CharlieFRuan commited on
Commit
8862500
1 Parent(s): f6b6088

Initial commit

Browse files
added_tokens.json ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ {
2
+ "<|im_end|>": 32002,
3
+ "<|im_start|>": 32001,
4
+ "[PAD]": 32000
5
+ }
logs.txt ADDED
The diff for this file is too large to render. See raw diff
 
mlc-chat-config.json ADDED
@@ -0,0 +1,43 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "model_type": "llama",
3
+ "quantization": "q4f32_1",
4
+ "model_config": {
5
+ "hidden_size": 2048,
6
+ "intermediate_size": 5632,
7
+ "num_attention_heads": 32,
8
+ "num_hidden_layers": 22,
9
+ "rms_norm_eps": 1e-05,
10
+ "vocab_size": 32003,
11
+ "position_embedding_base": 10000.0,
12
+ "context_window_size": 2048,
13
+ "prefill_chunk_size": 2048,
14
+ "num_key_value_heads": 4,
15
+ "head_dim": 64,
16
+ "tensor_parallel_shards": 1,
17
+ "max_batch_size": 1
18
+ },
19
+ "vocab_size": 32003,
20
+ "context_window_size": 2048,
21
+ "sliding_window_size": -1,
22
+ "prefill_chunk_size": 2048,
23
+ "attention_sink_size": -1,
24
+ "tensor_parallel_shards": 1,
25
+ "max_batch_size": 80,
26
+ "mean_gen_len": 128,
27
+ "max_gen_len": 512,
28
+ "shift_fill_factor": 0.3,
29
+ "temperature": 0.7,
30
+ "repetition_penalty": 1.0,
31
+ "top_p": 0.95,
32
+ "conv_template": "chatml",
33
+ "pad_token_id": 0,
34
+ "bos_token_id": 1,
35
+ "eos_token_id": 2,
36
+ "tokenizer_files": [
37
+ "tokenizer.model",
38
+ "tokenizer.json",
39
+ "added_tokens.json",
40
+ "tokenizer_config.json"
41
+ ],
42
+ "version": "0.1.0"
43
+ }
ndarray-cache-b16.json ADDED
@@ -0,0 +1,2631 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metadata": {
3
+ "ParamSize": 225,
4
+ "ParamBytes": 687848960.0,
5
+ "BitsPerParam": 5.002261984328079
6
+ },
7
+ "records": [
8
+ {
9
+ "dataPath": "params_shard_0.bin",
10
+ "format": "raw-shard",
11
+ "nbytes": 32771072,
12
+ "records": [
13
+ {
14
+ "name": "model.embed_tokens.q_weight",
15
+ "shape": [
16
+ 32003,
17
+ 256
18
+ ],
19
+ "dtype": "uint32",
20
+ "format": "f32-to-bf16",
21
+ "nbytes": 32771072,
22
+ "byteOffset": 0
23
+ }
24
+ ],
25
+ "md5sum": "5d04d3daaea233773d55c738ca893695"
26
+ },
27
+ {
28
+ "dataPath": "params_shard_1.bin",
29
+ "format": "raw-shard",
30
+ "nbytes": 31826304,
31
+ "records": [
32
+ {
33
+ "name": "model.embed_tokens.q_scale",
34
+ "shape": [
35
+ 32003,
36
+ 64
37
+ ],
38
+ "dtype": "bfloat16",
39
+ "format": "raw",
40
+ "nbytes": 4096384,
41
+ "byteOffset": 0
42
+ },
43
+ {
44
+ "name": "model.layers.0.self_attn.qkv_proj.q_weight",
45
+ "shape": [
46
+ 2560,
47
+ 256
48
+ ],
49
+ "dtype": "uint32",
50
+ "format": "f32-to-bf16",
51
+ "nbytes": 2621440,
52
+ "byteOffset": 4096384
53
+ },
54
+ {
55
+ "name": "model.layers.0.self_attn.qkv_proj.q_scale",
56
+ "shape": [
57
+ 2560,
58
+ 64
59
+ ],
60
+ "dtype": "bfloat16",
61
+ "format": "raw",
62
+ "nbytes": 327680,
63
+ "byteOffset": 6717824
64
+ },
65
+ {
66
+ "name": "model.layers.0.self_attn.o_proj.q_weight",
67
+ "shape": [
68
+ 2048,
69
+ 256
70
+ ],
71
+ "dtype": "uint32",
72
+ "format": "f32-to-bf16",
73
+ "nbytes": 2097152,
74
+ "byteOffset": 7045504
75
+ },
76
+ {
77
+ "name": "model.layers.0.self_attn.o_proj.q_scale",
78
+ "shape": [
79
+ 2048,
80
+ 64
81
+ ],
82
+ "dtype": "bfloat16",
83
+ "format": "raw",
84
+ "nbytes": 262144,
85
+ "byteOffset": 9142656
86
+ },
87
+ {
88
+ "name": "model.layers.0.mlp.gate_up_proj.q_weight",
89
+ "shape": [
90
+ 11264,
91
+ 256
92
+ ],
93
+ "dtype": "uint32",
94
+ "format": "f32-to-bf16",
95
+ "nbytes": 11534336,
96
+ "byteOffset": 9404800
97
+ },
98
+ {
99
+ "name": "model.layers.0.mlp.gate_up_proj.q_scale",
100
+ "shape": [
101
+ 11264,
102
+ 64
103
+ ],
104
+ "dtype": "bfloat16",
105
+ "format": "raw",
106
+ "nbytes": 1441792,
107
+ "byteOffset": 20939136
108
+ },
109
+ {
110
+ "name": "model.layers.0.mlp.down_proj.q_weight",
111
+ "shape": [
112
+ 2048,
113
+ 704
114
+ ],
115
+ "dtype": "uint32",
116
+ "format": "f32-to-bf16",
117
+ "nbytes": 5767168,
118
+ "byteOffset": 22380928
119
+ },
120
+ {
121
+ "name": "model.layers.0.mlp.down_proj.q_scale",
122
+ "shape": [
123
+ 2048,
124
+ 176
125
+ ],
126
+ "dtype": "bfloat16",
127
+ "format": "raw",
128
+ "nbytes": 720896,
129
+ "byteOffset": 28148096
130
+ },
131
+ {
132
+ "name": "model.layers.0.input_layernorm.weight",
133
+ "shape": [
134
+ 2048
135
+ ],
136
+ "dtype": "bfloat16",
137
+ "format": "raw",
138
+ "nbytes": 4096,
139
+ "byteOffset": 28868992
140
+ },
141
+ {
142
+ "name": "model.layers.0.post_attention_layernorm.weight",
143
+ "shape": [
144
+ 2048
145
+ ],
146
+ "dtype": "bfloat16",
147
+ "format": "raw",
148
+ "nbytes": 4096,
149
+ "byteOffset": 28873088
150
+ },
151
+ {
152
+ "name": "model.layers.1.self_attn.qkv_proj.q_weight",
153
+ "shape": [
154
+ 2560,
155
+ 256
156
+ ],
157
+ "dtype": "uint32",
158
+ "format": "f32-to-bf16",
159
+ "nbytes": 2621440,
160
+ "byteOffset": 28877184
161
+ },
162
+ {
163
+ "name": "model.layers.1.self_attn.qkv_proj.q_scale",
164
+ "shape": [
165
+ 2560,
166
+ 64
167
+ ],
168
+ "dtype": "bfloat16",
169
+ "format": "raw",
170
+ "nbytes": 327680,
171
+ "byteOffset": 31498624
172
+ }
173
+ ],
174
+ "md5sum": "680c11ff1744d5bfa3cab96425d32284"
175
+ },
176
+ {
177
+ "dataPath": "params_shard_2.bin",
178
+ "format": "raw-shard",
179
+ "nbytes": 27140096,
180
+ "records": [
181
+ {
182
+ "name": "model.layers.1.self_attn.o_proj.q_weight",
183
+ "shape": [
184
+ 2048,
185
+ 256
186
+ ],
187
+ "dtype": "uint32",
188
+ "format": "f32-to-bf16",
189
+ "nbytes": 2097152,
190
+ "byteOffset": 0
191
+ },
192
+ {
193
+ "name": "model.layers.1.self_attn.o_proj.q_scale",
194
+ "shape": [
195
+ 2048,
196
+ 64
197
+ ],
198
+ "dtype": "bfloat16",
199
+ "format": "raw",
200
+ "nbytes": 262144,
201
+ "byteOffset": 2097152
202
+ },
203
+ {
204
+ "name": "model.layers.1.mlp.gate_up_proj.q_weight",
205
+ "shape": [
206
+ 11264,
207
+ 256
208
+ ],
209
+ "dtype": "uint32",
210
+ "format": "f32-to-bf16",
211
+ "nbytes": 11534336,
212
+ "byteOffset": 2359296
213
+ },
214
+ {
215
+ "name": "model.layers.1.mlp.gate_up_proj.q_scale",
216
+ "shape": [
217
+ 11264,
218
+ 64
219
+ ],
220
+ "dtype": "bfloat16",
221
+ "format": "raw",
222
+ "nbytes": 1441792,
223
+ "byteOffset": 13893632
224
+ },
225
+ {
226
+ "name": "model.layers.1.mlp.down_proj.q_weight",
227
+ "shape": [
228
+ 2048,
229
+ 704
230
+ ],
231
+ "dtype": "uint32",
232
+ "format": "f32-to-bf16",
233
+ "nbytes": 5767168,
234
+ "byteOffset": 15335424
235
+ },
236
+ {
237
+ "name": "model.layers.1.mlp.down_proj.q_scale",
238
+ "shape": [
239
+ 2048,
240
+ 176
241
+ ],
242
+ "dtype": "bfloat16",
243
+ "format": "raw",
244
+ "nbytes": 720896,
245
+ "byteOffset": 21102592
246
+ },
247
+ {
248
+ "name": "model.layers.1.input_layernorm.weight",
249
+ "shape": [
250
+ 2048
251
+ ],
252
+ "dtype": "bfloat16",
253
+ "format": "raw",
254
+ "nbytes": 4096,
255
+ "byteOffset": 21823488
256
+ },
257
+ {
258
+ "name": "model.layers.1.post_attention_layernorm.weight",
259
+ "shape": [
260
+ 2048
261
+ ],
262
+ "dtype": "bfloat16",
263
+ "format": "raw",
264
+ "nbytes": 4096,
265
+ "byteOffset": 21827584
266
+ },
267
+ {
268
+ "name": "model.layers.2.self_attn.qkv_proj.q_weight",
269
+ "shape": [
270
+ 2560,
271
+ 256
272
+ ],
273
+ "dtype": "uint32",
274
+ "format": "f32-to-bf16",
275
+ "nbytes": 2621440,
276
+ "byteOffset": 21831680
277
+ },
278
+ {
279
+ "name": "model.layers.2.self_attn.qkv_proj.q_scale",
280
+ "shape": [
281
+ 2560,
282
+ 64
283
+ ],
284
+ "dtype": "bfloat16",
285
+ "format": "raw",
286
+ "nbytes": 327680,
287
+ "byteOffset": 24453120
288
+ },
289
+ {
290
+ "name": "model.layers.2.self_attn.o_proj.q_weight",
291
+ "shape": [
292
+ 2048,
293
+ 256
294
+ ],
295
+ "dtype": "uint32",
296
+ "format": "f32-to-bf16",
297
+ "nbytes": 2097152,
298
+ "byteOffset": 24780800
299
+ },
300
+ {
301
+ "name": "model.layers.2.self_attn.o_proj.q_scale",
302
+ "shape": [
303
+ 2048,
304
+ 64
305
+ ],
306
+ "dtype": "bfloat16",
307
+ "format": "raw",
308
+ "nbytes": 262144,
309
+ "byteOffset": 26877952
310
+ }
311
+ ],
312
+ "md5sum": "8c28be42014a0308dacc2f2ecdf10aa3"
313
+ },
314
+ {
315
+ "dataPath": "params_shard_3.bin",
316
+ "format": "raw-shard",
317
+ "nbytes": 24780800,
318
+ "records": [
319
+ {
320
+ "name": "model.layers.2.mlp.gate_up_proj.q_weight",
321
+ "shape": [
322
+ 11264,
323
+ 256
324
+ ],
325
+ "dtype": "uint32",
326
+ "format": "f32-to-bf16",
327
+ "nbytes": 11534336,
328
+ "byteOffset": 0
329
+ },
330
+ {
331
+ "name": "model.layers.2.mlp.gate_up_proj.q_scale",
332
+ "shape": [
333
+ 11264,
334
+ 64
335
+ ],
336
+ "dtype": "bfloat16",
337
+ "format": "raw",
338
+ "nbytes": 1441792,
339
+ "byteOffset": 11534336
340
+ },
341
+ {
342
+ "name": "model.layers.2.mlp.down_proj.q_weight",
343
+ "shape": [
344
+ 2048,
345
+ 704
346
+ ],
347
+ "dtype": "uint32",
348
+ "format": "f32-to-bf16",
349
+ "nbytes": 5767168,
350
+ "byteOffset": 12976128
351
+ },
352
+ {
353
+ "name": "model.layers.2.mlp.down_proj.q_scale",
354
+ "shape": [
355
+ 2048,
356
+ 176
357
+ ],
358
+ "dtype": "bfloat16",
359
+ "format": "raw",
360
+ "nbytes": 720896,
361
+ "byteOffset": 18743296
362
+ },
363
+ {
364
+ "name": "model.layers.2.input_layernorm.weight",
365
+ "shape": [
366
+ 2048
367
+ ],
368
+ "dtype": "bfloat16",
369
+ "format": "raw",
370
+ "nbytes": 4096,
371
+ "byteOffset": 19464192
372
+ },
373
+ {
374
+ "name": "model.layers.2.post_attention_layernorm.weight",
375
+ "shape": [
376
+ 2048
377
+ ],
378
+ "dtype": "bfloat16",
379
+ "format": "raw",
380
+ "nbytes": 4096,
381
+ "byteOffset": 19468288
382
+ },
383
+ {
384
+ "name": "model.layers.3.self_attn.qkv_proj.q_weight",
385
+ "shape": [
386
+ 2560,
387
+ 256
388
+ ],
389
+ "dtype": "uint32",
390
+ "format": "f32-to-bf16",
391
+ "nbytes": 2621440,
392
+ "byteOffset": 19472384
393
+ },
394
+ {
395
+ "name": "model.layers.3.self_attn.qkv_proj.q_scale",
396
+ "shape": [
397
+ 2560,
398
+ 64
399
+ ],
400
+ "dtype": "bfloat16",
401
+ "format": "raw",
402
+ "nbytes": 327680,
403
+ "byteOffset": 22093824
404
+ },
405
+ {
406
+ "name": "model.layers.3.self_attn.o_proj.q_weight",
407
+ "shape": [
408
+ 2048,
409
+ 256
410
+ ],
411
+ "dtype": "uint32",
412
+ "format": "f32-to-bf16",
413
+ "nbytes": 2097152,
414
+ "byteOffset": 22421504
415
+ },
416
+ {
417
+ "name": "model.layers.3.self_attn.o_proj.q_scale",
418
+ "shape": [
419
+ 2048,
420
+ 64
421
+ ],
422
+ "dtype": "bfloat16",
423
+ "format": "raw",
424
+ "nbytes": 262144,
425
+ "byteOffset": 24518656
426
+ }
427
+ ],
428
+ "md5sum": "a2fe11ae5b5c809f09adfcb9f58422be"
429
+ },
430
+ {
431
+ "dataPath": "params_shard_4.bin",
432
+ "format": "raw-shard",
433
+ "nbytes": 24780800,
434
+ "records": [
435
+ {
436
+ "name": "model.layers.3.mlp.gate_up_proj.q_weight",
437
+ "shape": [
438
+ 11264,
439
+ 256
440
+ ],
441
+ "dtype": "uint32",
442
+ "format": "f32-to-bf16",
443
+ "nbytes": 11534336,
444
+ "byteOffset": 0
445
+ },
446
+ {
447
+ "name": "model.layers.3.mlp.gate_up_proj.q_scale",
448
+ "shape": [
449
+ 11264,
450
+ 64
451
+ ],
452
+ "dtype": "bfloat16",
453
+ "format": "raw",
454
+ "nbytes": 1441792,
455
+ "byteOffset": 11534336
456
+ },
457
+ {
458
+ "name": "model.layers.3.mlp.down_proj.q_weight",
459
+ "shape": [
460
+ 2048,
461
+ 704
462
+ ],
463
+ "dtype": "uint32",
464
+ "format": "f32-to-bf16",
465
+ "nbytes": 5767168,
466
+ "byteOffset": 12976128
467
+ },
468
+ {
469
+ "name": "model.layers.3.mlp.down_proj.q_scale",
470
+ "shape": [
471
+ 2048,
472
+ 176
473
+ ],
474
+ "dtype": "bfloat16",
475
+ "format": "raw",
476
+ "nbytes": 720896,
477
+ "byteOffset": 18743296
478
+ },
479
+ {
480
+ "name": "model.layers.3.input_layernorm.weight",
481
+ "shape": [
482
+ 2048
483
+ ],
484
+ "dtype": "bfloat16",
485
+ "format": "raw",
486
+ "nbytes": 4096,
487
+ "byteOffset": 19464192
488
+ },
489
+ {
490
+ "name": "model.layers.3.post_attention_layernorm.weight",
491
