.v_proj.weight": "model-00005-of-00011.safetensors",
187
+ "layers.27.input_layernorm.weight": "model-00005-of-00011.safetensors",
188
+ "layers.27.mlp.down_proj.weight": "model-00005-of-00011.safetensors",
189
+ "layers.27.mlp.gate_proj.weight": "model-00005-of-00011.safetensors",
190
+ "layers.27.mlp.up_proj.weight": "model-00005-of-00011.safetensors",
191
+ "layers.27.post_attention_layernorm.weight": "model-00005-of-00011.safetensors",
192
+ "layers.27.self_attn.k_proj.weight": "model-00005-of-00011.safetensors",
193
+ "layers.27.self_attn.o_proj.weight": "model-00005-of-00011.safetensors",
194
+ "layers.27.self_attn.q_proj.weight": "model-00005-of-00011.safetensors",
195
+ "layers.27.self_attn.v_proj.weight": "model-00005-of-00011.safetensors",
196
+ "layers.28.input_layernorm.weight": "model-00005-of-00011.safetensors",
197
+ "layers.28.mlp.down_proj.weight": "model-00005-of-00011.safetensors",
198
+ "layers.28.mlp.gate_proj.weight": "model-00005-of-00011.safetensors",
199
+ "layers.28.mlp.up_proj.weight": "model-00005-of-00011.safetensors",
200
+ "layers.28.post_attention_layernorm.weight": "model-00005-of-00011.safetensors",
201
+ "layers.28.self_attn.k_proj.weight": "model-00005-of-00011.safetensors",
202
+ "layers.28.self_attn.o_proj.weight": "model-00005-of-00011.safetensors",
203
+ "layers.28.self_attn.q_proj.weight": "model-00005-of-00011.safetensors",
204
+ "layers.28.self_attn.v_proj.weight": "model-00005-of-00011.safetensors",
205
+ "layers.29.input_layernorm.weight": "model-00006-of-00011.safetensors",
206
+ "layers.29.mlp.down_proj.weight": "model-00006-of-00011.safetensors",
207
+ "layers.29.mlp.gate_proj.weight": "model-00006-of-00011.safetensors",
208
+ "layers.29.mlp.up_proj.weight": "model-00006-of-00011.safetensors",
209
+ "layers.29.post_attention_layernorm.weight": "model-00006-of-00011.safetensors",
210
+ "layers.29.self_attn.k_proj.weight": "model-00005-of-00011.safetensors",
211
+ "layers.29.self_attn.o_proj.weight": "model-00005-of-00011.safetensors",
212
+ "layers.29.self_attn.q_proj.weight": "model-00005-of-00011.safetensors",
213
+ "layers.29.self_attn.v_proj.weight": "model-00005-of-00011.safetensors",
214
+ "layers.3.input_layernorm.weight": "model-00001-of-00011.safetensors",
215
+ "layers.3.mlp.down_proj.weight": "model-00001-of-00011.safetensors",
216
+ "layers.3.mlp.gate_proj.weight": "model-00001-of-00011.safetensors",
217
+ "layers.3.mlp.up_proj.weight": "model-00001-of-00011.safetensors",
218
+ "layers.3.post_attention_layernorm.weight": "model-00001-of-00011.safetensors",
219
+ "layers.3.self_attn.k_proj.weight": "model-00001-of-00011.safetensors",
220
+ "layers.3.self_attn.o_proj.weight": "model-00001-of-00011.safetensors",
221
+ "layers.3.self_attn.q_proj.weight": "model-00001-of-00011.safetensors",
222
+ "layers.3.self_attn.v_proj.weight": "model-00001-of-00011.safetensors",
223
+ "layers.30.input_layernorm.weight": "model-00006-of-00011.safetensors",
224
+ "layers.30.mlp.down_proj.weight": "model-00006-of-00011.safetensors",
225
+ "layers.30.mlp.gate_proj.weight": "model-00006-of-00011.safetensors",
226
+ "layers.30.mlp.up_proj.weight": "model-00006-of-00011.safetensors",
227
+ "layers.30.post_attention_layernorm.weight": "model-00006-of-00011.safetensors",
228
+ "layers.30.self_attn.k_proj.weight": "model-00006-of-00011.safetensors",
229
+ "layers.30.self_attn.o_proj.weight": "model-00006-of-00011.safetensors",
230
+ "layers.30.self_attn.q_proj.weight": "model-00006-of-00011.safetensors",
231
+ "layers.30.self_attn.v_proj.weight": "model-00006-of-00011.safetensors",
232
+ "layers.31.input_layernorm.weight": "model-00006-of-00011.safetensors",
233
+ "layers.31.mlp.down_proj.weight": "model-00006-of-00011.safetensors",
234
+ "layers.31.mlp.gate_proj.weight": "model-00006-of-00011.safetensors",
235
+ "layers.31.mlp.up_proj.weight": "model-00006-of-00011.safetensors",
236
+ "layers.31.post_attention_layernorm.weight": "model-00006-of-00011.safetensors",
237
+ "layers.31.self_attn.k_proj.weight": "model-00006-of-00011.safetensors",
238
+ "layers.31.self_attn.o_proj.weight": "model-00006-of-00011.safetensors",
239
+ "layers.31.self_attn.q_proj.weight": "model-00006-of-00011.safetensors",
240
+ "layers.31.self_attn.v_proj.weight": "model-00006-of-00011.safetensors",
241
+ "layers.4.input_layernorm.weight": "model-00001-of-00011.safetensors",