+ "shape": [
492
+ 2048
493
+ ],
494
+ "dtype": "bfloat16",
495
+ "format": "raw",
496
+ "nbytes": 4096,
497
+ "byteOffset": 19468288
498
+ },
499
+ {
500
+ "name": "model.layers.4.self_attn.qkv_proj.q_weight",
501
+ "shape": [
502
+ 2560,
503
+ 256
504
+ ],
505
+ "dtype": "uint32",
506
+ "format": "f32-to-bf16",
507
+ "nbytes": 2621440,
508
+ "byteOffset": 19472384
509
+ },
510
+ {
511
+ "name": "model.layers.4.self_attn.qkv_proj.q_scale",
512
+ "shape": [
513
+ 2560,
514
+ 64
515
+ ],
516
+ "dtype": "bfloat16",
517
+ "format": "raw",
518
+ "nbytes": 327680,
519
+ "byteOffset": 22093824
520
+ },
521
+ {
522
+ "name": "model.layers.4.self_attn.o_proj.q_weight",
523
+ "shape": [
524
+ 2048,
525
+ 256
526
+ ],
527
+ "dtype": "uint32",
528
+ "format": "f32-to-bf16",
529
+ "nbytes": 2097152,
530
+ "byteOffset": 22421504
531
+ },
532
+ {
533
+ "name": "model.layers.4.self_attn.o_proj.q_scale",
534
+ "shape": [
535
+ 2048,
536
+ 64
537
+ ],
538
+ "dtype": "bfloat16",
539
+ "format": "raw",
540
+ "nbytes": 262144,
541
+ "byteOffset": 24518656
542
+ }
543
+ ],
544
+ "md5sum": "4cddfca8f748085f6fc7deeabda1ce27"
545
+ },
546
+ {
547
+ "dataPath": "params_shard_5.bin",
548
+ "format": "raw-shard",
549
+ "nbytes": 24780800,
550
+ "records": [
551
+ {
552
+ "name": "model.layers.4.mlp.gate_up_proj.q_weight",
553
+ "shape": [
554
+ 11264,
555
+ 256
556
+ ],
557
+ "dtype": "uint32",
558
+ "format": "f32-to-bf16",
559
+ "nbytes": 11534336,
560
+ "byteOffset": 0
561
+ },
562
+ {
563
+ "name": "model.layers.4.mlp.gate_up_proj.q_scale",
564
+ "shape": [
565
+ 11264,
566
+ 64
567
+ ],
568
+ "dtype": "bfloat16",
569
+ "format": "raw",
570
+ "nbytes": 1441792,
571
+ "byteOffset": 11534336
572
+ },
573
+ {
574
+ "name": "model.layers.4.mlp.down_proj.q_weight",
575
+ "shape": [
576
+ 2048,
577
+ 704
578
+ ],
579
+ "dtype": "uint32",
580
+ "format": "f32-to-bf16",
581
+ "nbytes": 5767168,
582
+ "byteOffset": 12976128
583
+ },
584
+ {
585
+ "name": "model.layers.4.mlp.down_proj.q_scale",
586
+ "shape": [
587
+ 2048,
588
+ 176
589
+ ],
590
+ "dtype": "bfloat16",
591
+ "format": "raw",
592
+ "nbytes": 720896,
593
+ "byteOffset": 18743296
594
+ },
595
+ {
596
+ "name": "model.layers.4.input_layernorm.weight",
597
+ "shape": [
598
+ 2048
599
+ ],
600
+ "dtype": "bfloat16",
601
+ "format": "raw",
602
+ "nbytes": 4096,
603
+ "byteOffset": 19464192
604
+ },
605
+ {
606
+ "name": "model.layers.4.post_attention_layernorm.weight",
607
+ "shape": [
608
+ 2048
609
+ ],
610
+ "dtype": "bfloat16",
611
+ "format": "raw",
612
+ "nbytes": 4096,
613
+ "byteOffset": 19468288
614
+ },
615
+ {
616
+ "name": "model.layers.5.self_attn.qkv_proj.q_weight",
617
+ "shape": [
618
+ 2560,
619
+ 256
620
+ ],
621
+ "dtype": "uint32",
622
+ "format": "f32-to-bf16",
623
+ "nbytes": 2621440,
624
+ "byteOffset": 19472384
625
+ },
626
+ {
627
+ "name": "model.layers.5.self_attn.qkv_proj.q_scale",
628
+ "shape": [
629
+ 2560,
630
+ 64
631
+ ],
632
+ "dtype": "bfloat16",
633
+ "format": "raw",
634
+ "nbytes": 327680,
635
+ "byteOffset": 22093824
636
+ },
637
+ {
638
+ "name": "model.layers.5.self_attn.o_proj.q_weight",
639
+ "shape": [
640
+ 2048,
641
+ 256
642
+ ],
643
+ "dtype": "uint32",
644
+ "format": "f32-to-bf16",
645
+ "nbytes": 2097152,
646
+ "byteOffset": 22421504
647
+ },
648
+ {
649
+ "name": "model.layers.5.self_attn.o_proj.q_scale",
650
+ "shape": [
651
+ 2048,
652
+ 64
653
+ ],
654
+ "dtype": "bfloat16",
655
+ "format": "raw",
656
+ "nbytes": 262144,
657
+ "byteOffset": 24518656
658
+ }
659
+ ],
660
+ "md5sum": "e120e89f444fd0c436baf3068f479762"
661
+ },
662
+ {
663
+ "dataPath": "params_shard_6.bin",
664
+ "format": "raw-shard",
665
+ "nbytes": 24780800,
666
+ "records": [
667
+ {
668
+ "name": "model.layers.5.mlp.gate_up_proj.q_weight",
669
+ "shape": [
670
+ 11264,
671
+ 256
672
+ ],
673
+ "dtype": "uint32",
674
+ "format": "f32-to-bf16",
675
+ "nbytes": 11534336,
676
+ "byteOffset": 0
677
+ },
678
+ {
679
+ "name": "model.layers.5.mlp.gate_up_proj.q_scale",
680
+ "shape": [
681
+ 11264,
682
+ 64
683
+ ],
684
+ "dtype": "bfloat16",
685
+ "format": "raw",
686
+ "nbytes": 1441792,
687
+ "byteOffset": 11534336
688
+ },
689
+ {
690
+ "name": "model.layers.5.mlp.down_proj.q_weight",
691
+ "shape": [
692
+ 2048,
693
+ 704
694
+ ],
695
+ "dtype": "uint32",
696
+ "format": "f32-to-bf16",
697
+ "nbytes": 5767168,
698
+ "byteOffset": 12976128
699
+ },
700
+ {
701
+ "name": "model.layers.5.mlp.down_proj.q_scale",
702
+ "shape": [
703
+ 2048,
704
+ 176
705
+ ],
706
+ "dtype": "bfloat16",
707
+ "format": "raw",
708
+ "nbytes": 720896,
709
+ "byteOffset": 18743296
710
+ },
711
+ {
712
+ "name": "model.layers.5.input_layernorm.weight",
713
+ "shape": [
714
+ 2048
715
+ ],
716
+ "dtype": "bfloat16",
717
+ "format": "raw",
718
+ "nbytes": 4096,
719
+ "byteOffset": 19464192
720
+ },
721
+ {
722
+ "name": "model.layers.5.post_attention_layernorm.weight",
723
+ "shape": [
724
+ 2048
725
+ ],
726
+ "dtype": "bfloat16",
727
+ "format": "raw",
728
+ "nbytes": 4096,
729
+ "byteOffset": 19468288
730
+ },
731
+ {
732
+ "name": "model.layers.6.self_attn.qkv_proj.q_weight",
733
+ "shape": [
734
+ 2560,
735
+ 256
736
+ ],
737
+ "dtype": "uint32",
738
+ "format": "f32-to-bf16",
739
+ "nbytes": 2621440,
740
+ "byteOffset": 19472384
741
+ },
742
+ {
743
+ "name": "model.layers.6.self_attn.qkv_proj.q_scale",
744
+ "shape": [
745
+ 2560,
746
+ 64
747
+ ],
748
+ "dtype": "bfloat16",
749
+ "format": "raw",
750
+ "nbytes": 327680,
751
+ "byteOffset": 22093824
752
+ },
753
+ {
754
+ "name": "model.layers.6.self_attn.o_proj.q_weight",
755
+ "shape": [
756
+ 2048,
757
+ 256
758
+ ],
759
+ "dtype": "uint32",
760
+ "format": "f32-to-bf16",
761
+ "nbytes": 2097152,
762
+ "byteOffset": 22421504
763
+ },
764
+ {
765
+ "name": "model.layers.6.self_attn.o_proj.q_scale",
766
+ "shape": [
767
+ 2048,
768
+ 64
769
+ ],
770
+ "dtype": "bfloat16",
771
+ "format": "raw",
772
+ "nbytes": 262144,
773
+ "byteOffset": 24518656
774
+ }
775
+ ],
776
+ "md5sum": "5a8837813b094d491b5a14205460886e"
777
+ },
778
+ {
779
+ "dataPath": "params_shard_7.bin",
780
+ "format": "raw-shard",
781
+ "nbytes": 24780800,
782
+ "records": [
783
+ {
784
+ "name": "model.layers.6.mlp.gate_up_proj.q_weight",
785
+ "shape": [
786
+ 11264,
787
+ 256
788
+ ],
789
+ "dtype": "uint32",
790
+ "format": "f32-to-bf16",
791
+ "nbytes": 11534336,
792
+ "byteOffset": 0
793
+ },
794
+ {
795
+ "name": "model.layers.6.mlp.gate_up_proj.q_scale",
796
+ "shape": [
797
+ 11264,
798
+ 64
799
+ ],
800
+ "dtype": "bfloat16",
801
+ "format": "raw",
802
+ "nbytes": 1441792,
803
+ "byteOffset": 11534336
804
+ },
805
+ {
806
+ "name": "model.layers.6.mlp.down_proj.q_weight",
807
+ "shape": [
808
+ 2048,
809
+ 704
810
+ ],
811
+ "dtype": "uint32",
812
+ "format": "f32-to-bf16",
813
+ "nbytes": 5767168,
814
+ "byteOffset": 12976128
815
+ },
816
+ {
817
+ "name": "model.layers.6.mlp.down_proj.q_scale",
818
+ "shape": [
819
+ 2048,
820
+ 176
821
+ ],
822
+ "dtype": "bfloat16",
823
+ "format": "raw",
824
+ "nbytes": 720896,
825
+ "byteOffset": 18743296
826
+ },
827
+ {
828
+ "name": "model.layers.6.input_layernorm.weight",
829
+ "shape": [
830
+ 2048
831
+ ],
832
+ "dtype": "bfloat16",
833
+ "format": "raw",
834
+ "nbytes": 4096,
835
+ "byteOffset": 19464192
836
+ },
837
+ {
838
+ "name": "model.layers.6.post_attention_layernorm.weight",
839
+ "shape": [
840
+ 2048
841
+ ],
842
+ "dtype": "bfloat16",
843
+ "format": "raw",
844
+ "nbytes": 4096,
845
+ "byteOffset": 19468288
846
+ },
847
+ {
848
+ "name": "model.layers.7.self_attn.qkv_proj.q_weight",
849
+ "shape": [
850
+ 2560,
851
+ 256
852
+ ],
853
+ "dtype": "uint32",
854
+ "format": "f32-to-bf16",
855
+ "nbytes": 2621440,
856
+ "byteOffset": 19472384
857
+ },
858
+ {
859
+ "name": "model.layers.7.self_attn.qkv_proj.q_scale",
860
+ "shape": [
861
+ 2560,
862
+ 64
863
+ ],
864
+ "dtype": "bfloat16",
865
+ "format": "raw",
866
+ "nbytes": 327680,
867
+ "byteOffset": 22093824
868
+ },
869
+ {
870
+ "name": "model.layers.7.self_attn.o_proj.q_weight",
871
+ "shape": [
872
+ 2048,
873
+ 256
874
+ ],
875
+ "dtype": "uint32",
876
+ "format": "f32-to-bf16",
877
+ "nbytes": 2097152,
878
+ "byteOffset": 22421504
879
+ },
880
+ {
881
+ "name": "model.layers.7.self_attn.o_proj.q_scale",
882
+ "shape": [
883
+ 2048,
884
+ 64
885
+ ],
886
+ "dtype": "bfloat16",
887
+ "format": "raw",
888
+ "nbytes": 262144,
889
+ "byteOffset": 24518656
890
+ }
891
+ ],
892
+ "md5sum": "4de03c54f7105a9e8fb0da0f275f40a3"
893
+ },
894
+ {
895
+ "dataPath": "params_shard_8.bin",
896
+ "format": "raw-shard",
897
+ "nbytes": 24780800,
898
+ "records": [
899
+ {
900
+ "name": "model.layers.7.mlp.gate_up_proj.q_weight",
901
+ "shape": [
902
+ 11264,
903
+ 256
904
+ ],
905
+ "dtype": "uint32",
906
+ "format": "f32-to-bf16",
907
+ "nbytes": 11534336,
908
+ "byteOffset": 0
909
+ },
910
+ {
911
+ "name": "model.layers.7.mlp.gate_up_proj.q_scale",
912
+ "shape": [
913
+ 11264,
914
+ 64
915
+ ],
916
+ "dtype": "bfloat16",
917
+ "format": "raw",
918
+ "nbytes": 1441792,
919
+ "byteOffset": 11534336
920
+ },
921
+ {
922
+ "name": "model.layers.7.mlp.down_proj.q_weight",
923
+ "shape": [
924
+ 2048,
925
+ 704
926
+ ],
927
+ "dtype": "uint32",
928
+ "format": "f32-to-bf16",
929
+ "nbytes": 5767168,
930
+ "byteOffset": 12976128
931
+ },
932
+ {
933
+ "name": "model.layers.7.mlp.down_proj.q_scale",
934
+ "shape": [
935
+ 2048,
936
+ 176
937
+ ],
938
+ "dtype": "bfloat16",
939
+ "format": "raw",
940
+ "nbytes": 720896,
941
+ "byteOffset": 18743296
942
+ },
943
+ {
944
+ "name": "model.layers.7.input_layernorm.weight",
945
+ "shape": [
946
+ 2048
947
+ ],
948
+ "dtype": "bfloat16",
949
+ "format": "raw",
950
+ "nbytes": 4096,
951
+ "byteOffset": 19464192
952
+ },
953
+ {
954
+ "name": "model.layers.7.post_attention_layernorm.weight",
955
+ "shape": [
956
+ 2048
957
+ ],
958
+ "dtype": "bfloat16",
959
+ "format": "raw",
960
+ "nbytes": 4096,
961
+ "byteOffset": 19468288
962
+ },
963
+ {
964
+ "name": "model.layers.8.self_attn.qkv_proj.q_weight",
965
+ "shape": [
966
+ 2560,
967
+ 256
968
+ ],
969
+ "dtype": "uint32",
970
+ "format": "f32-to-bf16",
971
+ "nbytes": 2621440,
972
+ "byteOffset": 19472384
973
+ },
974
+ {
975
+ "name": "model.layers.8.self_attn.qkv_proj.q_scale",
976
+ "shape": [
977
+ 2560,
978
+ 64
979
+ ],
980
+ "dtype": "bfloat16",
981
+ "format": "raw",
982
+ "nbytes": 327680,
983
+ "byteOffset": 22093824
984
+ },
985
+ {
986
+ "name": "model.layers.8.self_attn.o_proj.q_weight",
987
+ "shape": [
988
+ 2048,
989
+ 256
990
+ ],
991
+ "dtype": "uint32",
992
+ "format": "f32-to-bf16",
993
+ "nbytes": 2097152,
994
+ "byteOffset": 22421504
995
+ },
996
+ {
997
+ "name": "model.layers.8.self_attn.o_proj.q_scale",
998
+ "shape": [
999
+ 2048,
1000
+ 64
1001
+ ],
1002
+ "dtype": "bfloat16",
1003
+ "format": "raw",
1004
+ "nbytes": 262144,
1005
+ "byteOffset": 24518656
1006
+ }
1007
+ ],
1008
+ "md5sum": "ab488c2050fdfe91cd78eb473120e649"
1009
+ },
1010
+ {
1011
+ "dataPath": "params_shard_9.bin",
1012
+ "format": "raw-shard",
1013
+ "nbytes": 24780800,
1014
+ "records": [
1015
+ {
1016
+ "name": "model.layers.8.mlp.gate_up_proj.q_weight",
1017
+ "shape": [
1018
+ 11264,
1019
+ 256
1020
+ ],
1021
+ "dtype": "uint32",
1022
+ "format": "f32-to-bf16",
1023
+ "nbytes": 11534336,
1024
+ "byteOffset": 0
1025
+ },
1026
+ {
1027
+ "name": "model.layers.8.mlp.gate_up_proj.q_scale",
1028
+ "shape": [
1029
+ 11264,
1030
+ 64
1031
+ ],
1032
+ "dtype": "bfloat16",
1033
+ "format": "raw",
1034
+ "nbytes": 1441792,
1035
+ "byteOffset": 11534336
1036
+ },
1037
+ {
1038
+ "name": "model.layers.8.mlp.down_proj.q_weight",
1039
+ "shape": [
1040
+ 2048,
1041
+ 704
1042
+ ],
1043
+ "dtype": "uint32",
1044
+ "format": "f32-to-bf16",
1045
+ "nbytes": 5767168,
1046
+ "byteOffset": 12976128
1047
+ },
1048
+ {
1049
+ "name": "model.layers.8.mlp.down_proj.q_scale",
1050
+ "shape": [
1051
+ 2048,