242
+ "layers.4.mlp.down_proj.weight": "model-00001-of-00011.safetensors",
243
+ "layers.4.mlp.gate_proj.weight": "model-00001-of-00011.safetensors",
244
+ "layers.4.mlp.up_proj.weight": "model-00001-of-00011.safetensors",
245
+ "layers.4.post_attention_layernorm.weight": "model-00001-of-00011.safetensors",
246
+ "layers.4.self_attn.k_proj.weight": "model-00001-of-00011.safetensors",
247
+ "layers.4.self_attn.o_proj.weight": "model-00001-of-00011.safetensors",
248
+ "layers.4.self_attn.q_proj.weight": "model-00001-of-00011.safetensors",
249
+ "layers.4.self_attn.v_proj.weight": "model-00001-of-00011.safetensors",
250
+ "layers.5.input_layernorm.weight": "model-00002-of-00011.safetensors",
251
+ "layers.5.mlp.down_proj.weight": "model-00002-of-00011.safetensors",
252
+ "layers.5.mlp.gate_proj.weight": "model-00002-of-00011.safetensors",
253
+ "layers.5.mlp.up_proj.weight": "model-00002-of-00011.safetensors",
254
+ "layers.5.post_attention_layernorm.weight": "model-00002-of-00011.safetensors",
255
+ "layers.5.self_attn.k_proj.weight": "model-00001-of-00011.safetensors",
256
+ "layers.5.self_attn.o_proj.weight": "model-00001-of-00011.safetensors",
257
+ "layers.5.self_attn.q_proj.weight": "model-00001-of-00011.safetensors",
258
+ "layers.5.self_attn.v_proj.weight": "model-00001-of-00011.safetensors",
259
+ "layers.6.input_layernorm.weight": "model-00002-of-00011.safetensors",
260
+ "layers.6.mlp.down_proj.weight": "model-00002-of-00011.safetensors",
261
+ "layers.6.mlp.gate_proj.weight": "model-00002-of-00011.safetensors",
262
+ "layers.6.mlp.up_proj.weight": "model-00002-of-00011.safetensors",
263
+ "layers.6.post_attention_layernorm.weight": "model-00002-of-00011.safetensors",
264
+ "layers.6.self_attn.k_proj.weight": "model-00002-of-00011.safetensors",
265
+ "layers.6.self_attn.o_proj.weight": "model-00002-of-00011.safetensors",
266
+ "layers.6.self_attn.q_proj.weight": "model-00002-of-00011.safetensors",
267
+ "layers.6.self_attn.v_proj.weight": "model-00002-of-00011.safetensors",
268
+ "layers.7.input_layernorm.weight": "model-00002-of-00011.safetensors",
269
+ "layers.7.mlp.down_proj.weight": "model-00002-of-00011.safetensors",
270
+ "layers.7.mlp.gate_proj.weight": "model-00002-of-00011.safetensors",
271
+ "layers.7.mlp.up_proj.weight": "model-00002-of-00011.safetensors",
272
+ "layers.7.post_attention_layernorm.weight": "model-00002-of-00011.safetensors",
273
+ "layers.7.self_attn.k_proj.weight": "model-00002-of-00011.safetensors",
274
+ "layers.7.self_attn.o_proj.weight": "model-00002-of-00011.safetensors",
275
+ "layers.7.self_attn.q_proj.weight": "model-00002-of-00011.safetensors",
276
+ "layers.7.self_attn.v_proj.weight": "model-00002-of-00011.safetensors",
277
+ "layers.8.input_layernorm.weight": "model-00002-of-00011.safetensors",
278
+ "layers.8.mlp.down_proj.weight": "model-00002-of-00011.safetensors",
279
+ "layers.8.mlp.gate_proj.weight": "model-00002-of-00011.safetensors",
280
+ "layers.8.mlp.up_proj.weight": "model-00002-of-00011.safetensors",
281
+ "layers.8.post_attention_layernorm.weight": "model-00002-of-00011.safetensors",
282
+ "layers.8.self_attn.k_proj.weight": "model-00002-of-00011.safetensors",
283
+ "layers.8.self_attn.o_proj.weight": "model-00002-of-00011.safetensors",
284
+ "layers.8.self_attn.q_proj.weight": "model-00002-of-00011.safetensors",
285
+ "layers.8.self_attn.v_proj.weight": "model-00002-of-00011.safetensors",
286
+ "layers.9.input_layernorm.weight": "model-00002-of-00011.safetensors",
287
+ "layers.9.mlp.down_proj.weight": "model-00002-of-00011.safetensors",
288
+ "layers.9.mlp.gate_proj.weight": "model-00002-of-00011.safetensors",
289
+ "layers.9.mlp.up_proj.weight": "model-00002-of-00011.safetensors",
290
+ "layers.9.post_attention_layernorm.weight": "model-00002-of-00011.safetensors",
291
+ "layers.9.self_attn.k_proj.weight": "model-00002-of-00011.safetensors",
292
+ "layers.9.self_attn.o_proj.weight": "model-00002-of-00011.safetensors",
293
+ "layers.9.self_attn.q_proj.weight": "model-00002-of-00011.safetensors",
294
+ "layers.9.self_attn.v_proj.weight": "model-00002-of-00011.safetensors",
295
+ "norm.weight": "model-00006-of-00011.safetensors",
296
+ "transformer.embed_tokens.weight": "model-00006-of-00011.safetensors",
297
+ "transformer.layers.0.input_layernorm.weight": "model-00006-of-00011.safetensors",