1052
+ 176
1053
+ ],
1054
+ "dtype": "bfloat16",
1055
+ "format": "raw",
1056
+ "nbytes": 720896,
1057
+ "byteOffset": 18743296
1058
+ },
1059
+ {
1060
+ "name": "model.layers.8.input_layernorm.weight",
1061
+ "shape": [
1062
+ 2048
1063
+ ],
1064
+ "dtype": "bfloat16",
1065
+ "format": "raw",
1066
+ "nbytes": 4096,
1067
+ "byteOffset": 19464192
1068
+ },
1069
+ {
1070
+ "name": "model.layers.8.post_attention_layernorm.weight",
1071
+ "shape": [
1072
+ 2048
1073
+ ],
1074
+ "dtype": "bfloat16",
1075
+ "format": "raw",
1076
+ "nbytes": 4096,
1077
+ "byteOffset": 19468288
1078
+ },
1079
+ {
1080
+ "name": "model.layers.9.self_attn.qkv_proj.q_weight",
1081
+ "shape": [
1082
+ 2560,
1083
+ 256
1084
+ ],
1085
+ "dtype": "uint32",
1086
+ "format": "f32-to-bf16",
1087
+ "nbytes": 2621440,
1088
+ "byteOffset": 19472384
1089
+ },
1090
+ {
1091
+ "name": "model.layers.9.self_attn.qkv_proj.q_scale",
1092
+ "shape": [
1093
+ 2560,
1094
+ 64
1095
+ ],
1096
+ "dtype": "bfloat16",
1097
+ "format": "raw",
1098
+ "nbytes": 327680,
1099
+ "byteOffset": 22093824
1100
+ },
1101
+ {
1102
+ "name": "model.layers.9.self_attn.o_proj.q_weight",
1103
+ "shape": [
1104
+ 2048,
1105
+ 256
1106
+ ],
1107
+ "dtype": "uint32",
1108
+ "format": "f32-to-bf16",
1109
+ "nbytes": 2097152,
1110
+ "byteOffset": 22421504
1111
+ },
1112
+ {
1113
+ "name": "model.layers.9.self_attn.o_proj.q_scale",
1114
+ "shape": [
1115
+ 2048,
1116
+ 64
1117
+ ],
1118
+ "dtype": "bfloat16",
1119
+ "format": "raw",
1120
+ "nbytes": 262144,
1121
+ "byteOffset": 24518656
1122
+ }
1123
+ ],
1124
+ "md5sum": "ff443a9a0d0db2a220f332b8c6f4fa6a"
1125
+ },
1126
+ {
1127
+ "dataPath": "params_shard_10.bin",
1128
+ "format": "raw-shard",
1129
+ "nbytes": 24780800,
1130
+ "records": [
1131
+ {
1132
+ "name": "model.layers.9.mlp.gate_up_proj.q_weight",
1133
+ "shape": [
1134
+ 11264,
1135
+ 256
1136
+ ],
1137
+ "dtype": "uint32",
1138
+ "format": "f32-to-bf16",
1139
+ "nbytes": 11534336,
1140
+ "byteOffset": 0
1141
+ },
1142
+ {
1143
+ "name": "model.layers.9.mlp.gate_up_proj.q_scale",
1144
+ "shape": [
1145
+ 11264,
1146
+ 64
1147
+ ],
1148
+ "dtype": "bfloat16",
1149
+ "format": "raw",
1150
+ "nbytes": 1441792,
1151
+ "byteOffset": 11534336
1152
+ },
1153
+ {
1154
+ "name": "model.layers.9.mlp.down_proj.q_weight",
1155
+ "shape": [
1156
+ 2048,
1157
+ 704
1158
+ ],
1159
+ "dtype": "uint32",
1160
+ "format": "f32-to-bf16",
1161
+ "nbytes": 5767168,
1162
+ "byteOffset": 12976128
1163
+ },
1164
+ {
1165
+ "name": "model.layers.9.mlp.down_proj.q_scale",
1166
+ "shape": [
1167
+ 2048,
1168
+ 176
1169
+ ],
1170
+ "dtype": "bfloat16",
1171
+ "format": "raw",
1172
+ "nbytes": 720896,
1173
+ "byteOffset": 18743296
1174
+ },
1175
+ {
1176
+ "name": "model.layers.9.input_layernorm.weight",
1177
+ "shape": [
1178
+ 2048
1179
+ ],
1180
+ "dtype": "bfloat16",
1181
+ "format": "raw",
1182
+ "nbytes": 4096,
1183
+ "byteOffset": 19464192
1184
+ },
1185
+ {
1186
+ "name": "model.layers.9.post_attention_layernorm.weight",
1187
+ "shape": [
1188
+ 2048
1189
+ ],
1190
+ "dtype": "bfloat16",
1191
+ "format": "raw",
1192
+ "nbytes": 4096,
1193
+ "byteOffset": 19468288
1194
+ },
1195
+ {
1196
+ "name": "model.layers.10.self_attn.qkv_proj.q_weight",
1197
+ "shape": [
1198
+ 2560,
1199
+ 256
1200
+ ],
1201
+ "dtype": "uint32",
1202
+ "format": "f32-to-bf16",
1203
+ "nbytes": 2621440,
1204
+ "byteOffset": 19472384
1205
+ },
1206
+ {
1207
+ "name": "model.layers.10.self_attn.qkv_proj.q_scale",
1208
+ "shape": [
1209
+ 2560,
1210
+ 64
1211
+ ],
1212
+ "dtype": "bfloat16",
1213
+ "format": "raw",
1214
+ "nbytes": 327680,
1215
+ "byteOffset": 22093824
1216
+ },
1217
+ {
1218
+ "name": "model.layers.10.self_attn.o_proj.q_weight",
1219
+ "shape": [
1220
+ 2048,
1221
+ 256
1222
+ ],
1223
+ "dtype": "uint32",
1224
+ "format": "f32-to-bf16",
1225
+ "nbytes": 2097152,
1226
+ "byteOffset": 22421504
1227
+ },
1228
+ {
1229
+ "name": "model.layers.10.self_attn.o_proj.q_scale",
1230
+ "shape": [
1231
+ 2048,
1232
+ 64
1233
+ ],
1234
+ "dtype": "bfloat16",
1235
+ "format": "raw",
1236
+ "nbytes": 262144,
1237
+ "byteOffset": 24518656
1238
+ }
1239
+ ],
1240
+ "md5sum": "b3e98ffeb6a1180a5ec8ab7fa89b15f1"
1241
+ },
1242
+ {
1243
+ "dataPath": "params_shard_11.bin",
1244
+ "format": "raw-shard",
1245
+ "nbytes": 24780800,
1246
+ "records": [
1247
+ {
1248
+ "name": "model.layers.10.mlp.gate_up_proj.q_weight",
1249
+ "shape": [
1250
+ 11264,
1251
+ 256
1252
+ ],
1253
+ "dtype": "uint32",
1254
+ "format": "f32-to-bf16",
1255
+ "nbytes": 11534336,
1256
+ "byteOffset": 0
1257
+ },
1258
+ {
1259
+ "name": "model.layers.10.mlp.gate_up_proj.q_scale",
1260
+ "shape": [
1261
+ 11264,
1262
+ 64
1263
+ ],
1264
+ "dtype": "bfloat16",
1265
+ "format": "raw",
1266
+ "nbytes": 1441792,
1267
+ "byteOffset": 11534336
1268
+ },
1269
+ {
1270
+ "name": "model.layers.10.mlp.down_proj.q_weight",
1271
+ "shape": [
1272
+ 2048,
1273
+ 704
1274
+ ],
1275
+ "dtype": "uint32",
1276
+ "format": "f32-to-bf16",
1277
+ "nbytes": 5767168,
1278
+ "byteOffset": 12976128
1279
+ },
1280
+ {
1281
+ "name": "model.layers.10.mlp.down_proj.q_scale",
1282
+ "shape": [
1283
+ 2048,
1284
+ 176
1285
+ ],
1286
+ "dtype": "bfloat16",
1287
+ "format": "raw",
1288
+ "nbytes": 720896,
1289
+ "byteOffset": 18743296
1290
+ },
1291
+ {
1292
+ "name": "model.layers.10.input_layernorm.weight",
1293
+ "shape": [
1294
+ 2048
1295
+ ],
1296
+ "dtype": "bfloat16",
1297
+ "format": "raw",
1298
+ "nbytes": 4096,
1299
+ "byteOffset": 19464192
1300
+ },
1301
+ {
1302
+ "name": "model.layers.10.post_attention_layernorm.weight",
1303
+ "shape": [
1304
+ 2048
1305
+ ],
1306
+ "dtype": "bfloat16",
1307
+ "format": "raw",
1308
+ "nbytes": 4096,
1309
+ "byteOffset": 19468288
1310
+ },
1311
+ {
1312
+ "name": "model.layers.11.self_attn.qkv_proj.q_weight",
1313
+ "shape": [
1314
+ 2560,
1315
+ 256
1316
+ ],
1317
+ "dtype": "uint32",
1318
+ "format": "f32-to-bf16",
1319
+ "nbytes": 2621440,
1320
+ "byteOffset": 19472384
1321
+ },
1322
+ {
1323
+ "name": "model.layers.11.self_attn.qkv_proj.q_scale",
1324
+ "shape": [
1325
+ 2560,
1326
+ 64
1327
+ ],
1328
+ "dtype": "bfloat16",
1329
+ "format": "raw",
1330
+ "nbytes": 327680,
1331
+ "byteOffset": 22093824
1332
+ },
1333
+ {
1334
+ "name": "model.layers.11.self_attn.o_proj.q_weight",
1335
+ "shape": [
1336
+ 2048,
1337
+ 256
1338
+ ],
1339
+ "dtype": "uint32",
1340
+ "format": "f32-to-bf16",
1341
+ "nbytes": 2097152,
1342
+ "byteOffset": 22421504
1343
+ },
1344
+ {
1345
+ "name": "model.layers.11.self_attn.o_proj.q_scale",
1346
+ "shape": [
1347
+ 2048,
1348
+ 64
1349
+ ],
1350
+ "dtype": "bfloat16",
1351
+ "format": "raw",
1352
+ "nbytes": 262144,
1353
+ "byteOffset": 24518656
1354
+ }
1355
+ ],
1356
+ "md5sum": "466137e0af99496138348a221df36f30"
1357
+ },
1358
+ {
1359
+ "dataPath": "params_shard_12.bin",
1360
+ "format": "raw-shard",
1361
+ "nbytes": 24780800,
1362
+ "records": [
1363
+ {
1364
+ "name": "model.layers.11.mlp.gate_up_proj.q_weight",
1365
+ "shape": [
1366
+ 11264,
1367
+ 256
1368
+ ],
1369
+ "dtype": "uint32",
1370
+ "format": "f32-to-bf16",
1371
+ "nbytes": 11534336,
1372
+ "byteOffset": 0
1373
+ },
1374
+ {
1375
+ "name": "model.layers.11.mlp.gate_up_proj.q_scale",
1376
+ "shape": [
1377
+ 11264,
1378
+ 64
1379
+ ],
1380
+ "dtype": "bfloat16",
1381
+ "format": "raw",
1382
+ "nbytes": 1441792,
1383
+ "byteOffset": 11534336
1384
+ },
1385
+ {
1386
+ "name": "model.layers.11.mlp.down_proj.q_weight",
1387
+ "shape": [
1388
+ 2048,
1389
+ 704
1390
+ ],
1391
+ "dtype": "uint32",
1392
+ "format": "f32-to-bf16",
1393
+ "nbytes": 5767168,
1394
+ "byteOffset": 12976128
1395
+ },
1396
+ {
1397
+ "name": "model.layers.11.mlp.down_proj.q_scale",
1398
+ "shape": [
1399
+ 2048,
1400
+ 176
1401
+ ],
1402
+ "dtype": "bfloat16",
1403
+ "format": "raw",
1404
+ "nbytes": 720896,
1405
+ "byteOffset": 18743296
1406
+ },
1407
+ {
1408
+ "name": "model.layers.11.input_layernorm.weight",
1409
+ "shape": [
1410
+ 2048
1411
+ ],
1412
+ "dtype": "bfloat16",
1413
+ "format": "raw",
1414
+ "nbytes": 4096,
1415
+ "byteOffset": 19464192
1416
+ },
1417
+ {
1418
+ "name": "model.layers.11.post_attention_layernorm.weight",
1419
+ "shape": [
1420
+ 2048
1421
+ ],
1422
+ "dtype": "bfloat16",
1423
+ "format": "raw",
1424
+ "nbytes": 4096,
1425
+ "byteOffset": 19468288
1426
+ },
1427
+ {
1428
+ "name": "model.layers.12.self_attn.qkv_proj.q_weight",
1429
+ "shape": [
1430
+ 2560,
1431
+ 256
1432
+ ],
1433
+ "dtype": "uint32",
1434
+ "format": "f32-to-bf16",
1435
+ "nbytes": 2621440,
1436
+ "byteOffset": 19472384
1437
+ },
1438
+ {
1439
+ "name": "model.layers.12.self_attn.qkv_proj.q_scale",
1440
+ "shape": [
1441
+ 2560,
1442
+ 64
1443
+ ],
1444
+ "dtype": "bfloat16",
1445
+ "format": "raw",
1446
+ "nbytes": 327680,
1447
+ "byteOffset": 22093824
1448
+ },
1449
+ {
1450
+ "name": "model.layers.12.self_attn.o_proj.q_weight",
1451
+ "shape": [
1452
+ 2048,
1453
+ 256
1454
+ ],
1455
+ "dtype": "uint32",
1456
+ "format": "f32-to-bf16",
1457
+ "nbytes": 2097152,
1458
+ "byteOffset": 22421504
1459
+ },
1460
+ {
1461
+ "name": "model.layers.12.self_attn.o_proj.q_scale",
1462
+ "shape": [
1463
+ 2048,
1464
+ 64
1465
+ ],
1466
+ "dtype": "bfloat16",
1467
+ "format": "raw",
1468
+ "nbytes": 262144,
1469
+ "byteOffset": 24518656
1470
+ }
1471
+ ],
1472
+ "md5sum": "8097c1ddc6ac569ae089e4e1735709d1"
1473
+ },
1474
+ {
1475
+ "dataPath": "params_shard_13.bin",
1476
+ "format": "raw-shard",
1477
+ "nbytes": 24780800,
1478
+ "records": [
1479
+ {
1480
+ "name": "model.layers.12.mlp.gate_up_proj.q_weight",
1481
+ "shape": [
1482
+ 11264,
1483
+ 256
1484
+ ],
1485
+ "dtype": "uint32",
1486
+ "format": "f32-to-bf16",
1487
+ "nbytes": 11534336,
1488
+ "byteOffset": 0
1489
+ },
1490
+ {
1491
+ "name": "model.layers.12.mlp.gate_up_proj.q_scale",
1492
+ "shape": [
1493
+ 11264,
1494
+ 64
1495
+ ],
1496
+ "dtype": "bfloat16",
1497
+ "format": "raw",
1498
+ "nbytes": 1441792,
1499
+ "byteOffset": 11534336
1500
+ },
1501
+ {
1502
+ "name": "model.layers.12.mlp.down_proj.q_weight",
1503
+ "shape": [
1504
+ 2048,
1505
+ 704
1506
+ ],
1507
+ "dtype": "uint32",
1508
+ "format": "f32-to-bf16",
1509
+ "nbytes": 5767168,
1510
+ "byteOffset": 12976128
1511
+ },
1512
+ {
1513
+ "name": "model.layers.12.mlp.down_proj.q_scale",
1514
+ "shape": [
1515
+ 2048,
1516
+ 176
1517
+ ],
1518
+ "dtype": "bfloat16",
1519
+ "format": "raw",
1520
+ "nbytes": 720896,
1521
+ "byteOffset": 18743296
1522
+ },
1523
+ {
1524
+ "name": "model.layers.12.input_layernorm.weight",
1525
+ "shape": [
1526
+ 2048
1527
+ ],
1528
+ "dtype": "bfloat16",
1529
+ "format": "raw",
1530
+ "nbytes": 4096,
1531
+ "byteOffset": 19464192
1532
+ },
1533
+ {
1534
+ "name": "model.layers.12.post_attention_layernorm.weight",
1535
+ "shape": [
1536
+ 2048
1537
+ ],
1538
+ "dtype": "bfloat16",
1539
+ "format": "raw",
1540
+ "nbytes": 4096,
1541
+ "byteOffset": 19468288
1542
+ },
1543
+ {
1544
+ "name": "model.layers.13.self_attn.qkv_proj.q_weight",
1545
+ "shape": [
1546
+ 2560,
1547
+ 256
1548
+ ],
1549
+ "dtype": "uint32",
1550
+ "format": "f32-to-bf16",
1551
+ "nbytes": 2621440,
1552
+ "byteOffset": 19472384
1553
+ },
1554
+ {
1555
+ "name": "model.layers.13.self_attn.qkv_proj.q_scale",
1556
+ "shape": [
1557
+ 2560,
1558
+ 64
1559
+ ],
1560
+ "dtype": "bfloat16",
1561
+ "format": "raw",
1562
+ "nbytes": 327680,
1563
+ "byteOffset": 22093824
1564
+ },
1565
+ {
1566
+ "name": "model.layers.13.self_attn.o_proj.q_weight",
1567
+ "shape": [
1568
+ 2048,
1569
+ 256
1570
+ ],
1571
+ "dtype": "uint32",
1572
+ "format": "f32-to-bf16",
1573
+ "nbytes": 2097152,
1574
+ "byteOffset": 22421504
1575
+ },
1576
+ {
1577
+ "name": "model.layers.13.self_attn.o_proj.q_scale",
1578
+ "shape": [
1579
+ 2048,
1580
+ 64
1581
+ ],
1582
+ "dtype": "bfloat16",
1583
+ "format": "raw",
1584
+ "nbytes": 262144,
1585
+ "byteOffset": 24518656
1586
+ }
1587
+ ],
1588
+ "md5sum": "cb4d8ac3523d4a78c40f6e05e6a75a3f"
1589
+ },
1590
+ {
1591
+ "dataPath": "params_shard_14.bin",
1592
+ "format": "raw-shard",
1593
+ "nbytes": 24780800,
1594
+ "records": [
1595
+ {
1596
+ "name": "model.layers.13.mlp.gate_up_proj.q_weight",
1597
+ "shape": [
1598
+ 11264,
1599
+ 256
1600
+ ],
1601
+ "dtype": "uint32",
1602
+ "format": "f32-to-bf16",
1603
+ "nbytes": 11534336,
1604
+ "byteOffset": 0
1605
+ },
1606
+ {
1607
+ "name": "model.layers.13.mlp.gate_up_proj.q_scale",
1608
+ "shape": [
1609
+ 11264,
1610
+ 64
1611
+ ],
1612
+ "dtype": "bfloat16",
1613
+ "format": "raw",
1614
+ "nbytes": 1441792,
1615
+ "byteOffset": 11534336
1616
+ },
1617
+ {
1618
+ "name": "model.layers.13.mlp.down_proj.q_weight",
1619
+ "shape": [
1620
+ 2048,
1621
+ 704
1622
+ ],
1623
+ "dtype": "uint32",
1624
+ "format": "f32-to-bf16",
1625
+ "nbytes": 5767168,
1626
+ "byteOffset": 12976128
1627
+ },
1628
+ {
1629
+ "name": "model.layers.13.mlp.down_proj.q_scale",
1630
+ "shape": [
1631
+ 2048,
1632
+ 176
1633
+ ],
1634
+ "dtype": "bfloat16",
1635
+ "format": "raw",
1636
+ "nbytes": 720896,
1637
+ "byteOffset": 18743296
1638
+ },
1639
+ {
1640
+ "name": "model.layers.13.input_layernorm.weight",
1641
+ "shape": [
1642
+ 2048
1643
+ ],
1644
+ "dtype": "bfloat16",
1645
+ "format": "raw",
1646
+ "nbytes": 4096,
1647
+ "byteOffset": 19464192
1648
+ },
1649
+ {
1650
+ "name": "model.layers.13.post_attention_layernorm.weight",
1651
+ "shape": [
1652
+ 2048
1653
+ ],
1654
+ "dtype": "bfloat16",
1655
+ "format": "raw",
1656
+ "nbytes": 4096,
1657
+ "byteOffset": 19468288
1658
+ },
1659
+ {
1660
+ "name": "model.layers.14.self_attn.qkv_proj.q_weight",
1661
+ "shape": [
1662
+ 2560,
1663
+ 256
1664
+ ],
1665
+ "dtype": "uint32",
1666
+ "format": "f32-to-bf16",
1667
+ "nbytes": 2621440,
1668
+ "byteOffset": 19472384
1669
+ },
1670
+ {
1671
+ "name": "model.layers.14.self_attn.qkv_proj.q_scale",
1672
+ "shape": [
1673
+ 2560,
1674
+ 64
1675
+ ],
1676
+ "dtype": "bfloat16",
1677
+ "format": "raw",
1678
+ "nbytes": 327680,
1679
+ "byteOffset": 22093824
1680
+ },
1681
+ {
1682
+ "name": "model.layers.14.self_attn.o_proj.q_weight",
1683
+ "shape": [
1684
+ 2048,
1685
+ 256
1686
+ ],
1687
+ "dtype": "uint32",
1688
+ "format": "f32-to-bf16",
1689
+ "nbytes": 2097152,
1690
+ "byteOffset": 22421504
1691
+ },
1692
+ {
1693
+ "name": "model.layers.14.self_attn.o_proj.q_scale",
1694
+ "shape": [
1695
+ 2048,
1696
+ 64
1697
+ ],
1698
+ "dtype": "bfloat16",
1699
+ "format": "raw",
1700
+ "nbytes": 262144,
1701
+ "byteOffset": 24518656
1702
+ }
1703
+ ],
1704
+ "md5sum": "97fec4eb81b0c8ba326d6bd7c2d0d01e"
1705
+ },
1706
+ {
1707
+ "dataPath": "params_shard_15.bin",
1708
+ "format": "raw-shard",
1709
+ "nbytes": 24780800,
1710
+ "records": [
1711
+ {
1712
+ "name": "model.layers.14.mlp.gate_up_proj.q_weight",
1713
+ "shape": [
1714
+ 11264,
1715
+ 256
1716
+ ],
1717
+ "dtype": "uint32",
1718
+ "format": "f32-to-bf16",
1719
+ "nbytes": 11534336,
1720
+ "byteOffset": 0
1721
+ },
1722
+ {
1723
+ "name": "model.layers.14.mlp.gate_up_proj.q_scale",
1724
+ "shape": [
1725
+ 11264,
1726
+ 64
1727
+ ],
1728
+ "dtype": "bfloat16",
1729
+ "format": "raw",
1730
+ "nbytes": 1441792,
1731
+ "byteOffset": 11534336
1732
+ },
1733
+ {
1734
+ "name": "model.layers.14.mlp.down_proj.q_weight",
1735
+ "shape": [
1736
+ 2048,
1737
+ 704
1738
+ ],
1739
+ "dtype": "uint32",
1740
+ "format": "f32-to-bf16",
1741
+ "nbytes": 5767168,
1742
+ "byteOffset": 12976128
1743
+ },
1744
+ {
1745
+ "name": "model.layers.14.mlp.down_proj.q_scale",
1746
+ "shape": [
1747
+ 2048,
1748
+ 176
1749
+ ],
1750
+ "dtype": "bfloat16",
1751
+ "format": "raw",
1752
+ "nbytes": 720896,
1753
+ "byteOffset": 18743296
1754
+ },
1755
+ {
1756
+ "name": "model.layers.14.input_layernorm.weight",
1757
+ "shape": [
1758
+ 2048
1759
+ ],
1760
+ "dtype": "bfloat16",
1761
+ "format": "raw",
1762
+ "nbytes": 4096,
1763
+ "byteOffset": 19464192
1764
+ },
1765
+ {
1766
+ "name": "model.layers.14.post_attention_layernorm.weight",
1767
+ "shape": [
1768
+ 2048
1769
+ ],
1770
+ "dtype": "bfloat16",
1771
+ "format": "raw",
1772
+ "nbytes": 4096,
1773
+ "byteOffset": 19468288
1774
+ },
1775
+ {
1776
+ "name": "model.layers.15.self_attn.qkv_proj.q_weight",
1777
+ "shape": [
1778
+ 2560,
1779
+ 256
1780
+ ],
1781
+ "dtype": "uint32",
1782
+ "format": "f32-to-bf16",
1783
+ "nbytes": 2621440,
1784
+ "byteOffset": 19472384
1785
+ },
1786
+ {
1787
+ "name": "model.layers.15.self_attn.qkv_proj.q_scale",
1788
+ "shape": [
1789
+ 2560,
1790
+ 64
1791
+ ],
1792
+ "dtype": "bfloat16",
1793
+ "format": "raw",
1794
+ "nbytes": 327680,
1795
+ "byteOffset": 22093824
1796
+ },
1797
+ {
1798
+ "name": "model.layers.15.self_attn.o_proj.q_weight",
1799
+ "shape": [
1800
+ 2048,
1801
+ 256
1802
+ ],
1803
+ "dtype": "uint32",
1804
+ "format": "f32-to-bf16",
1805
+ "nbytes": 2097152,
1806
+ "byteOffset": 22421504
1807
+ },
1808
+ {
1809
+ "name": "model.layers.15.self_attn.o_proj.q_scale",
1810
+ "shape": [
1811
+ 2048,
1812
+ 64
1813
+ ],
1814
+ "dtype": "bfloat16",
1815
+ "format": "raw",
1816
+ "nbytes": 262144,
1817
+ "byteOffset": 24518656
1818
+ }
1819
+ ],
1820
+ "md5sum": "5942e747ef08c2f1f6aee113d2fdb843"
1821
+ },
1822
+ {
1823
+ "dataPath": "params_shard_16.bin",
1824
+ "format": "raw-shard",
1825
+ "nbytes": 24780800,
1826
+ "records": [
1827
+ {
1828
+ "name": "model.layers.15.mlp.gate_up_proj.q_weight",
1829
+ "shape": [
1830
+ 11264,
1831
+ 256
1832
+ ],
1833
+ "dtype": "uint32",
1834
+ "format": "f32-to-bf16",
1835
+ "nbytes": 11534336,
1836
+ "byteOffset": 0
1837
+ },
1838
+ {
1839
+ "name": "model.layers.15.mlp.gate_up_proj.q_scale",
1840
+ "shape": [
1841
+ 11264,
1842
+ 64
1843
+ ],
1844
+ "dtype": "bfloat16",
1845
+ "format": "raw",
1846
+ "nbytes": 1441792,
1847
+ "byteOffset": 11534336
1848
+ },
1849
+ {
1850
+ "name": "model.layers.15.mlp.down_proj.q_weight",
1851
+ "shape": [
1852
+ 2048,
1853
+ 704
1854
+ ],
1855
+ "dtype": "uint32",
1856
+ "format": "f32-to-bf16",
1857
+ "nbytes": 5767168,
1858
+ "byteOffset": 12976128
1859
+ },
1860
+ {
1861
+ "name": "model.layers.15.mlp.down_proj.q_scale",
1862
+ "shape": [
1863
+ 2048,
1864
+ 176
1865
+ ],
1866
+ "dtype": "bfloat16",
1867
+ "format": "raw",
1868
+ "nbytes": 720896,
1869
+ "byteOffset": 18743296
1870
+ },
1871
+ {
1872
+ "name": "model.layers.15.input_layernorm.weight",
1873
+ "shape": [
1874
+ 2048
1875
+ ],
1876
+ "dtype": "bfloat16",
1877
+ "format": "raw",
1878
+ "nbytes": 4096,
1879
+ "byteOffset": 19464192
1880
+ },
1881
+ {
1882
+ "name": "model.layers.15.post_attention_layernorm.weight",
1883
+ "shape": [
1884
+ 2048
1885
+ ],
1886
+ "dtype": "bfloat16",
1887
+ "format": "raw",
1888
+ "nbytes": 4096,
1889
+ "byteOffset": 19468288
1890
+ },
1891
+ {
1892
+ "name": "model.layers.16.self_attn.qkv_proj.q_weight",
1893
+ "shape": [
1894
+ 2560,
1895
+ 256
1896
+ ],
1897
+ "dtype": "uint32",
1898
+ "format": "f32-to-bf16",
1899
+ "nbytes": 2621440,
1900
+ "byteOffset": 19472384
1901
+ },
1902
+ {
1903
+ "name": "model.layers.16.self_attn.qkv_proj.q_scale",
1904
+ "shape": [
1905
+ 2560,
1906
+ 64
1907
+ ],
1908
+ "dtype": "bfloat16",
1909
+ "format": "raw",
1910
+ "nbytes": 327680,
1911
+ "byteOffset": 22093824
1912
+ },
1913
+ {
1914
+ "name": "model.layers.16.self_attn.o_proj.q_weight",
1915
+ "shape": [
1916
+ 2048,
1917
+ 256
1918
+ ],
1919
+ "dtype": "uint32",
1920
+ "format": "f32-to-bf16",
1921
+ "nbytes": 2097152,
1922
+ "byteOffset": 22421504
1923
+ },
1924
+ {
1925
+ "name": "model.layers.16.self_attn.o_proj.q_scale",
1926
+ "shape": [
1927
+ 2048,
1928
+ 64
1929
+ ],
1930
+ "dtype": "bfloat16",
1931
+ "format": "raw",
1932
+ "nbytes": 262144,
1933
+ "byteOffset": 24518656
1934
+ }
1935
+ ],
1936
+ "md5sum": "fa1b0776d20499dcf35ca524e65addcd"
1937
+ },
1938
+ {
1939
+ "dataPath": "params_shard_17.bin",
1940
+ "format": "raw-shard",
1941
+ "nbytes": 24780800,
1942
+ "records": [
1943
+ {
1944
+ "name": "model.layers.16.mlp.gate_up_proj.q_weight",
1945
+ "shape": [
1946
+ 11264,
1947
+ 256
1948
+ ],
1949
+ "dtype": "uint32",
1950
+ "format": "f32-to-bf16",
1951
+ "nbytes": 11534336,
1952
+ "byteOffset": 0
1953
+ },
1954
+ {
1955
+ "name": "model.layers.16.mlp.gate_up_proj.q_scale",
1956
+ "shape": [
1957
+ 11264,
1958
+ 64
1959
+ ],
1960
+ "dtype": "bfloat16",
1961
+ "format": "raw",
1962
+ "nbytes": 1441792,
1963
+ "byteOffset": 11534336
1964
+ },
1965
+ {
1966
+ "name": "model.layers.16.mlp.down_proj.q_weight",
1967
+ "shape": [
1968
+ 2048,
1969
+ 704
1970
+ ],
1971
+ "dtype": "uint32",
1972
+ "format": "f32-to-bf16",
1973
+ "nbytes": 5767168,
1974
+ "byteOffset": 12976128
1975
+ },
1976
+ {
1977
+ "name": "model.layers.16.mlp.down_proj.q_scale",
1978
+ "shape": [
1979
+ 2048,
1980
+ 176
1981
+ ],
1982
+ "dtype": "bfloat16",
1983
+ "format": "raw",
1984
+ "nbytes": 720896,
1985
+ "byteOffset": 18743296
1986
+ },
1987
+ {
1988
+ "name": "model.layers.16.input_layernorm.weight",
1989
+ "shape": [
1990
+ 2048
1991
+ ],
1992
+ "dtype": "bfloat16",
1993
+ "format": "raw",
1994
+ "nbytes": 4096,
1995
+ "byteOffset": 19464192
1996
+ },
1997
+ {
1998
+ "name": "model.layers.16.post_attention_layernorm.weight",
1999
+ "shape": [
2000
+ 2048
2001
+ ],
2002
+ "dtype": "bfloat16",
2003
+ "format": "raw",
2004
+ "nbytes": 4096,
2005
+ "byteOffset": 19468288
2006
+ },
2007
+ {
2008
+ "name": "model.layers.17.self_attn.qkv_proj.q_weight",
2009
+ "shape": [
2010
+ 2560,
2011
+ 256
2012
+ ],
2013
+ "dtype": "uint32",
2014
+ "format": "f32-to-bf16",
2015
+ "nbytes": 2621440,
2016
+ "byteOffset": 19472384
2017
+ },
2018
+ {
2019
+ "name": "model.layers.17.self_attn.qkv_proj.q_scale",
2020
+ "shape": [
2021
+ 2560,
2022
+ 64
2023
+ ],
2024
+ "dtype": "bfloat16",
2025
+ "format": "raw",
2026
+ "nbytes": 327680,
2027
+ "byteOffset": 22093824
2028
+ },
2029
+ {
2030
+ "name": "model.layers.17.self_attn.o_proj.q_weight",
2031
+ "shape": [
2032
+ 2048,
2033
+ 256
2034
+ ],
2035
+ "dtype": "uint32",
2036
+ "format": "f32-to-bf16",
2037
+ "nbytes": 2097152,
2038
+ "byteOffset": 22421504
2039
+ },
2040
+ {
2041
+ "name": "model.layers.17.self_attn.o_proj.q_scale",
2042
+ "shape": [
2043
+ 2048,
2044
+ 64
2045
+ ],
2046
+ "dtype": "bfloat16",
2047
+ "format": "raw",
2048
+ "nbytes": 262144,
2049
+ "byteOffset": 24518656
2050
+ }
2051
+ ],
2052
+ "md5sum": "4eec190a3558aad16647bedadf63ad68"
2053
+ },
2054
+ {
2055
+ "dataPath": "params_shard_18.bin",
2056
+ "format": "raw-shard",
2057
+ "nbytes": 24780800,
2058
+ "records": [
2059
+ {
2060
+ "name": "model.layers.17.mlp.gate_up_proj.q_weight",
2061
+ "shape": [
2062
+ 11264,
2063
+ 256
2064
+ ],
2065
+ "dtype": "uint32",
2066
+ "format": "f32-to-bf16",
2067
+ "nbytes": 11534336,
2068
+ "byteOffset": 0
2069
+ },
2070
+ {
2071
+ "name": "model.layers.17.mlp.gate_up_proj.q_scale",
2072
+ "shape": [
2073
+ 11264,
2074
+ 64
2075
+ ],
2076
+ "dtype": "bfloat16",
2077
+ "format": "raw",
2078
+ "nbytes": 1441792,
2079
+ "byteOffset": 11534336
2080
+ },
2081
+ {
2082
+ "name": "model.layers.17.mlp.down_proj.q_weight",
2083
+ "shape": [
2084
+ 2048,
2085
+ 704
2086
+ ],
2087
+ "dtype": "uint32",
2088
+ "format": "f32-to-bf16",
2089
+ "nbytes": 5767168,
2090
+ "byteOffset": 12976128
2091
+ },
2092
+ {
2093
+ "name": "model.layers.17.mlp.down_proj.q_scale",
2094
+ "shape": [
2095
+ 2048,
2096
+ 176
2097
+ ],
2098
+ "dtype": "bfloat16",
2099
+ "format": "raw",
2100
+ "nbytes": 720896,
2101
+ "byteOffset": 18743296
2102
+ },
2103
+ {
2104
+ "name": "model.layers.17.input_layernorm.weight",
2105
+ "shape": [
2106
+ 2048
2107
+ ],
2108
+ "dtype": "bfloat16",
2109
+ "format": "raw",
2110
+ "nbytes": 4096,
2111
+ "byteOffset": 19464192
2112
+ },
2113
+ {
2114
+ "name": "model.layers.17.post_attention_layernorm.weight",
2115
+ "shape": [
2116
+ 2048
2117
+ ],
2118
+ "dtype": "bfloat16",
2119
+ "format": "raw",
2120
+ "nbytes": 4096,
2121
+ "byteOffset": 19468288