298
+ "transformer.layers.0.mlp.down_proj.weight": "model-00006-of-00011.safetensors",
299
+ "transformer.layers.0.mlp.gate_proj.weight": "model-00006-of-00011.safetensors",
300
+ "transformer.layers.0.mlp.up_proj.weight": "model-00006-of-00011.safetensors",
301
+ "transformer.layers.0.post_attention_layernorm.weight": "model-00006-of-00011.safetensors",
302
+ "transformer.layers.0.self_attn.k_proj.weight": "model-00006-of-00011.safetensors",
303
+ "transformer.layers.0.self_attn.o_proj.weight": "model-00006-of-00011.safetensors",
304
+ "transformer.layers.0.self_attn.q_proj.weight": "model-00006-of-00011.safetensors",
305
+ "transformer.layers.0.self_attn.v_proj.weight": "model-00006-of-00011.safetensors",
306
+ "transformer.layers.1.input_layernorm.weight": "model-00006-of-00011.safetensors",
307
+ "transformer.layers.1.mlp.down_proj.weight": "model-00006-of-00011.safetensors",
308
+ "transformer.layers.1.mlp.gate_proj.weight": "model-00006-of-00011.safetensors",
309
+ "transformer.layers.1.mlp.up_proj.weight": "model-00006-of-00011.safetensors",
310
+ "transformer.layers.1.post_attention_layernorm.weight": "model-00006-of-00011.safetensors",
311
+ "transformer.layers.1.self_attn.k_proj.weight": "model-00006-of-00011.safetensors",
312
+ "transformer.layers.1.self_attn.o_proj.weight": "model-00006-of-00011.safetensors",
313
+ "transformer.layers.1.self_attn.q_proj.weight": "model-00006-of-00011.safetensors",
314
+ "transformer.layers.1.self_attn.v_proj.weight": "model-00006-of-00011.safetensors",
315
+ "transformer.layers.10.input_layernorm.weight": "model-00008-of-00011.safetensors",
316
+ "transformer.layers.10.mlp.down_proj.weight": "model-00008-of-00011.safetensors",
317
+ "transformer.layers.10.mlp.gate_proj.weight": "model-00008-of-00011.safetensors",
318
+ "transformer.layers.10.mlp.up_proj.weight": "model-00008-of-00011.safetensors",
319
+ "transformer.layers.10.post_attention_layernorm.weight": "model-00008-of-00011.safetensors",
320
+ "transformer.layers.10.self_attn.k_proj.weight": "model-00008-of-00011.safetensors",
321
+ "transformer.layers.10.self_attn.o_proj.weight": "model-00008-of-00011.safetensors",
322
+ "transformer.layers.10.self_attn.q_proj.weight": "model-00008-of-00011.safetensors",
323
+ "transformer.layers.10.self_attn.v_proj.weight": "model-00008-of-00011.safetensors",
324
+ "transformer.layers.11.input_layernorm.weight": "model-00008-of-00011.safetensors",
325
+ "transformer.layers.11.mlp.down_proj.weight": "model-00008-of-00011.safetensors",
326
+ "transformer.layers.11.mlp.gate_proj.weight": "model-00008-of-00011.safetensors",
327
+ "transformer.layers.11.mlp.up_proj.weight": "model-00008-of-00011.safetensors",
328
+ "transformer.layers.11.post_attention_layernorm.weight": "model-00008-of-00011.safetensors",
329
+ "transformer.layers.11.self_attn.k_proj.weight": "model-00008-of-00011.safetensors",
330
+ "transformer.layers.11.self_attn.o_proj.weight": "model-00008-of-00011.safetensors",
331
+ "transformer.layers.11.self_attn.q_proj.weight": "model-00008-of-00011.safetensors",
332
+ "transformer.layers.11.self_attn.v_proj.weight": "model-00008-of-00011.safetensors",
333
+ "transformer.layers.12.input_layernorm.weight": "model-00008-of-00011.safetensors",
334
+ "transformer.layers.12.mlp.down_proj.weight": "model-00008-of-00011.safetensors",
335
+ "transformer.layers.12.mlp.gate_proj.weight": "model-00008-of-00011.safetensors",
336
+ "transformer.layers.12.mlp.up_proj.weight": "model-00008-of-00011.safetensors",
337
+ "transformer.layers.12.post_attention_layernorm.weight": "model-00008-of-00011.safetensors",
338
+ "transformer.layers.12.self_attn.k_proj.weight": "model-00008-of-00011.safetensors",
339
+ "transformer.layers.12.self_attn.o_proj.weight": "model-00008-of-00011.safetensors",
340
+ "transformer.layers.12.self_attn.q_proj.weight": "model-00008-of-00011.safetensors",
341
+ "transformer.layers.12.self_attn.v_proj.weight": "model-00008-of-00011.safetensors",
342
+ "transformer.layers.13.input_layernorm.weight": "model-00008-of-00011.safetensors",
343
+ "transformer.layers.13.mlp.down_proj.weight": "model-00008-of-00011.safetensors",
344
+ "transformer.layers.13.mlp.gate_proj.weight": "model-00008-of-00011.safetensors",
345
+ "transformer.layers.13.mlp.up_proj.weight": "model-00008-of-00011.safetensors",
346
+ "transformer.layers.13.post_attention_layernorm.weight": "model-00008-of-00011.safetensors",
347
+ "transformer.layers.13.self_attn.k_proj.weight": "model-00008-of-00011.safetensors",
348
+ "transformer.layers.13.self_attn.o_proj.weight": "model-00008-of-00011.safetensors",
349
+ "transformer.layers.13.self_attn.q_proj.weight": "model-00008-of-00011.safetensors",
350
+ "transformer.layers.13.self_attn.v_proj.weight": "model-00008-of-00011.safetensors",
351
+ "transformer.layers.14.input_layernorm.weight": "model-00009-of-00011.safetensors",
352
+ "transformer.layers.14.mlp.down_proj.weight": "model-00009-of-00011.safetensors",
353
+ "transformer.layers.14.mlp.gate_proj.weight": "model-00008-of-00011.safetensors",
354
+ "transformer.layers.14.mlp.up_proj.weight": "model-00008-of-00011.safetensors",
355
+ "transformer.layers.14.post_attention_layernorm.weight": "model-00009-of-00011.safetensors",
356
+ "transformer.layers.14.self_attn.k_proj.weight": "model-00008-of-00011.safetensors",
357
+ "transformer.layers.14.self_attn.o_proj.weight": "model-00008-of-00011.safetensors",
358
+ "transformer.layers.14.self_attn.q_proj.weight": "model-00008-of-00011.safetensors",
359
+ "transformer.layers.14.self_attn.v_proj.weight": "model-00008-of-00011.safetensors",
360
+ "transformer.layers.15.input_layernorm.weight": "model-00009-of-00011.safetensors",
361
+ "transformer.layers.15.mlp.down_proj.weight": "model-00009-of-00011.safetensors",
362
+ "transformer.layers.15.mlp.gate_proj.weight": "model-00009-of-00011.safetensors",
363
+ "transformer.layers.15.mlp.up_proj.weight": "model-00009-of-00011.safetensors",
364
+ "transformer.layers.15.post_attention_layernorm.weight": "model-00009-of-00011.safetensors",
365
+ "transformer.layers.15.self_attn.k_proj.weight": "model-00009-of-00011.safetensors",
366
+ "transformer.layers.15.self_attn.o_proj.weight": "model-00009-of-00011.safetensors",
367
+ "transformer.layers.15.self_attn.q_proj.weight": "model-00009-of-00011.safetensors",
368
+ "transformer.layers.15.self_attn.v_proj.weight": "model-00009-of-00011.safetensors",
369
+ "transformer.layers.16.input_layernorm.weight": "model-00009-of-00011.safetensors",
370
+ "transformer.layers.16.mlp.down_proj.weight": "model-00009-of-00011.safetensors",
371
+ "transformer.layers.16.mlp.gate_proj.weight": "model-00009-of-00011.safetensors",
372
+ "transformer.layers.16.mlp.up_proj.weight": "model-00009-of-00011.safetensors",
373
+ "transformer.layers.16.post_attention_layernorm.weight": "model-00009-of-00011.safetensors",
374
+ "transformer.layers.16.self_attn.k_proj.weight": "model-00009-of-00011.safetensors",
375
+ "transformer.layers.16.self_attn.o_proj.weight": "model-00009-of-00011.safetensors",
376
+ "transformer.layers.16.self_attn.q_proj.weight": "model-00009-of-00011.safetensors",
377
+ "transformer.layers.16.self_attn.v_proj.weight": "model-00009-of-00011.safetensors",
378
+ "transformer.layers.17.input_layernorm.weight": "model-00009-of-00011.safetensors",
379
+ "transformer.layers.17.mlp.down_proj.weight": "model-00009-of-00011.safetensors",
380
+ "transformer.layers.17.mlp.gate_proj.weight": "model-00009-of-00011.safetensors",
381
+ "transformer.layers.17.mlp.up_proj.weight": "model-00009-of-00011.safetensors",
382
+ "transformer.layers.17.post_attention_layernorm.weight": "model-00009-of-00011.safetensors",
383
+ "transformer.layers.17.self_attn.k_proj.weight": "model-00009-of-00011.safetensors",
384
+ "transformer.layers.17.self_attn.o_proj.weight": "model-00009-of-00011.safetensors",
385
+ "transformer.layers.17.self_attn.q_proj.weight": "model-00009-of-00011.safetensors",
386
+ "transformer.layers.17.self_attn.v_proj.weight": "model-00009-of-00011.safetensors",
387
+ "transformer.layers.18.input_layernorm.weight": "model-00009-of-00011.safetensors",
388
+ "transformer.layers.18.mlp.down_proj.weight": "model-00009-of-00011.safetensors",
389
+ "transformer.layers.18.mlp.gate_proj.weight": "model-00009-of-00011.safetensors",
390
+ "transformer.layers.18.mlp.up_proj.weight": "model-00009-of-00011.safetensors",
391
+ "transformer.layers.18.post_attention_layernorm.weight": "model-00009-of-00011.safetensors",
392
+ "transformer.layers.18.self_attn.k_proj.weight": "model-00009-of-00011.safetensors",
393
+ "transformer.layers.18.self_attn.o_proj.weight": "model-00009-of-00011.safetensors",