2122
+ },
2123
+ {
2124
+ "name": "model.layers.18.self_attn.qkv_proj.q_weight",
2125
+ "shape": [
2126
+ 2560,
2127
+ 256
2128
+ ],
2129
+ "dtype": "uint32",
2130
+ "format": "f32-to-bf16",
2131
+ "nbytes": 2621440,
2132
+ "byteOffset": 19472384
2133
+ },
2134
+ {
2135
+ "name": "model.layers.18.self_attn.qkv_proj.q_scale",
2136
+ "shape": [
2137
+ 2560,
2138
+ 64
2139
+ ],
2140
+ "dtype": "bfloat16",
2141
+ "format": "raw",
2142
+ "nbytes": 327680,
2143
+ "byteOffset": 22093824
2144
+ },
2145
+ {
2146
+ "name": "model.layers.18.self_attn.o_proj.q_weight",
2147
+ "shape": [
2148
+ 2048,
2149
+ 256
2150
+ ],
2151
+ "dtype": "uint32",
2152
+ "format": "f32-to-bf16",
2153
+ "nbytes": 2097152,
2154
+ "byteOffset": 22421504
2155
+ },
2156
+ {
2157
+ "name": "model.layers.18.self_attn.o_proj.q_scale",
2158
+ "shape": [
2159
+ 2048,
2160
+ 64
2161
+ ],
2162
+ "dtype": "bfloat16",
2163
+ "format": "raw",
2164
+ "nbytes": 262144,
2165
+ "byteOffset": 24518656
2166
+ }
2167
+ ],
2168
+ "md5sum": "d34c9a3f1d0cb3900907226eca95af8d"
2169
+ },
2170
+ {
2171
+ "dataPath": "params_shard_19.bin",
2172
+ "format": "raw-shard",
2173
+ "nbytes": 24780800,
2174
+ "records": [
2175
+ {
2176
+ "name": "model.layers.18.mlp.gate_up_proj.q_weight",
2177
+ "shape": [
2178
+ 11264,
2179
+ 256
2180
+ ],
2181
+ "dtype": "uint32",
2182
+ "format": "f32-to-bf16",
2183
+ "nbytes": 11534336,
2184
+ "byteOffset": 0
2185
+ },
2186
+ {
2187
+ "name": "model.layers.18.mlp.gate_up_proj.q_scale",
2188
+ "shape": [
2189
+ 11264,
2190
+ 64
2191
+ ],
2192
+ "dtype": "bfloat16",
2193
+ "format": "raw",
2194
+ "nbytes": 1441792,
2195
+ "byteOffset": 11534336
2196
+ },
2197
+ {
2198
+ "name": "model.layers.18.mlp.down_proj.q_weight",
2199
+ "shape": [
2200
+ 2048,
2201
+ 704
2202
+ ],
2203
+ "dtype": "uint32",
2204
+ "format": "f32-to-bf16",
2205
+ "nbytes": 5767168,
2206
+ "byteOffset": 12976128
2207
+ },
2208
+ {
2209
+ "name": "model.layers.18.mlp.down_proj.q_scale",
2210
+ "shape": [
2211
+ 2048,
2212
+ 176
2213
+ ],
2214
+ "dtype": "bfloat16",
2215
+ "format": "raw",
2216
+ "nbytes": 720896,
2217
+ "byteOffset": 18743296
2218
+ },
2219
+ {
2220
+ "name": "model.layers.18.input_layernorm.weight",
2221
+ "shape": [
2222
+ 2048
2223
+ ],
2224
+ "dtype": "bfloat16",
2225
+ "format": "raw",
2226
+ "nbytes": 4096,
2227
+ "byteOffset": 19464192
2228
+ },
2229
+ {
2230
+ "name": "model.layers.18.post_attention_layernorm.weight",
2231
+ "shape": [
2232
+ 2048
2233
+ ],
2234
+ "dtype": "bfloat16",
2235
+ "format": "raw",
2236
+ "nbytes": 4096,
2237
+ "byteOffset": 19468288
2238
+ },
2239
+ {
2240
+ "name": "model.layers.19.self_attn.qkv_proj.q_weight",
2241
+ "shape": [
2242
+ 2560,
2243
+ 256
2244
+ ],
2245
+ "dtype": "uint32",
2246
+ "format": "f32-to-bf16",
2247
+ "nbytes": 2621440,
2248
+ "byteOffset": 19472384
2249
+ },
2250
+ {
2251
+ "name": "model.layers.19.self_attn.qkv_proj.q_scale",
2252
+ "shape": [
2253
+ 2560,
2254
+ 64
2255
+ ],
2256
+ "dtype": "bfloat16",
2257
+ "format": "raw",
2258
+ "nbytes": 327680,
2259
+ "byteOffset": 22093824
2260
+ },
2261
+ {
2262
+ "name": "model.layers.19.self_attn.o_proj.q_weight",
2263
+ "shape": [
2264
+ 2048,
2265
+ 256
2266
+ ],
2267
+ "dtype": "uint32",
2268
+ "format": "f32-to-bf16",
2269
+ "nbytes": 2097152,
2270
+ "byteOffset": 22421504
2271
+ },
2272
+ {
2273
+ "name": "model.layers.19.self_attn.o_proj.q_scale",
2274
+ "shape": [
2275
+ 2048,
2276
+ 64
2277
+ ],
2278
+ "dtype": "bfloat16",
2279
+ "format": "raw",
2280
+ "nbytes": 262144,
2281
+ "byteOffset": 24518656
2282
+ }
2283
+ ],
2284
+ "md5sum": "829dd454ac1242bc973177bc2248a1f8"
2285
+ },
2286
+ {
2287
+ "dataPath": "params_shard_20.bin",
2288
+ "format": "raw-shard",
2289
+ "nbytes": 24780800,
2290
+ "records": [
2291
+ {
2292
+ "name": "model.layers.19.mlp.gate_up_proj.q_weight",
2293
+ "shape": [
2294
+ 11264,
2295
+ 256
2296
+ ],
2297
+ "dtype": "uint32",
2298
+ "format": "f32-to-bf16",
2299
+ "nbytes": 11534336,
2300
+ "byteOffset": 0
2301
+ },
2302
+ {
2303
+ "name": "model.layers.19.mlp.gate_up_proj.q_scale",
2304
+ "shape": [
2305
+ 11264,
2306
+ 64
2307
+ ],
2308
+ "dtype": "bfloat16",
2309
+ "format": "raw",
2310
+ "nbytes": 1441792,
2311
+ "byteOffset": 11534336
2312
+ },
2313
+ {
2314
+ "name": "model.layers.19.mlp.down_proj.q_weight",
2315
+ "shape": [
2316
+ 2048,
2317
+ 704
2318
+ ],
2319
+ "dtype": "uint32",
2320
+ "format": "f32-to-bf16",
2321
+ "nbytes": 5767168,
2322
+ "byteOffset": 12976128
2323
+ },
2324
+ {
2325
+ "name": "model.layers.19.mlp.down_proj.q_scale",
2326
+ "shape": [
2327
+ 2048,
2328
+ 176
2329
+ ],
2330
+ "dtype": "bfloat16",
2331
+ "format": "raw",
2332
+ "nbytes": 720896,
2333
+ "byteOffset": 18743296
2334
+ },
2335
+ {
2336
+ "name": "model.layers.19.input_layernorm.weight",
2337
+ "shape": [
2338
+ 2048
2339
+ ],
2340
+ "dtype": "bfloat16",
2341
+ "format": "raw",
2342
+ "nbytes": 4096,
2343
+ "byteOffset": 19464192
2344
+ },
2345
+ {
2346
+ "name": "model.layers.19.post_attention_layernorm.weight",
2347
+ "shape": [
2348
+ 2048
2349
+ ],
2350
+ "dtype": "bfloat16",
2351
+ "format": "raw",
2352
+ "nbytes": 4096,
2353
+ "byteOffset": 19468288
2354
+ },
2355
+ {
2356
+ "name": "model.layers.20.self_attn.qkv_proj.q_weight",
2357
+ "shape": [
2358
+ 2560,
2359
+ 256
2360
+ ],
2361
+ "dtype": "uint32",
2362
+ "format": "f32-to-bf16",
2363
+ "nbytes": 2621440,
2364
+ "byteOffset": 19472384
2365
+ },
2366
+ {
2367
+ "name": "model.layers.20.self_attn.qkv_proj.q_scale",
2368
+ "shape": [
2369
+ 2560,
2370
+ 64
2371
+ ],
2372
+ "dtype": "bfloat16",
2373
+ "format": "raw",
2374
+ "nbytes": 327680,
2375
+ "byteOffset": 22093824
2376
+ },
2377
+ {
2378
+ "name": "model.layers.20.self_attn.o_proj.q_weight",
2379
+ "shape": [
2380
+ 2048,
2381
+ 256
2382
+ ],
2383
+ "dtype": "uint32",
2384
+ "format": "f32-to-bf16",
2385
+ "nbytes": 2097152,
2386
+ "byteOffset": 22421504
2387
+ },
2388
+ {
2389
+ "name": "model.layers.20.self_attn.o_proj.q_scale",
2390
+ "shape": [
2391
+ 2048,
2392
+ 64
2393
+ ],
2394
+ "dtype": "bfloat16",
2395
+ "format": "raw",
2396
+ "nbytes": 262144,
2397
+ "byteOffset": 24518656
2398
+ }
2399
+ ],
2400
+ "md5sum": "4b6e5593f9e640846c38651fd89f279c"
2401
+ },
2402
+ {
2403
+ "dataPath": "params_shard_21.bin",
2404
+ "format": "raw-shard",
2405
+ "nbytes": 24780800,
2406
+ "records": [
2407
+ {
2408
+ "name": "model.layers.20.mlp.gate_up_proj.q_weight",
2409
+ "shape": [
2410
+ 11264,
2411
+ 256
2412
+ ],
2413
+ "dtype": "uint32",
2414
+ "format": "f32-to-bf16",
2415
+ "nbytes": 11534336,
2416
+ "byteOffset": 0
2417
+ },
2418
+ {
2419
+ "name": "model.layers.20.mlp.gate_up_proj.q_scale",
2420
+ "shape": [
2421
+ 11264,
2422
+ 64
2423
+ ],
2424
+ "dtype": "bfloat16",
2425
+ "format": "raw",
2426
+ "nbytes": 1441792,
2427
+ "byteOffset": 11534336
2428
+ },
2429
+ {
2430
+ "name": "model.layers.20.mlp.down_proj.q_weight",
2431
+ "shape": [
2432
+ 2048,
2433
+ 704
2434
+ ],
2435
+ "dtype": "uint32",
2436
+ "format": "f32-to-bf16",
2437
+ "nbytes": 5767168,
2438
+ "byteOffset": 12976128
2439
+ },
2440
+ {
2441
+ "name": "model.layers.20.mlp.down_proj.q_scale",
2442
+ "shape": [
2443
+ 2048,
2444
+ 176
2445
+ ],
2446
+ "dtype": "bfloat16",
2447
+ "format": "raw",
2448
+ "nbytes": 720896,
2449
+ "byteOffset": 18743296
2450
+ },
2451
+ {
2452
+ "name": "model.layers.20.input_layernorm.weight",
2453
+ "shape": [
2454
+ 2048
2455
+ ],
2456
+ "dtype": "bfloat16",
2457
+ "format": "raw",
2458
+ "nbytes": 4096,
2459
+ "byteOffset": 19464192
2460
+ },
2461
+ {
2462
+ "name": "model.layers.20.post_attention_layernorm.weight",
2463
+ "shape": [
2464
+ 2048
2465
+ ],
2466
+ "dtype": "bfloat16",
2467
+ "format": "raw",
2468
+ "nbytes": 4096,
2469
+ "byteOffset": 19468288
2470
+ },
2471
+ {
2472
+ "name": "model.layers.21.self_attn.qkv_proj.q_weight",
2473
+ "shape": [
2474
+ 2560,
2475
+ 256
2476
+ ],
2477
+ "dtype": "uint32",
2478
+ "format": "f32-to-bf16",
2479
+ "nbytes": 2621440,
2480
+ "byteOffset": 19472384
2481
+ },
2482
+ {
2483
+ "name": "model.layers.21.self_attn.qkv_proj.q_scale",
2484
+ "shape": [
2485
+ 2560,
2486
+ 64
2487
+ ],
2488
+ "dtype": "bfloat16",
2489
+ "format": "raw",
2490
+ "nbytes": 327680,
2491
+ "byteOffset": 22093824
2492
+ },
2493
+ {
2494
+ "name": "model.layers.21.self_attn.o_proj.q_weight",
2495
+ "shape": [
2496
+ 2048,
2497
+ 256
2498
+ ],
2499
+ "dtype": "uint32",
2500
+ "format": "f32-to-bf16",
2501
+ "nbytes": 2097152,
2502
+ "byteOffset": 22421504
2503
+ },
2504
+ {
2505
+ "name": "model.layers.21.self_attn.o_proj.q_scale",
2506
+ "shape": [
2507
+ 2048,
2508
+ 64
2509
+ ],
2510
+ "dtype": "bfloat16",
2511
+ "format": "raw",
2512
+ "nbytes": 262144,
2513
+ "byteOffset": 24518656
2514
+ }
2515
+ ],
2516
+ "md5sum": "a74f48bc17bc757ea9d4c75002229241"
2517
+ },
2518
+ {
2519
+ "dataPath": "params_shard_22.bin",
2520
+ "format": "raw-shard",
2521
+ "nbytes": 32771072,
2522
+ "records": [
2523
+ {
2524
+ "name": "lm_head.q_weight",
2525
+ "shape": [
2526
+ 32003,
2527
+ 256
2528
+ ],
2529
+ "dtype": "uint32",
2530
+ "format": "f32-to-bf16",
2531
+ "nbytes": 32771072,
2532
+ "byteOffset": 0
2533
+ }
2534
+ ],
2535
+ "md5sum": "9a1c53a4e8ad27cf8d5fb1fcc5c333c2"
2536
+ },
2537
+ {
2538
+ "dataPath": "params_shard_23.bin",
2539
+ "format": "raw-shard",
2540
+ "nbytes": 23572864,
2541
+ "records": [
2542
+ {
2543
+ "name": "model.layers.21.mlp.gate_up_proj.q_weight",
2544
+ "shape": [
2545
+ 11264,
2546
+ 256
2547
+ ],
2548
+ "dtype": "uint32",
2549
+ "format": "f32-to-bf16",
2550
+ "nbytes": 11534336,
2551
+ "byteOffset": 0
2552
+ },
2553
+ {
2554
+ "name": "model.layers.21.mlp.gate_up_proj.q_scale",
2555
+ "shape": [
2556
+ 11264,
2557
+ 64
2558
+ ],
2559
+ "dtype": "bfloat16",
2560
+ "format": "raw",
2561
+ "nbytes": 1441792,
2562
+ "byteOffset": 11534336
2563
+ },
2564
+ {
2565
+ "name": "model.layers.21.mlp.down_proj.q_weight",
2566
+ "shape": [
2567
+ 2048,
2568
+ 704
2569
+ ],
2570
+ "dtype": "uint32",
2571
+ "format": "f32-to-bf16",
2572
+ "nbytes": 5767168,
2573
+ "byteOffset": 12976128
2574
+ },
2575
+ {
2576
+ "name": "model.layers.21.mlp.down_proj.q_scale",
2577
+ "shape": [
2578
+ 2048,
2579
+ 176
2580
+ ],
2581
+ "dtype": "bfloat16",
2582
+ "format": "raw",
2583
+ "nbytes": 720896,
2584
+ "byteOffset": 18743296
2585
+ },
2586
+ {
2587
+ "name": "model.layers.21.input_layernorm.weight",
2588
+ "shape": [
2589
+ 2048
2590
+ ],
2591
+ "dtype": "bfloat16",
2592
+ "format": "raw",
2593
+ "nbytes": 4096,
2594
+ "byteOffset": 19464192
2595
+ },
2596
+ {
2597
+ "name": "model.layers.21.post_attention_layernorm.weight",
2598
+ "shape": [
2599
+ 2048
2600
+ ],
2601
+ "dtype": "bfloat16",
2602
+ "format": "raw",
2603
+ "nbytes": 4096,
2604
+ "byteOffset": 19468288
2605
+ },
2606
+ {
2607
+ "name": "model.norm.weight",
2608
+ "shape": [
2609
+ 2048
2610
+ ],
2611
+ "dtype": "bfloat16",
2612
+ "format": "raw",
2613
+ "nbytes": 4096,
2614
+ "byteOffset": 19472384
2615
+ },
2616
+ {
2617
+ "name": "lm_head.q_scale",
2618
+ "shape": [
2619
+ 32003,
2620
+ 64
2621
+ ],
2622
+ "dtype": "bfloat16",
2623
+ "format": "raw",
2624
+ "nbytes": 4096384,
2625
+ "byteOffset": 19476480
2626
+ }
2627
+ ],
2628
+ "md5sum": "cc2bb39dd7372753ef316485aa555550"
2629
+ }
2630
+ ]
2631
+ }
ndarray-cache.json ADDED
@@ -0,0 +1,2631 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metadata": {
3
+ "ParamSize": 225,
4
+ "ParamBytes": 687848960.0,
5
+ "BitsPerParam": 5.002261984328079
6
+ },
7
+ "records": [
8
+ {
9
+ "dataPath": "params_shard_0.bin",
10
+ "format": "raw-shard",
11
+ "nbytes": 32771072,
12
+ "records": [
13
+ {
14
+ "name": "model.embed_tokens.q_weight",
15
+ "shape": [
16
+ 32003,
17
+ 256
18
+ ],
19
+ "dtype": "uint32",
20
+ "format": "f32-to-bf16",
21
+ "nbytes": 32771072,
22
+ "byteOffset": 0
23
+ }
24
+ ],
25
+ "md5sum": "5d04d3daaea233773d55c738ca893695"
26
+ },
27
+ {
28
+ "dataPath": "params_shard_1.bin",
29
+ "format": "raw-shard",
30
+ "nbytes": 31826304,
31
+ "records": [
32
+ {
33
+ "name": "model.embed_tokens.q_scale",
34
+ "shape": [
35
+ 32003,
36
+ 64
37
+ ],
38
+ "dtype": "float32",
39
+ "format": "f32-to-bf16",
40
+ "nbytes": 4096384,
41
+ "byteOffset": 0
42
+ },
43
+ {
44
+ "name": "model.layers.0.self_attn.qkv_proj.q_weight",
45
+ "shape": [
46
+ 2560,
47
+ 256
48
+ ],
49
+ "dtype": "uint32",
50
+ "format": "f32-to-bf16",
51
+ "nbytes": 2621440,
52
+ "byteOffset": 4096384
53
+ },
54
+ {
55
+ "name": "model.layers.0.self_attn.qkv_proj.q_scale",
56
+ "shape": [
57
+ 2560,
58
+ 64
59
+ ],
60
+ "dtype": "float32",
61
+ "format": "f32-to-bf16",
62
+ "nbytes": 327680,
63
+ "byteOffset": 6717824
64
+ },
65
+ {
66
+ "name": "model.layers.0.self_attn.o_proj.q_weight",
67
+ "shape": [
68
+ 2048,
69
+ 256
70
+ ],
71
+ "dtype": "uint32",
72
+ "format": "f32-to-bf16",
73
+ "nbytes": 2097152,
74
+ "byteOffset": 7045504
75
+ },
76
+ {
77
+ "name": "model.layers.0.self_attn.o_proj.q_scale",
78
+ "shape": [
79
+ 2048,
80
+ 64
81
+ ],
82
+ "dtype": "float32",
83
+ "format": "f32-to-bf16",
84
+ "nbytes": 262144,
85
+ "byteOffset": 9142656
86
+ },
87
+ {
88
+ "name": "model.layers.0.mlp.gate_up_proj.q_weight",
89
+ "shape": [
90
+ 11264,
91
+ 256
92
+ ],
93
+ "dtype": "uint32",
94
+ "format": "f32-to-bf16",
95
+ "nbytes": 11534336,
96
+ "byteOffset": 9404800
97
+ },
98
+ {
99
+ "name": "model.layers.0.mlp.gate_up_proj.q_scale",
100
+ "shape": [
101
+ 11264,
102
+ 64
103
+ ],
104
+ "dtype": "float32",
105
+ "format": "f32-to-bf16",
106
+ "nbytes": 1441792,
107
+ "byteOffset": 20939136
108
+ },
109
+ {
110
+ "name": "model.layers.0.mlp.down_proj.q_weight",
111
+ "shape": [
112
+ 2048,
113
+ 704
114
+ ],
115
+ "dtype": "uint32",
116
+ "format": "f32-to-bf16",
117
+ "nbytes": 5767168,
118
+ "byteOffset": 22380928
119
+ },
120
+ {
121
+ "name": "model.layers.0.mlp.down_proj.q_scale",
122
+ "shape": [
123
+ 2048,
124
+ 176
125
+ ],
126
+ "dtype": "float32",
127
+ "format": "f32-to-bf16",
128
+ "nbytes": 720896,
129
+ "byteOffset": 28148096
130
+ },
131
+ {
132
+ "name": "model.layers.0.input_layernorm.weight",
133
+ "shape": [
134
+ 2048
135
+ ],
136
+ "dtype": "float32",
137
+ "format": "f32-to-bf16",
138
+ "nbytes": 4096,
139
+ "byteOffset": 28868992
140
+ },
141
+ {
142
+ "name": "model.layers.0.post_attention_layernorm.weight",
143
+ "shape": [
144
+ 2048
145
+ ],
146
+ "dtype": "float32",
147
+ "format": "f32-to-bf16",
148
+ "nbytes": 4096,
149
+ "byteOffset": 28873088
150
+ },
151
+ {
152
+ "name": "model.layers.1.self_attn.qkv_proj.q_weight",
153
+ "shape": [
154
+ 2560,
155
+ 256
156
+ ],
157
+ "dtype": "uint32",
158
+ "format": "f32-to-bf16",
159
+ "nbytes": 2621440,
160
+ "byteOffset": 28877184
161
+ },
162
+ {
163
+ "name": "model.layers.1.self_attn.qkv_proj.q_scale",
164
+ "shape": [
165
+ 2560,
166
+ 64
167
+ ],
168
+ "dtype": "float32",
169
+ "format": "f32-to-bf16",
170
+ "nbytes": 327680,
171
+ "byteOffset": 31498624
172
+ }
173
+ ],
174
+ "md5sum": "680c11ff1744d5bfa3cab96425d32284"
175
+ },
176
+ {
177
+ "dataPath": "params_shard_2.bin",
178
+ "format": "raw-shard",
179
+ "nbytes": 27140096,
180
+ "records": [
181
+ {
182
+ "name": "model.layers.1.self_attn.o_proj.q_weight",
183
+ "shape": [
184
+ 2048,
185
+ 256
186
+ ],
187
+ "dtype": "uint32",
188
+ "format": "f32-to-bf16",
189
+ "nbytes": 2097152,
190
+ "byteOffset": 0
191
+ },
192
+ {
193
+ "name": "model.layers.1.self_attn.o_proj.q_scale",
194
+ "shape": [
195
+ 2048,
196
+ 64
197
+ ],
198
+ "dtype": "float32",
199
+ "format": "f32-to-bf16",
200
+ "nbytes": 262144,
201
+ "byteOffset": 2097152
202
+ },
203
+ {
204
+ "name": "model.layers.1.mlp.gate_up_proj.q_weight",
205
+ "shape": [
206
+ 11264,
207
+ 256
208
+ ],
209
+ "dtype": "uint32",
210
+ "format": "f32-to-bf16",
211
+ "nbytes": 11534336,
212
+ "byteOffset": 2359296
213
+ },
214
+ {
215
+ "name": "model.layers.1.mlp.gate_up_proj.q_scale",
216
+ "shape": [
217
+ 11264,
218
+ 64
219
+ ],
220
+ "dtype": "float32",
221
+ "format": "f32-to-bf16",
222
+ "nbytes": 1441792,
223
+ "byteOffset": 13893632
224
+ },
225
+ {
226
+ "name": "model.layers.1.mlp.down_proj.q_weight",
227
+ "shape": [
228
+ 2048,
229
+ 704
230
+ ],
231
+ "dtype": "uint32",
232
+ "format": "f32-to-bf16",
233
+ "nbytes": 5767168,
234
+ "byteOffset": 15335424
235
+ },
236
+ {
237
+ "name": "model.layers.1.mlp.down_proj.q_scale",
238
+ "shape": [
239
+ 2048,
240
+ 176
241
+ ],
242
+ "dtype": "float32",
243
+ "format": "f32-to-bf16",
244
+ "nbytes": 720896,
245
+ "byteOffset": 21102592
246
+ },
247
+ {
248
+ "name": "model.layers.1.input_layernorm.weight",
249
+ "shape": [
250
+ 2048
251
+ ],
252
+ "dtype": "float32",
253
+ "format": "f32-to-bf16",
254
+ "nbytes": 4096,
255
+ "byteOffset": 21823488
256
+ },
257
+ {
258
+ "name": "model.layers.1.post_attention_layernorm.weight",
259
+ "shape": [
260
+ 2048
261
+ ],
262
+ "dtype": "float32",
263
+ "format": "f32-to-bf16",
264
+ "nbytes": 4096,
265
+ "byteOffset": 21827584
266
+ },
267
+ {
268
+ "name": "model.layers.2.self_attn.qkv_proj.q_weight",
269
+ "shape": [
270
+ 2560,
271
+ 256
272
+ ],
273
+ "dtype": "uint32",
274
+ "format": "f32-to-bf16",
275
+ "nbytes": 2621440,
276
+ "byteOffset": 21831680
277
+ },
278
+ {
279
+ "name": "model.layers.2.self_attn.qkv_proj.q_scale",
280
+ "shape": [
281
+ 2560,
282
+ 64
283
+ ],
284
+ "dtype": "float32",
285
+ "format": "f32-to-bf16",
286
+ "nbytes": 327680,
287
+ "byteOffset": 24453120
288
+ },
289
+ {
290
+ "name": "model.layers.2.self_attn.o_proj.q_weight",
291
+ "shape": [
292
+ 2048,
293
+ 256
294
+ ],
295
+ "dtype": "uint32",
296
+ "format": "f32-to-bf16",
297
+ "nbytes": 2097152,
298
+ "byteOffset": 24780800
299
+ },
300
+ {
301
+ "name": "model.layers.2.self_attn.o_proj.q_scale",
302
+ "shape": [
303
+ 2048,
304
+ 64
305
+ ],
306
+ "dtype": "float32",
307
+ "format": "f32-to-bf16",
308
+ "nbytes": 262144,
309
+ "byteOffset": 26877952
310
+ }
311
+ ],
312
+ "md5sum": "8c28be42014a0308dacc2f2ecdf10aa3"
313
+ },
314
+ {
315
+ "dataPath": "params_shard_3.bin",
316
+ "format": "raw-shard",
317
+ "nbytes": 24780800,
318
+ "records": [
319
+ {
320
+ "name": "model.layers.2.mlp.gate_up_proj.q_weight",
321
+ "shape": [
322
+ 11264,
323
+ 256
324
+ ],
325
+ "dtype": "uint32",
326
+ "format": "f32-to-bf16",
327
+ "nbytes": 11534336,
328
+ "byteOffset": 0
329
+ },
330
+ {
331
+ "name": "model.layers.2.mlp.gate_up_proj.q_scale",
332
+ "shape": [
333
+ 11264,
334
+ 64
335
+ ],
336
+ "dtype": "float32",
337
+ "format": "f32-to-bf16",
338
+ "nbytes": 1441792,
339
+ "byteOffset": 11534336
340
+ },
341
+ {
342
+ "name": "model.layers.2.mlp.down_proj.q_weight",
343
+ "shape": [
344
+ 2048,
345
+ 704
346
+ ],
347
+ "dtype": "uint32",
348
+ "format": "f32-to-bf16",
349
+ "nbytes": 5767168,
350
+ "byteOffset": 12976128
351
+ },
352
+ {
353
+ "name": "model.layers.2.mlp.down_proj.q_scale",
354
+ "shape": [
355
+ 2048,
356
+ 176
357
+ ],
358
+ "dtype": "float32",
359
+ "format": "f32-to-bf16",
360
+ "nbytes": 720896,
361
+ "byteOffset": 18743296
362
+ },
363
+ {
364
+ "name": "model.layers.2.input_layernorm.weight",
365
+ "shape": [
366
+ 2048
367
+ ],
368
+ "dtype": "float32",
369
+ "format": "f32-to-bf16",
370
+ "nbytes": 4096,
371
+ "byteOffset": 19464192
372
+ },
373
+ {
374
+ "name": "model.layers.2.post_attention_layernorm.weight",
375
+ "shape": [
376
+ 2048
377
+ ],
378
+ "dtype": "float32",
379
+ "format": "f32-to-bf16",
380
+ "nbytes": 4096,
381
+ "byteOffset": 19468288
382
+ },
383
+ {
384
+ "name": "model.layers.3.self_attn.qkv_proj.q_weight",
385
+ "shape": [
386
+ 2560,
387
+ 256
388
+ ],
389
+ "dtype": "uint32",
390
+ "format": "f32-to-bf16",
391
+ "nbytes": 2621440,
392
+ "byteOffset": 19472384
393
+ },
394
+ {
395
+ "name": "model.layers.3.self_attn.qkv_proj.q_scale",
396
+ "shape": [
397
+ 2560,
398
+ 64
399
+ ],
400
+ "dtype": "float32",
401
+ "format": "f32-to-bf16",
402
+ "nbytes": 327680,
403
+ "byteOffset": 22093824
404
+ },
405
+ {
406
+ "name": "model.layers.3.self_attn.o_proj.q_weight",
407
+ "shape": [
408
+ 2048,
409
+ 256
410
+ ],
411
+ "dtype": "uint32",
412
+ "format": "f32-to-bf16",
413
+ "nbytes": 2097152,
414
+ "byteOffset": 22421504
415
+ },
416
+ {
417
+ "name": "model.layers.3.self_attn.o_proj.q_scale",
418
+ "shape": [
419
+ 2048,
420
+ 64
421
+ ],
422
+ "dtype": "float32",
423
+ "format": "f32-to-bf16",
424
+ "nbytes": 262144,
425
+ "byteOffset": 24518656
426
+ }
427
+ ],
428
+ "md5sum": "a2fe11ae5b5c809f09adfcb9f58422be"
429
+ },
430
+ {
431
+ "dataPath": "params_shard_4.bin",
432
+ "format": "raw-shard",
433
+ "nbytes": 24780800,
434
+ "records": [
435
+ {
436
+ "name": "model.layers.3.mlp.gate_up_proj.q_weight",
437
+ "shape": [
438
+ 11264,
439
+ 256
440
+ ],
441
+ "dtype": "uint32",
442
+ "format": "f32-to-bf16",
443
+ "nbytes": 11534336,
444
+ "byteOffset": 0
445
+ },
446
+ {
447
+ "name": "model.layers.3.mlp.gate_up_proj.q_scale",
448
+ "shape": [
449
+ 11264,
450
+ 64
451
+ ],
452
+ "dtype": "float32",
453
+ "format": "f32-to-bf16",
454
+ "nbytes": 1441792,
455
+ "byteOffset": 11534336
456
+ },
457
+ {
458
+ "name": "model.layers.3.mlp.down_proj.q_weight",
459
+ "shape": [
460
+ 2048,
461
+ 704
462
+ ],
463
+ "dtype": "uint32",
464
+ "format": "f32-to-bf16",
465
+ "nbytes": 5767168,
466
+ "byteOffset": 12976128
467
+ },
468
+ {
469
+ "name": "model.layers.3.mlp.down_proj.q_scale",
470
+ "shape": [
471
+ 2048,
472
+ 176
473
+ ],
474
+ "dtype": "float32",
475
+ "format": "f32-to-bf16",
476
+ "nbytes": 720896,
477
+ "byteOffset": 18743296
478
+ },
479
+ {
480
+ "name": "model.layers.3.input_layernorm.weight",
481
+ "shape": [
482
+ 2048
483
+ ],
484
+ "dtype": "float32",
485
+ "format": "f32-to-bf16",
486
+ "nbytes": 4096,
487
+ "byteOffset": 19464192
488
+ },
489
+ {
490
+ "name": "model.layers.3.post_attention_layernorm.weight",
491
+ "shape": [
492
+ 2048
493
+ ],
494
+ "dtype": "float32",
495
+ "format": "f32-to-bf16",
496
+ "nbytes": 4096,
497
+ "byteOffset": 19468288
498
+ },
499
+ {
500
+ "name": "model.layers.4.self_attn.qkv_proj.q_weight",
501
+ "shape": [
502
+ 2560,
503
+ 256
504
+ ],
505
+ "dtype": "uint32",
506
+ "format": "f32-to-bf16",
507
+ "nbytes": 2621440,
508
+ "byteOffset": 19472384
509
+ },
510
+ {
511
+ "name": "model.layers.4.self_attn.qkv_proj.q_scale",
512
+ "shape": [
513
+ 2560,
514
+ 64
515
+ ],
516
+ "dtype": "float32",
517
+ "format": "f32-to-bf16",
518
+ "nbytes": 327680,
519
+ "byteOffset": 22093824
520
+ },
521
+ {
522
+ "name": "model.layers.4.self_attn.o_proj.q_weight",
523
+ "shape": [
524
+ 2048,
525
+ 256
526
+ ],
527
+ "dtype": "uint32",
528
+ "format": "f32-to-bf16",
529
+ "nbytes": 2097152,
530
+ "byteOffset": 22421504
531
+ },
532
+ {
533
+ "name": "model.layers.4.self_attn.o_proj.q_scale",
534
+ "shape": [
535
+ 2048,
536
+ 64
537
+ ],
538
+ "dtype": "float32",
539
+ "format": "f32-to-bf16",
540
+ "nbytes": 262144,
541
+ "byteOffset": 24518656
542
+ }
543
+ ],
544
+ "md5sum": "4cddfca8f748085f6fc7deeabda1ce27"
545
+ },
546
+ {
547
+ "dataPath": "params_shard_5.bin",
548
+ "format": "raw-shard",
549
+ "nbytes": 24780800,
550
+ "records": [
551
+ {
552
+ "name": "model.layers.4.mlp.gate_up_proj.q_weight",
553
+ "shape": [
554
+ 11264,
555
+ 256
556
+ ],
557
+ "dtype": "uint32",
558
+ "format": "f32-to-bf16",
559
+ "nbytes": 11534336,
560
+ "byteOffset": 0
561
+ },
562
+ {
563
+ "name": "model.layers.4.mlp.gate_up_proj.q_scale",
564
+ "shape": [
565
+ 11264,
566
+ 64
567
+ ],
568
+ "dtype": "float32",
569
+ "format": "f32-to-bf16",
570
+ "nbytes": 1441792,
571
+ "byteOffset": 11534336
572
+ },
573
+ {
574
+ "name": "model.layers.4.mlp.down_proj.q_weight",
575
+ "shape": [
576
+ 2048,
577
+ 704
578
+ ],
579
+ "dtype": "uint32",
580
+ "format": "f32-to-bf16",
581
+ "nbytes": 5767168,
582
+ "byteOffset": 12976128
583
+ },
584
+ {
585
+ "name": "model.layers.4.mlp.down_proj.q_scale",