394
+ "transformer.layers.18.self_attn.q_proj.weight": "model-00009-of-00011.safetensors",
395
+ "transformer.layers.18.self_attn.v_proj.weight": "model-00009-of-00011.safetensors",
396
+ "transformer.layers.19.input_layernorm.weight": "model-00009-of-00011.safetensors",
397
+ "transformer.layers.19.mlp.down_proj.weight": "model-00009-of-00011.safetensors",
398
+ "transformer.layers.19.mlp.gate_proj.weight": "model-00009-of-00011.safetensors",
399
+ "transformer.layers.19.mlp.up_proj.weight": "model-00009-of-00011.safetensors",
400
+ "transformer.layers.19.post_attention_layernorm.weight": "model-00009-of-00011.safetensors",
401
+ "transformer.layers.19.self_attn.k_proj.weight": "model-00009-of-00011.safetensors",
402
+ "transformer.layers.19.self_attn.o_proj.weight": "model-00009-of-00011.safetensors",
403
+ "transformer.layers.19.self_attn.q_proj.weight": "model-00009-of-00011.safetensors",
404
+ "transformer.layers.19.self_attn.v_proj.weight": "model-00009-of-00011.safetensors",
405
+ "transformer.layers.2.input_layernorm.weight": "model-00007-of-00011.safetensors",
406
+ "transformer.layers.2.mlp.down_proj.weight": "model-00007-of-00011.safetensors",
407
+ "transformer.layers.2.mlp.gate_proj.weight": "model-00006-of-00011.safetensors",
408
+ "transformer.layers.2.mlp.up_proj.weight": "model-00006-of-00011.safetensors",
409
+ "transformer.layers.2.post_attention_layernorm.weight": "model-00007-of-00011.safetensors",
410
+ "transformer.layers.2.self_attn.k_proj.weight": "model-00006-of-00011.safetensors",
411
+ "transformer.layers.2.self_attn.o_proj.weight": "model-00006-of-00011.safetensors",
412
+ "transformer.layers.2.self_attn.q_proj.weight": "model-00006-of-00011.safetensors",
413
+ "transformer.layers.2.self_attn.v_proj.weight": "model-00006-of-00011.safetensors",
414
+ "transformer.layers.20.input_layernorm.weight": "model-00010-of-00011.safetensors",
415
+ "transformer.layers.20.mlp.down_proj.weight": "model-00010-of-00011.safetensors",
416
+ "transformer.layers.20.mlp.gate_proj.weight": "model-00009-of-00011.safetensors",
417
+ "transformer.layers.20.mlp.up_proj.weight": "model-00009-of-00011.safetensors",
418
+ "transformer.layers.20.post_attention_layernorm.weight": "model-00010-of-00011.safetensors",
419
+ "transformer.layers.20.self_attn.k_proj.weight": "model-00009-of-00011.safetensors",
420
+ "transformer.layers.20.self_attn.o_proj.weight": "model-00009-of-00011.safetensors",
421
+ "transformer.layers.20.self_attn.q_proj.weight": "model-00009-of-00011.safetensors",
422
+ "transformer.layers.20.self_attn.v_proj.weight": "model-00009-of-00011.safetensors",
423
+ "transformer.layers.21.input_layernorm.weight": "model-00010-of-00011.safetensors",
424
+ "transformer.layers.21.mlp.down_proj.weight": "model-00010-of-00011.safetensors",
425
+ "transformer.layers.21.mlp.gate_proj.weight": "model-00010-of-00011.safetensors",
426
+ "transformer.layers.21.mlp.up_proj.weight": "model-00010-of-00011.safetensors",
427
+ "transformer.layers.21.post_attention_layernorm.weight": "model-00010-of-00011.safetensors",
428
+ "transformer.layers.21.self_attn.k_proj.weight": "model-00010-of-00011.safetensors",
429
+ "transformer.layers.21.self_attn.o_proj.weight": "model-00010-of-00011.safetensors",
430
+ "transformer.layers.21.self_attn.q_proj.weight": "model-00010-of-00011.safetensors",
431
+ "transformer.layers.21.self_attn.v_proj.weight": "model-00010-of-00011.safetensors",
432
+ "transformer.layers.22.input_layernorm.weight": "model-00010-of-00011.safetensors",
433
+ "transformer.layers.22.mlp.down_proj.weight": "model-00010-of-00011.safetensors",
434
+ "transformer.layers.22.mlp.gate_proj.weight": "model-00010-of-00011.safetensors",
435
+ "transformer.layers.22.mlp.up_proj.weight": "model-00010-of-00011.safetensors",
436
+ "transformer.layers.22.post_attention_layernorm.weight": "model-00010-of-00011.safetensors",
437
+ "transformer.layers.22.self_attn.k_proj.weight": "model-00010-of-00011.safetensors",
438
+ "transformer.layers.22.self_attn.o_proj.weight": "model-00010-of-00011.safetensors",
439
+ "transformer.layers.22.self_attn.q_proj.weight": "model-00010-of-00011.safetensors",
440
+ "transformer.layers.22.self_attn.v_proj.weight": "model-00010-of-00011.safetensors",
441
+ "transformer.layers.23.input_layernorm.weight": "model-00010-of-00011.safetensors",
442