586
+ "shape": [
587
+ 2048,
588
+ 176
589
+ ],
590
+ "dtype": "float32",
591
+ "format": "f32-to-bf16",
592
+ "nbytes": 720896,
593
+ "byteOffset": 18743296
594
+ },
595
+ {
596
+ "name": "model.layers.4.input_layernorm.weight",
597
+ "shape": [
598
+ 2048
599
+ ],
600
+ "dtype": "float32",
601
+ "format": "f32-to-bf16",
602
+ "nbytes": 4096,
603
+ "byteOffset": 19464192
604
+ },
605
+ {
606
+ "name": "model.layers.4.post_attention_layernorm.weight",
607
+ "shape": [
608
+ 2048
609
+ ],
610
+ "dtype": "float32",
611
+ "format": "f32-to-bf16",
612
+ "nbytes": 4096,
613
+ "byteOffset": 19468288
614
+ },
615
+ {
616
+ "name": "model.layers.5.self_attn.qkv_proj.q_weight",
617
+ "shape": [
618
+ 2560,
619
+ 256
620
+ ],
621
+ "dtype": "uint32",
622
+ "format": "f32-to-bf16",
623
+ "nbytes": 2621440,
624
+ "byteOffset": 19472384
625
+ },
626
+ {
627
+ "name": "model.layers.5.self_attn.qkv_proj.q_scale",
628
+ "shape": [
629
+ 2560,
630
+ 64
631
+ ],
632
+ "dtype": "float32",
633
+ "format": "f32-to-bf16",
634
+ "nbytes": 327680,
635
+ "byteOffset": 22093824
636
+ },
637
+ {
638
+ "name": "model.layers.5.self_attn.o_proj.q_weight",
639
+ "shape": [
640
+ 2048,
641
+ 256
642
+ ],
643
+ "dtype": "uint32",
644
+ "format": "f32-to-bf16",
645
+ "nbytes": 2097152,
646
+ "byteOffset": 22421504
647
+ },
648
+ {
649
+ "name": "model.layers.5.self_attn.o_proj.q_scale",
650
+ "shape": [
651
+ 2048,
652
+ 64
653
+ ],
654
+ "dtype": "float32",
655
+ "format": "f32-to-bf16",
656
+ "nbytes": 262144,
657
+ "byteOffset": 24518656
658
+ }
659
+ ],
660
+ "md5sum": "e120e89f444fd0c436baf3068f479762"
661
+ },
662
+ {
663
+ "dataPath": "params_shard_6.bin",
664
+ "format": "raw-shard",
665
+ "nbytes": 24780800,
666
+ "records": [
667
+ {
668
+ "name": "model.layers.5.mlp.gate_up_proj.q_weight",
669
+ "shape": [
670
+ 11264,
671
+ 256
672
+ ],
673
+ "dtype": "uint32",
674
+ "format": "f32-to-bf16",
675
+ "nbytes": 11534336,
676
+ "byteOffset": 0
677
+ },
678
+ {
679
+ "name": "model.layers.5.mlp.gate_up_proj.q_scale",
680
+ "shape": [
681
+ 11264,
682
+ 64
683
+ ],
684
+ "dtype": "float32",
685
+ "format": "f32-to-bf16",
686
+ "nbytes": 1441792,
687
+ "byteOffset": 11534336
688
+ },
689
+ {
690
+ "name": "model.layers.5.mlp.down_proj.q_weight",
691
+ "shape": [
692
+ 2048,
693
+ 704
694
+ ],
695
+ "dtype": "uint32",
696
+ "format": "f32-to-bf16",
697
+ "nbytes": 5767168,
698
+ "byteOffset": 12976128
699
+ },
700
+ {
701
+ "name": "model.layers.5.mlp.down_proj.q_scale",
702
+ "shape": [
703
+ 2048,
704
+ 176
705
+ ],
706
+ "dtype": "float32",
707
+ "format": "f32-to-bf16",
708
+ "nbytes": 720896,
709
+ "byteOffset": 18743296
710
+ },
711
+ {
712
+ "name": "model.layers.5.input_layernorm.weight",
713
+ "shape": [
714
+ 2048
715
+ ],
716
+ "dtype": "float32",
717
+ "format": "f32-to-bf16",
718
+ "nbytes": 4096,
719
+ "byteOffset": 19464192
720
+ },
721
+ {
722
+ "name": "model.layers.5.post_attention_layernorm.weight",
723
+ "shape": [
724
+ 2048
725
+ ],
726
+ "dtype": "float32",
727
+ "format": "f32-to-bf16",
728
+ "nbytes": 4096,
729
+ "byteOffset": 19468288
730
+ },
731
+ {
732
+ "name": "model.layers.6.self_attn.qkv_proj.q_weight",
733
+ "shape": [
734
+ 2560,
735
+ 256
736
+ ],
737
+ "dtype": "uint32",
738
+ "format": "f32-to-bf16",
739
+ "nbytes": 2621440,
740
+ "byteOffset": 19472384
741
+ },
742
+ {
743
+ "name": "model.layers.6.self_attn.qkv_proj.q_scale",
744
+ "shape": [
745
+ 2560,
746
+ 64
747
+ ],
748
+ "dtype": "float32",
749
+ "format": "f32-to-bf16",
750
+ "nbytes": 327680,
751
+ "byteOffset": 22093824
752
+ },
753
+ {
754
+ "name": "model.layers.6.self_attn.o_proj.q_weight",
755
+ "shape": [
756
+ 2048,
757
+ 256
758
+ ],
759
+ "dtype": "uint32",
760
+ "format": "f32-to-bf16",
761
+ "nbytes": 2097152,
762
+ "byteOffset": 22421504
763
+ },
764
+ {
765
+ "name": "model.layers.6.self_attn.o_proj.q_scale",
766
+ "shape": [
767
+ 2048,
768
+ 64
769
+ ],
770
+ "dtype": "float32",
771
+ "format": "f32-to-bf16",
772
+ "nbytes": 262144,
773
+ "byteOffset": 24518656
774
+ }
775
+ ],
776
+ "md5sum": "5a8837813b094d491b5a14205460886e"
777
+ },
778
+ {
779
+ "dataPath": "params_shard_7.bin",
780
+ "format": "raw-shard",
781
+ "nbytes": 24780800,
782
+ "records": [
783
+ {
784
+ "name": "model.layers.6.mlp.gate_up_proj.q_weight",
785
+ "shape": [
786
+ 11264,
787
+ 256
788
+ ],
789
+ "dtype": "uint32",
790
+ "format": "f32-to-bf16",
791
+ "nbytes": 11534336,
792
+ "byteOffset": 0
793
+ },
794
+ {
795
+ "name": "model.layers.6.mlp.gate_up_proj.q_scale",
796
+ "shape": [
797
+ 11264,
798
+ 64
799
+ ],
800
+ "dtype": "float32",
801
+ "format": "f32-to-bf16",
802
+ "nbytes": 1441792,
803
+ "byteOffset": 11534336
804
+ },
805
+ {
806
+ "name": "model.layers.6.mlp.down_proj.q_weight",
807
+ "shape": [
808
+ 2048,
809
+ 704
810
+ ],
811
+ "dtype": "uint32",
812
+ "format": "f32-to-bf16",
813
+ "nbytes": 5767168,
814
+ "byteOffset": 12976128
815
+ },
816
+ {
817
+ "name": "model.layers.6.mlp.down_proj.q_scale",
818
+ "shape": [
819
+ 2048,
820
+ 176
821
+ ],
822
+ "dtype": "float32",
823
+ "format": "f32-to-bf16",
824
+ "nbytes": 720896,
825
+ "byteOffset": 18743296
826
+ },
827
+ {
828
+ "name": "model.layers.6.input_layernorm.weight",
829
+ "shape": [
830
+ 2048
831
+ ],
832
+ "dtype": "float32",
833
+ "format": "f32-to-bf16",
834
+ "nbytes": 4096,
835
+ "byteOffset": 19464192
836
+ },
837
+ {
838
+ "name": "model.layers.6.post_attention_layernorm.weight",
839
+ "shape": [
840
+ 2048
841
+ ],
842
+ "dtype": "float32",
843
+ "format": "f32-to-bf16",
844
+ "nbytes": 4096,
845
+ "byteOffset": 19468288
846
+ },
847
+ {
848
+ "name": "model.layers.7.self_attn.qkv_proj.q_weight",
849
+ "shape": [
850
+ 2560,
851
+ 256
852
+ ],
853
+ "dtype": "uint32",
854
+ "format": "f32-to-bf16",
855
+ "nbytes": 2621440,
856
+ "byteOffset": 19472384
857
+ },
858
+ {
859
+ "name": "model.layers.7.self_attn.qkv_proj.q_scale",
860
+ "shape": [
861
+ 2560,
862
+ 64
863
+ ],
864
+ "dtype": "float32",
865
+ "format": "f32-to-bf16",
866
+ "nbytes": 327680,
867
+ "byteOffset": 22093824
868
+ },
869
+ {
870
+ "name": "model.layers.7.self_attn.o_proj.q_weight",
871
+ "shape": [
872
+ 2048,
873
+ 256
874
+ ],
875
+ "dtype": "uint32",
876
+ "format": "f32-to-bf16",
877
+ "nbytes": 2097152,
878
+ "byteOffset": 22421504
879
+ },
880
+ {
881
+ "name": "model.layers.7.self_attn.o_proj.q_scale",
882
+ "shape": [
883
+ 2048,
884
+ 64
885
+ ],
886
+ "dtype": "float32",
887
+ "format": "f32-to-bf16",
888
+ "nbytes": 262144,
889
+ "byteOffset": 24518656
890
+ }
891
+ ],
892
+ "md5sum": "4de03c54f7105a9e8fb0da0f275f40a3"
893
+ },
894
+ {
895
+ "dataPath": "params_shard_8.bin",
896
+ "format": "raw-shard",
897
+ "nbytes": 24780800,
898
+ "records": [
899
+ {
900
+ "name": "model.layers.7.mlp.gate_up_proj.q_weight",
901
+ "shape": [
902
+ 11264,
903
+ 256
904
+ ],
905
+ "dtype": "uint32",
906
+ "format": "f32-to-bf16",
907
+ "nbytes": 11534336,
908
+ "byteOffset": 0
909
+ },
910
+ {
911
+ "name": "model.layers.7.mlp.gate_up_proj.q_scale",
912
+ "shape": [
913
+ 11264,
914
+ 64
915
+ ],
916
+ "dtype": "float32",
917
+ "format": "f32-to-bf16",
918
+ "nbytes": 1441792,
919
+ "byteOffset": 11534336
920
+ },
921
+ {
922
+ "name": "model.layers.7.mlp.down_proj.q_weight",
923
+ "shape": [
924
+ 2048,
925
+ 704
926
+ ],
927
+ "dtype": "uint32",
928
+ "format": "f32-to-bf16",
929
+ "nbytes": 5767168,
930
+ "byteOffset": 12976128
931
+ },
932
+ {
933
+ "name": "model.layers.7.mlp.down_proj.q_scale",
934
+ "shape": [
935
+ 2048,
936
+ 176
937
+ ],
938
+ "dtype": "float32",
939
+ "format": "f32-to-bf16",
940
+ "nbytes": 720896,
941
+ "byteOffset": 18743296
942
+ },
943
+ {
944
+ "name": "model.layers.7.input_layernorm.weight",
945
+ "shape": [
946
+ 2048
947
+ ],
948
+ "dtype": "float32",
949
+ "format": "f32-to-bf16",
950
+ "nbytes": 4096,
951
+ "byteOffset": 19464192
952
+ },
953
+ {
954
+ "name": "model.layers.7.post_attention_layernorm.weight",
955
+ "shape": [
956
+ 2048
957
+ ],
958
+ "dtype": "float32",
959
+ "format": "f32-to-bf16",
960
+ "nbytes": 4096,
961
+ "byteOffset": 19468288
962
+ },
963
+ {
964
+ "name": "model.layers.8.self_attn.qkv_proj.q_weight",
965
+ "shape": [
966
+ 2560,
967
+ 256
968
+ ],
969
+ "dtype": "uint32",
970
+ "format": "f32-to-bf16",
971
+ "nbytes": 2621440,
972
+ "byteOffset": 19472384
973
+ },
974
+ {
975
+ "name": "model.layers.8.self_attn.qkv_proj.q_scale",
976
+ "shape": [
977
+ 2560,
978
+ 64
979
+ ],
980
+ "dtype": "float32",
981
+ "format": "f32-to-bf16",
982
+ "nbytes": 327680,
983
+ "byteOffset": 22093824
984
+ },
985
+ {
986
+ "name": "model.layers.8.self_attn.o_proj.q_weight",
987
+ "shape": [
988
+ 2048,
989
+ 256
990
+ ],
991
+ "dtype": "uint32",
992
+ "format": "f32-to-bf16",
993
+ "nbytes": 2097152,
994
+ "byteOffset": 22421504
995
+ },
996
+ {
997
+ "name": "model.layers.8.self_attn.o_proj.q_scale",
998
+ "shape": [
999
+ 2048,
1000
+ 64
1001
+ ],
1002
+ "dtype": "float32",
1003
+ "format": "f32-to-bf16",
1004
+ "nbytes": 262144,
1005
+ "byteOffset": 24518656
1006
+ }
1007
+ ],
1008
+ "md5sum": "ab488c2050fdfe91cd78eb473120e649"
1009
+ },
1010
+ {
1011
+ "dataPath": "params_shard_9.bin",
1012
+ "format": "raw-shard",
1013
+ "nbytes": 24780800,
1014
+ "records": [
1015
+ {
1016
+ "name": "model.layers.8.mlp.gate_up_proj.q_weight",
1017
+ "shape": [
1018
+ 11264,
1019
+ 256
1020
+ ],
1021
+ "dtype": "uint32",
1022
+ "format": "f32-to-bf16",
1023
+ "nbytes": 11534336,
1024
+ "byteOffset": 0
1025
+ },
1026
+ {
1027
+ "name": "model.layers.8.mlp.gate_up_proj.q_scale",
1028
+ "shape": [
1029
+ 11264,
1030
+ 64
1031
+ ],
1032
+ "dtype": "float32",
1033
+ "format": "f32-to-bf16",
1034
+ "nbytes": 1441792,
1035
+ "byteOffset": 11534336
1036
+ },
1037
+ {
1038
+ "name": "model.layers.8.mlp.down_proj.q_weight",
1039
+ "shape": [
1040
+ 2048,
1041
+ 704
1042
+ ],
1043
+ "dtype": "uint32",
1044
+ "format": "f32-to-bf16",
1045
+ "nbytes": 5767168,
1046
+ "byteOffset": 12976128
1047
+ },
1048
+ {
1049
+ "name": "model.layers.8.mlp.down_proj.q_scale",
1050
+ "shape": [
1051
+ 2048,
1052
+ 176
1053
+ ],
1054
+ "dtype": "float32",
1055
+ "format": "f32-to-bf16",
1056
+ "nbytes": 720896,
1057
+ "byteOffset": 18743296
1058
+ },
1059
+ {
1060
+ "name": "model.layers.8.input_layernorm.weight",
1061
+ "shape": [
1062
+ 2048
1063
+ ],
1064
+ "dtype": "float32",
1065
+ "format": "f32-to-bf16",
1066
+ "nbytes": 4096,
1067
+ "byteOffset": 19464192
1068
+ },
1069
+ {
1070
+ "name": "model.layers.8.post_attention_layernorm.weight",
1071
+ "shape": [
1072
+ 2048
1073
+ ],
1074
+ "dtype": "float32",
1075
+ "format": "f32-to-bf16",
1076
+ "nbytes": 4096,
1077
+ "byteOffset": 19468288
1078
+ },
1079
+ {
1080
+ "name": "model.layers.9.self_attn.qkv_proj.q_weight",
1081
+ "shape": [
1082
+ 2560,
1083
+ 256
1084
+ ],
1085
+ "dtype": "uint32",
1086
+ "format": "f32-to-bf16",
1087
+ "nbytes": 2621440,
1088
+ "byteOffset": 19472384
1089
+ },
1090
+ {
1091
+ "name": "model.layers.9.self_attn.qkv_proj.q_scale",
1092
+ "shape": [
1093
+ 2560,
1094
+ 64
1095
+ ],
1096
+ "dtype": "float32",
1097
+ "format": "f32-to-bf16",
1098
+ "nbytes": 327680,
1099
+ "byteOffset": 22093824
1100
+ },
1101
+ {
1102
+ "name": "model.layers.9.self_attn.o_proj.q_weight",
1103
+ "shape": [
1104
+ 2048,
1105
+ 256
1106
+ ],
1107
+ "dtype": "uint32",
1108
+ "format": "f32-to-bf16",
1109
+ "nbytes": 2097152,
1110
+ "byteOffset": 22421504
1111
+ },
1112
+ {
1113
+ "name": "model.layers.9.self_attn.o_proj.q_scale",
1114
+ "shape": [
1115
+ 2048,
1116
+ 64
1117
+ ],
1118
+ "dtype": "float32",
1119
+ "format": "f32-to-bf16",
1120
+ "nbytes": 262144,
1121
+ "byteOffset": 24518656
1122
+ }
1123
+ ],
1124
+ "md5sum": "ff443a9a0d0db2a220f332b8c6f4fa6a"
1125
+ },
1126
+ {
1127
+ "dataPath": "params_shard_10.bin",
1128
+ "format": "raw-shard",
1129
+ "nbytes": 24780800,
1130
+ "records": [
1131
+ {
1132