+ "transformer.layers.23.mlp.down_proj.weight": "model-00010-of-00011.safetensors",
443
+ "transformer.layers.23.mlp.gate_proj.weight": "model-00010-of-00011.safetensors",
444
+ "transformer.layers.23.mlp.up_proj.weight": "model-00010-of-00011.safetensors",
445
+ "transformer.layers.23.post_attention_layernorm.weight": "model-00010-of-00011.safetensors",
446
+ "transformer.layers.23.self_attn.k_proj.weight": "model-00010-of-00011.safetensors",
447
+ "transformer.layers.23.self_attn.o_proj.weight": "model-00010-of-00011.safetensors",
448
+ "transformer.layers.23.self_attn.q_proj.weight": "model-00010-of-00011.safetensors",
449
+ "transformer.layers.23.self_attn.v_proj.weight": "model-00010-of-00011.safetensors",
450
+ "transformer.layers.24.input_layernorm.weight": "model-00010-of-00011.safetensors",
451
+ "transformer.layers.24.mlp.down_proj.weight": "model-00010-of-00011.safetensors",
452
+ "transformer.layers.24.mlp.gate_proj.weight": "model-00010-of-00011.safetensors",
453
+ "transformer.layers.24.mlp.up_proj.weight": "model-00010-of-00011.safetensors",
454
+ "transformer.layers.24.post_attention_layernorm.weight": "model-00010-of-00011.safetensors",
455
+ "transformer.layers.24.self_attn.k_proj.weight": "model-00010-of-00011.safetensors",
456
+ "transformer.layers.24.self_attn.o_proj.weight": "model-00010-of-00011.safetensors",
457
+ "transformer.layers.24.self_attn.q_proj.weight": "model-00010-of-00011.safetensors",
458
+ "transformer.layers.24.self_attn.v_proj.weight": "model-00010-of-00011.safetensors",
459
+ "transformer.layers.25.input_layernorm.weight": "model-00010-of-00011.safetensors",
460
+ "transformer.layers.25.mlp.down_proj.weight": "model-00010-of-00011.safetensors",
461
+ "transformer.layers.25.mlp.gate_proj.weight": "model-00010-of-00011.safetensors",
462
+ "transformer.layers.25.mlp.up_proj.weight": "model-00010-of-00011.safetensors",
463
+ "transformer.layers.25.post_attention_layernorm.weight": "model-00010-of-00011.safetensors",
464
+ "transformer.layers.25.self_attn.k_proj.weight": "model-00010-of-00011.safetensors",
465
+ "transformer.layers.25.self_attn.o_proj.weight": "model-00010-of-00011.safetensors",
466
+ "transformer.layers.25.self_attn.q_proj.weight": "model-00010-of-00011.safetensors",
467
+ "transformer.layers.25.self_attn.v_proj.weight": "model-00010-of-00011.safetensors",
468
+ "transformer.layers.26.input_layernorm.weight": "model-00011-of-00011.safetensors",
469
+ "transformer.layers.26.mlp.down_proj.weight": "model-00011-of-00011.safetensors",
470
+ "transformer.layers.26.mlp.gate_proj.weight": "model-00010-of-00011.safetensors",
471
+ "transformer.layers.26.mlp.up_proj.weight": "model-00010-of-00011.safetensors",
472
+ "transformer.layers.26.post_attention_layernorm.weight": "model-00011-of-00011.safetensors",
473
+ "transformer.layers.26.self_attn.k_proj.weight": "model-00010-of-00011.safetensors",
474
+ "transformer.layers.26.self_attn.o_proj.weight": "model-00010-of-00011.safetensors",
475
+ "transformer.layers.26.self_attn.q_proj.weight": "model-00010-of-00011.safetensors",
476
+ "transformer.layers.26.self_attn.v_proj.weight": "model-00010-of-00011.safetensors",
477
+ "transformer.layers.27.input_layernorm.weight": "model-00011-of-00011.safetensors",
478
+ "transformer.layers.27.mlp.down_proj.weight": "model-00011-of-00011.safetensors",
479
+ "transformer.layers.27.mlp.gate_proj.weight": "model-00011-of-00011.safetensors",
480
+ "transformer.layers.27.mlp.up_proj.weight": "model-00011-of-00011.safetensors",
481
+ "transformer.layers.27.post_attention_layernorm.weight": "model-00011-of-00011.safetensors",
482
+ "transformer.layers.27.self_attn.k_proj.weight": "model-00011-of-00011.safetensors",
483
+ "transformer.layers.27.self_attn.o_proj.weight": "model-00011-of-00011.safetensors",
484
+ "transformer.layers.27.self_attn.q_proj.weight": "model-00011-of-00011.safetensors",
485
+ "transformer.layers.27.self_attn.v_proj.weight": "model-00011-of-00011.safetensors",
486
+ "transformer.layers.28.input_layernorm.weight": "model-00011-of-00011.safetensors",
487
+ "transformer.layers.28.mlp.down_proj.weight": "model-00011-of-00011.safetensors",
488
+ "transformer.layers.28.mlp.gate_proj.weight": "model-00011-of-00011.safetensors",
489
+ "transformer.layers.28.mlp.up_proj.weight": "model-00011-of-00011.safetensors",
490
+ "transformer.layers.28.post_attention_layernorm.weight": "model-00011-of-00011.safetensors",
491
+ "transformer.layers.28.self_attn.k_proj.weight": "model-00011-of-00011.safetensors",
492
+ "transformer.layers.28.self_attn.o_proj.weight": "model-00011-of-00011.safetensors",
493
+ "transformer.layers.28.self_attn.q_proj.weight": "model-00011-of-00011.safetensors",
494
+ "transformer.layers.28.self_attn.v_proj.weight": "model-00011-of-00011.safetensors",
495
+ "transformer.layers.29.input_layernorm.weight": "model-00011-of-00011.safetensors",
496
+ "transformer.layers.29.mlp.down_proj.weight": "model-00011-of-00011.safetensors",
497
+ "transformer.layers.29.mlp.gate_proj.weight": "model-00011-of-00011.safetensors",
498
+ "transformer.layers.29.mlp.up_proj.weight": "model-00011-of-00011.safetensors",
499
+ "transformer.layers.29.post_attention_layernorm.weight": "model-00011-of-00011.safetensors",
500
+ "transformer.layers.29.self_attn.k_proj.weight": "model-00011-of-00011.safetensors",
501
+ "transformer.layers.29.self_attn.o_proj.weight": "model-00011-of-00011.safetensors",
502
+ "transformer.layers.29.self_attn.q_proj.weight": "model-00011-of-00011.safetensors",
503
+ "transformer.layers.29.self_attn.v_proj.weight": "model-00011-of-00011.safetensors",
504
+ "transformer.layers.3.input_layernorm.weight": "model-00007-of-00011.safetensors",
505
+ "transformer.layers.3.mlp.down_proj.weight": "model-00007-of-00011.safetensors",
506
+ "transformer.layers.3.mlp.gate_proj.weight": "model-00007-of-00011.safetensors",
507
+ "transformer.layers.3.mlp.up_proj.weight": "model-00007-of-00011.safetensors",
508
+ "transformer.layers.3.post_attention_layernorm.weight": "model-00007-of-00011.safetensors",
509
+ "transformer.layers.3.self_attn.k_proj.weight": "model-00007-of-00011.safetensors",
510
+ "transformer.layers.3.self_attn.o_proj.weight": "model-00007-of-00011.safetensors",
511
+ "transformer.layers.3.self_attn.q_proj.weight": "model-00007-of-00011.safetensors",
512
+ "transformer.layers.3.self_attn.v_proj.weight": "model-00007-of-00011.safetensors",
513
+ "transformer.layers.30.input_layernorm.weight": "model-00011-of-00011.safetensors",
514
+ "transformer.layers.30.mlp.down_proj.weight": "model-00011-of-00011.safetensors",
515
+ "transformer.layers.30.mlp.gate_proj.weight": "model-00011-of-00011.safetensors",
516
+ "transformer.layers.30.mlp.up_proj.weight": "model-00011-of-00011.safetensors",
517
+ "transformer.layers.30.post_attention_layernorm.weight": "model-00011-of-00011.safetensors",
518
+ "transformer.layers.30.self_attn.k_proj.weight": "model-00011-of-00011.safetensors",
519
+ "transformer.layers.30.self_attn.o_proj.weight": "model-00011-of-00011.safetensors",
520
+ "transformer.layers.30.self_attn.q_proj.weight": "model-00011-of-00011.safetensors",
521
+ "transformer.layers.30.self_attn.v_proj.weight": "model-00011-of-00011.safetensors",
522
+ "transformer.layers.31.input_layernorm.weight": "model-00011-of-00011.safetensors",
523
+ "transformer.layers.31.mlp.down_proj.weight": "model-00011-of-00011.safetensors",
524
+ "transformer.layers.31.mlp.gate_proj.weight": "model-00011-of-00011.safetensors",
525
+ "transformer.layers.31.mlp.up_proj.weight": "model-00011-of-00011.safetensors",
526
+ "transformer.layers.31.post_attention_layernorm.weight": "model-00011-of-00011.safetensors",
527
+ "transformer.layers.31.self_attn.k_proj.weight": "model-00011-of-00011.safetensors",
528
+ "transformer.layers.31.self_attn.o_proj.weight": "model-00011-of-00011.safetensors",
529
+ "transformer.layers.31.self_attn.q_proj.weight": "model-00011-of-00011.safetensors",
530
+ "transformer.layers.31.self_attn.v_proj.weight": "model-00011-of-00011.safetensors",
531
+ "transformer.layers.4.input_layernorm.weight": "model-00007-of-00011.safetensors",
532
+ "transformer.layers.4.mlp.down_proj.weight": "model-00007-of-00011.safetensors",
533
+ "transformer.layers.4.mlp.gate_proj.weight": "model-00007-of-00011.safetensors",
534
+ "transformer.layers.4.mlp.up_proj.weight": "model-00007-of-00011.safetensors",
535
+ "transformer.layers.4.post_attention_layernorm.weight": "model-00007-of-00011.safetensors",
536
+ "transformer.layers.4.self_attn.k_proj.weight": "model-00007-of-00011.safetensors",
537
+ "transformer.layers.4.self_attn.o_proj.weight": "model-00007-of-00011.safetensors",
538
+ "transformer.layers.4.self_attn.q_proj.weight": "model-00007-of-00011.safetensors",
539
+ "transformer.layers.4.self_attn.v_proj.weight": "model-00007-of-00011.safetensors",
540
+ "transformer.layers.5.input_layernorm.weight": "model-00007-of-00011.safetensors",
541
+ "transformer.layers.5.mlp.down_proj.weight": "model-00007-of-00011.safetensors",
542
+ "transformer.layers.5.mlp.gate_proj.weight": "model-00007-of-00011.safetensors",
543
+ "transformer.layers.5.mlp.up_proj.weight": "model-00007-of-00011.safetensors",
544
+ "transformer.layers.5.post_attention_layernorm.weight": "model-00007-of-00011.safetensors",
545
+ "transformer.layers.5.self_attn.k_proj.weight": "model-00007-of-00011.safetensors",
546
+ "transformer.layers.5.self_attn.o_proj.weight": "model-00007-of-00011.safetensors",
547
+ "transformer.layers.5.self_attn.q_proj.weight": "model-00007-of-00011.safetensors",
548
+ "transformer.layers.5.self_attn.v_proj.weight": "model-00007-of-00011.safetensors",
549
+ "transformer.layers.6.input_layernorm.weight": "model-00007-of-00011.safetensors",
550
+ "transformer.layers.6.mlp.down_proj.weight": "model-00007-of-00011.safetensors",
551
+ "transformer.layers.6.mlp.gate_proj.weight": "model-00007-of-00011.safetensors",
552
+ "transformer.layers.6.mlp.up_proj.weight": "model-00007-of-00011.safetensors",
553
+ "transformer.layers.6.post_attention_layernorm.weight": "model-00007-of-00011.safetensors",
554
+ "transformer.layers.6.self_attn.k_proj.weight": "model-00007-of-00011.safetensors",
555
+ "transformer.layers.6.self_attn.o_proj.weight": "model-00007-of-00011.safetensors",
556
+ "transformer.layers.6.self_attn.q_proj.weight": "model-00007-of-00011.safetensors",
557
+ "transformer.layers.6.self_attn.v_proj.weight": "model-00007-of-00011.safetensors",
558
+ "transformer.layers.7.input_layernorm.weight": "model-00007-of-00011.safetensors",
559
+ "transformer.layers.7.mlp.down_proj.weight": "model-00007-of-00011.safetensors",
560
+ "transformer.layers.7.mlp.gate_proj.weight": "model-00007-of-00011.safetensors",
561
+ "transformer.layers.7.mlp.up_proj.weight": "model-00007-of-00011.safetensors",
562
+ "transformer.layers.7.post_attention_layernorm.weight": "model-00007-of-00011.safetensors",
563
+ "transformer.layers.7.self_attn.k_proj.weight": "model-00007-of-00011.safetensors",
564
+ "transformer.layers.7.self_attn.o_proj.weight": "model-00007-of-00011.safetensors",
565
+ "transformer.layers.7.self_attn.q_proj.weight": "model-00007-of-00011.safetensors",
566
+ "transformer.layers.7.self_attn.v_proj.weight": "model-00007-of-00011.safetensors",
567
+ "transformer.layers.8.input_layernorm.weight": "model-00008-of-00011.safetensors",
568
+ "transformer.layers.8.mlp.down_proj.weight": "model-00008-of-00011.safetensors",
569
+ "transformer.layers.8.mlp.gate_proj.weight": "model-00007-of-00011.safetensors",
570
+ "transformer.layers.8.mlp.up_proj.weight": "model-00007-of-00011.safetensors",
571
+ "transformer.layers.8.post_attention_layernorm.weight": "model-00008-of-00011.safetensors",
572
+ "transformer.layers.8.self_attn.k_proj.weight": "model-00007-of-00011.safetensors",
573
+ "transformer.layers.8.self_attn.o_proj.weight": "model-00007-of-00011.safetensors",
574
+ "transformer.layers.8.self_attn.q_proj.weight": "model-00007-of-00011.safetensors",
575
+ "transformer.layers.8.self_attn.v_proj.weight": "model-00007-of-00011.safetensors",
576
+ "transformer.layers.9.input_layernorm.weight": "model-00008-of-00011.safetensors",
577
+ "transformer.layers.9.mlp.down_proj.weight": "model-00008-of-00011.safetensors",
578
+ "transformer.layers.9.mlp.gate_proj.weight": "model-00008-of-00011.safetensors",
579
+ "transformer.layers.9.mlp.up_proj.weight": "model-00008-of-00011.safetensors",
580
+ "transformer.layers.9.post_attention_layernorm.weight": "model-00008-of-00011.safetensors",
581
+ "transformer.layers.9.self_attn.k_proj.weight": "model-00008-of-00011.safetensors",
582
+ "transformer.layers.9.self_attn.o_proj.weight": "model-00008-of-00011.safetensors",
583
+ "transformer.layers.9.self_attn.q_proj.weight": "model-00008-of-00011.safetensors",
584
+ "transformer.layers.9.self_attn.v_proj.weight": "model-00008-of-00011.safetensors",
585
+ "transformer.norm.weight": "model-00011-of-00011.safetensors",
586
+ "value_head.weight": "model-00011-of-00011.safetensors"
587
+ }
588
+ }