+ "name": "model.layers.9.mlp.gate_up_proj.q_weight",
1133
+ "shape": [
1134
+ 11264,
1135
+ 256
1136
+ ],
1137
+ "dtype": "uint32",
1138
+ "format": "f32-to-bf16",
1139
+ "nbytes": 11534336,
1140
+ "byteOffset": 0
1141
+ },
1142
+ {
1143
+ "name": "model.layers.9.mlp.gate_up_proj.q_scale",
1144
+ "shape": [
1145
+ 11264,
1146
+ 64
1147
+ ],
1148
+ "dtype": "float32",
1149
+ "format": "f32-to-bf16",
1150
+ "nbytes": 1441792,
1151
+ "byteOffset": 11534336
1152
+ },
1153
+ {
1154
+ "name": "model.layers.9.mlp.down_proj.q_weight",
1155
+ "shape": [
1156
+ 2048,
1157
+ 704
1158
+ ],
1159
+ "dtype": "uint32",
1160
+ "format": "f32-to-bf16",
1161
+ "nbytes": 5767168,
1162
+ "byteOffset": 12976128
1163
+ },
1164
+ {
1165
+ "name": "model.layers.9.mlp.down_proj.q_scale",
1166
+ "shape": [
1167
+ 2048,
1168
+ 176
1169
+ ],
1170
+ "dtype": "float32",
1171
+ "format": "f32-to-bf16",
1172
+ "nbytes": 720896,
1173
+ "byteOffset": 18743296
1174
+ },
1175
+ {
1176
+ "name": "model.layers.9.input_layernorm.weight",
1177
+ "shape": [
1178
+ 2048
1179
+ ],
1180
+ "dtype": "float32",
1181
+ "format": "f32-to-bf16",
1182
+ "nbytes": 4096,
1183
+ "byteOffset": 19464192
1184
+ },
1185
+ {
1186
+ "name": "model.layers.9.post_attention_layernorm.weight",
1187
+ "shape": [
1188
+ 2048
1189
+ ],
1190
+ "dtype": "float32",
1191
+ "format": "f32-to-bf16",
1192
+ "nbytes": 4096,
1193
+ "byteOffset": 19468288
1194
+ },
1195
+ {
1196
+ "name": "model.layers.10.self_attn.qkv_proj.q_weight",
1197
+ "shape": [
1198
+ 2560,
1199
+ 256
1200
+ ],
1201
+ "dtype": "uint32",
1202
+ "format": "f32-to-bf16",
1203
+ "nbytes": 2621440,
1204
+ "byteOffset": 19472384
1205
+ },
1206
+ {
1207
+ "name": "model.layers.10.self_attn.qkv_proj.q_scale",
1208
+ "shape": [
1209
+ 2560,
1210
+ 64
1211
+ ],
1212
+ "dtype": "float32",
1213
+ "format": "f32-to-bf16",
1214
+ "nbytes": 327680,
1215
+ "byteOffset": 22093824
1216
+ },
1217
+ {
1218
+ "name": "model.layers.10.self_attn.o_proj.q_weight",
1219
+ "shape": [
1220
+ 2048,
1221
+ 256
1222
+ ],
1223
+ "dtype": "uint32",
1224
+ "format": "f32-to-bf16",
1225
+ "nbytes": 2097152,
1226
+ "byteOffset": 22421504
1227
+ },
1228
+ {
1229
+ "name": "model.layers.10.self_attn.o_proj.q_scale",
1230
+ "shape": [
1231
+ 2048,
1232
+ 64
1233
+ ],
1234
+ "dtype": "float32",
1235
+ "format": "f32-to-bf16",
1236
+ "nbytes": 262144,
1237
+ "byteOffset": 24518656
1238
+ }
1239
+ ],
1240
+ "md5sum": "b3e98ffeb6a1180a5ec8ab7fa89b15f1"
1241
+ },
1242
+ {
1243
+ "dataPath": "params_shard_11.bin",
1244
+ "format": "raw-shard",
1245
+ "nbytes": 24780800,
1246
+ "records": [
1247
+ {
1248
+ "name": "model.layers.10.mlp.gate_up_proj.q_weight",
1249
+ "shape": [
1250
+ 11264,
1251
+ 256
1252
+ ],
1253
+ "dtype": "uint32",
1254
+ "format": "f32-to-bf16",
1255
+ "nbytes": 11534336,
1256
+ "byteOffset": 0
1257
+ },
1258
+ {
1259
+ "name": "model.layers.10.mlp.gate_up_proj.q_scale",
1260
+ "shape": [
1261
+ 11264,
1262
+ 64
1263
+ ],
1264
+ "dtype": "float32",
1265
+ "format": "f32-to-bf16",
1266
+ "nbytes": 1441792,
1267
+ "byteOffset": 11534336
1268
+ },
1269
+ {
1270
+ "name": "model.layers.10.mlp.down_proj.q_weight",
1271
+ "shape": [
1272
+ 2048,
1273
+ 704
1274
+ ],
1275
+ "dtype": "uint32",
1276
+ "format": "f32-to-bf16",
1277
+ "nbytes": 5767168,
1278
+ "byteOffset": 12976128
1279
+ },
1280
+ {
1281
+ "name": "model.layers.10.mlp.down_proj.q_scale",
1282
+ "shape": [
1283
+ 2048,
1284
+ 176
1285
+ ],
1286
+ "dtype": "float32",
1287
+ "format": "f32-to-bf16",
1288
+ "nbytes": 720896,
1289
+ "byteOffset": 18743296
1290
+ },
1291
+ {
1292
+ "name": "model.layers.10.input_layernorm.weight",
1293
+ "shape": [
1294
+ 2048
1295
+ ],
1296
+ "dtype": "float32",
1297
+ "format": "f32-to-bf16",
1298
+ "nbytes": 4096,
1299
+ "byteOffset": 19464192
1300
+ },
1301
+ {
1302
+ "name": "model.layers.10.post_attention_layernorm.weight",
1303
+ "shape": [
1304
+ 2048
1305
+ ],
1306
+ "dtype": "float32",
1307
+ "format": "f32-to-bf16",
1308
+ "nbytes": 4096,
1309
+ "byteOffset": 19468288
1310
+ },
1311
+ {
1312
+ "name": "model.layers.11.self_attn.qkv_proj.q_weight",
1313
+ "shape": [
1314
+ 2560,
1315
+ 256
1316
+ ],
1317
+ "dtype": "uint32",
1318
+ "format": "f32-to-bf16",
1319
+ "nbytes": 2621440,
1320
+ "byteOffset": 19472384
1321
+ },
1322
+ {
1323
+ "name": "model.layers.11.self_attn.qkv_proj.q_scale",
1324
+ "shape": [
1325
+ 2560,
1326
+ 64
1327
+ ],
1328
+ "dtype": "float32",
1329
+ "format": "f32-to-bf16",
1330
+ "nbytes": 327680,
1331
+ "byteOffset": 22093824
1332
+ },
1333
+ {
1334
+ "name": "model.layers.11.self_attn.o_proj.q_weight",
1335
+ "shape": [
1336
+ 2048,
1337
+ 256
1338
+ ],
1339
+ "dtype": "uint32",
1340
+ "format": "f32-to-bf16",
1341
+ "nbytes": 2097152,
1342
+ "byteOffset": 22421504
1343
+ },
1344
+ {
1345
+ "name": "model.layers.11.self_attn.o_proj.q_scale",
1346
+ "shape": [
1347
+ 2048,
1348
+ 64
1349
+ ],
1350
+ "dtype": "float32",
1351
+ "format": "f32-to-bf16",
1352
+ "nbytes": 262144,
1353
+ "byteOffset": 24518656
1354
+ }
1355
+ ],
1356
+ "md5sum": "466137e0af99496138348a221df36f30"
1357
+ },
1358
+ {
1359
+ "dataPath": "params_shard_12.bin",
1360
+ "format": "raw-shard",
1361
+ "nbytes": 24780800,
1362
+ "records": [
1363
+ {
1364
+ "name": "model.layers.11.mlp.gate_up_proj.q_weight",
1365
+ "shape": [
1366
+ 11264,
1367
+ 256
1368
+ ],
1369
+ "dtype": "uint32",
1370
+ "format": "f32-to-bf16",
1371
+ "nbytes": 11534336,
1372
+ "byteOffset": 0
1373
+ },
1374
+ {
1375
+ "name": "model.layers.11.mlp.gate_up_proj.q_scale",
1376
+ "shape": [
1377
+ 11264,
1378
+ 64
1379
+ ],
1380
+ "dtype": "float32",
1381
+ "format": "f32-to-bf16",
1382
+ "nbytes": 1441792,
1383
+ "byteOffset": 11534336
1384
+ },
1385
+ {
1386
+ "name": "model.layers.11.mlp.down_proj.q_weight",
1387
+ "shape": [
1388
+ 2048,
1389
+ 704
1390
+ ],
1391
+ "dtype": "uint32",
1392
+ "format": "f32-to-bf16",
1393
+ "nbytes": 5767168,
1394
+ "byteOffset": 12976128
1395
+ },
1396
+ {
1397
+ "name": "model.layers.11.mlp.down_proj.q_scale",
1398
+ "shape": [
1399
+ 2048,
1400
+ 176
1401
+ ],
1402
+ "dtype": "float32",
1403
+ "format": "f32-to-bf16",
1404
+ "nbytes": 720896,
1405
+ "byteOffset": 18743296
1406
+ },
1407
+ {
1408
+ "name": "model.layers.11.input_layernorm.weight",
1409
+ "shape": [
1410
+ 2048
1411
+ ],
1412
+ "dtype": "float32",
1413
+ "format": "f32-to-bf16",
1414
+ "nbytes": 4096,
1415
+ "byteOffset": 19464192
1416
+ },
1417
+ {
1418
+ "name": "model.layers.11.post_attention_layernorm.weight",
1419
+ "shape": [
1420
+ 2048
1421
+ ],
1422
+ "dtype": "float32",
1423
+ "format": "f32-to-bf16",
1424
+ "nbytes": 4096,
1425
+ "byteOffset": 19468288
1426
+ },
1427
+ {
1428
+ "name": "model.layers.12.self_attn.qkv_proj.q_weight",
1429
+ "shape": [
1430
+ 2560,
1431
+ 256
1432
+ ],
1433
+ "dtype": "uint32",
1434
+ "format": "f32-to-bf16",
1435
+ "nbytes": 2621440,
1436
+ "byteOffset": 19472384
1437
+ },
1438
+ {
1439
+ "name": "model.layers.12.self_attn.qkv_proj.q_scale",
1440
+ "shape": [
1441
+ 2560,
1442
+ 64
1443
+ ],
1444
+ "dtype": "float32",
1445
+ "format": "f32-to-bf16",
1446
+ "nbytes": 327680,
1447
+ "byteOffset": 22093824
1448
+ },
1449
+ {
1450
+ "name": "model.layers.12.self_attn.o_proj.q_weight",
1451
+ "shape": [
1452
+ 2048,
1453
+ 256
1454
+ ],
1455
+ "dtype": "uint32",
1456
+ "format": "f32-to-bf16",
1457
+ "nbytes": 2097152,
1458
+ "byteOffset": 22421504
1459
+ },
1460
+ {
1461
+ "name": "model.layers.12.self_attn.o_proj.q_scale",
1462
+ "shape": [
1463
+ 2048,
1464
+ 64
1465
+ ],
1466
+ "dtype": "float32",
1467
+ "format": "f32-to-bf16",
1468
+ "nbytes": 262144,
1469
+ "byteOffset": 24518656
1470
+ }
1471
+ ],
1472
+ "md5sum": "8097c1ddc6ac569ae089e4e1735709d1"
1473
+ },
1474
+ {
1475
+ "dataPath": "params_shard_13.bin",
1476
+ "format": "raw-shard",
1477
+ "nbytes": 24780800,
1478
+ "records": [
1479
+ {
1480
+ "name": "model.layers.12.mlp.gate_up_proj.q_weight",
1481
+ "shape": [
1482
+ 11264,
1483
+ 256
1484
+ ],
1485
+ "dtype": "uint32",
1486
+ "format": "f32-to-bf16",
1487
+ "nbytes": 11534336,
1488
+ "byteOffset": 0
1489
+ },
1490
+ {
1491
+ "name": "model.layers.12.mlp.gate_up_proj.q_scale",
1492
+ "shape": [
1493
+ 11264,
1494
+ 64
1495
+ ],
1496
+ "dtype": "float32",
1497
+ "format": "f32-to-bf16",
1498
+ "nbytes": 1441792,
1499
+ "byteOffset": 11534336
1500
+ },
1501
+ {
1502
+ "name": "model.layers.12.mlp.down_proj.q_weight",
1503
+ "shape": [
1504
+ 2048,
1505
+ 704
1506
+ ],
1507
+ "dtype": "uint32",
1508
+ "format": "f32-to-bf16",
1509
+ "nbytes": 5767168,
1510
+ "byteOffset": 12976128
1511
+ },
1512
+ {
1513
+ "name": "model.layers.12.mlp.down_proj.q_scale",
1514
+ "shape": [
1515
+ 2048,
1516
+ 176
1517
+ ],
1518
+ "dtype": "float32",
1519
+ "format": "f32-to-bf16",
1520
+ "nbytes": 720896,
1521
+ "byteOffset": 18743296
1522
+ },
1523
+ {
1524
+ "name": "model.layers.12.input_layernorm.weight",
1525
+ "shape": [
1526
+ 2048
1527
+ ],
1528
+ "dtype": "float32",
1529
+ "format": "f32-to-bf16",
1530
+ "nbytes": 4096,
1531
+ "byteOffset": 19464192
1532
+ },
1533
+ {
1534
+ "name": "model.layers.12.post_attention_layernorm.weight",
1535
+ "shape": [
1536
+ 2048
1537
+ ],
1538
+ "dtype": "float32",
1539
+ "format": "f32-to-bf16",
1540
+ "nbytes": 4096,
1541
+ "byteOffset": 19468288
1542
+ },
1543
+ {
1544
+ "name": "model.layers.13.self_attn.qkv_proj.q_weight",
1545
+ "shape": [
1546
+ 2560,
1547
+ 256
1548
+ ],
1549
+ "dtype": "uint32",
1550
+ "format": "f32-to-bf16",
1551
+ "nbytes": 2621440,
1552
+ "byteOffset": 19472384
1553
+ },
1554
+ {
1555
+ "name": "model.layers.13.self_attn.qkv_proj.q_scale",
1556
+ "shape": [
1557
+ 2560,
1558
+ 64
1559
+ ],
1560
+ "dtype": "float32",
1561
+ "format": "f32-to-bf16",
1562
+ "nbytes": 327680,
1563
+ "byteOffset": 22093824
1564
+ },
1565
+ {
1566
+ "name": "model.layers.13.self_attn.o_proj.q_weight",
1567
+ "shape": [
1568
+ 2048,
1569
+ 256
1570
+ ],
1571
+ "dtype": "uint32",
1572
+ "format": "f32-to-bf16",
1573
+ "nbytes": 2097152,
1574
+ "byteOffset": 22421504
1575
+ },
1576
+ {
1577
+ "name": "model.layers.13.self_attn.o_proj.q_scale",
1578
+ "shape": [
1579
+ 2048,
1580
+ 64
1581
+ ],
1582
+ "dtype": "float32",
1583
+ "format": "f32-to-bf16",
1584
+ "nbytes": 262144,
1585
+ "byteOffset": 24518656
1586
+ }
1587
+ ],
1588
+ "md5sum": "cb4d8ac3523d4a78c40f6e05e6a75a3f"
1589
+ },
1590
+ {
1591
+ "dataPath": "params_shard_14.bin",
1592
+ "format": "raw-shard",
1593
+ "nbytes": 24780800,
1594
+ "records": [
1595
+ {
1596
+ "name": "model.layers.13.mlp.gate_up_proj.q_weight",
1597
+ "shape": [
1598
+ 11264,
1599
+ 256
1600
+ ],
1601
+ "dtype": "uint32",
1602
+ "format": "f32-to-bf16",
1603
+ "nbytes": 11534336,
1604
+ "byteOffset": 0
1605
+ },
1606
+ {
1607
+ "name": "model.layers.13.mlp.gate_up_proj.q_scale",
1608
+ "shape": [
1609
+ 11264,
1610
+ 64
1611
+ ],
1612
+ "dtype": "float32",
1613
+ "format": "f32-to-bf16",
1614
+ "nbytes": 1441792,
1615
+ "byteOffset": 11534336
1616
+ },
1617
+ {
1618
+ "name": "model.layers.13.mlp.down_proj.q_weight",
1619
+ "shape": [
1620
+ 2048,
1621
+ 704
1622
+ ],
1623
+ "dtype": "uint32",
1624
+ "format": "f32-to-bf16",
1625
+ "nbytes": 5767168,
1626
+ "byteOffset": 12976128
1627
+ },
1628
+ {
1629
+ "name": "model.layers.13.mlp.down_proj.q_scale",
1630
+ "shape": [
1631
+ 2048,
1632
+ 176
1633
+ ],
1634
+ "dtype": "float32",
1635
+ "format": "f32-to-bf16",
1636
+ "nbytes": 720896,
1637
+ "byteOffset": 18743296
1638
+ },
1639
+ {
1640
+ "name": "model.layers.13.input_layernorm.weight",
1641
+ "shape": [
1642
+ 2048
1643
+ ],
1644
+ "dtype": "float32",
1645
+ "format": "f32-to-bf16",
1646
+ "nbytes": 4096,
1647
+ "byteOffset": 19464192
1648
+ },
1649
+ {
1650
+ "name": "model.layers.13.post_attention_layernorm.weight",
1651
+ "shape": [
1652
+ 2048
1653
+ ],
1654
+ "dtype": "float32",
1655
+ "format": "f32-to-bf16",
1656
+ "nbytes": 4096,
1657
+ "byteOffset": 19468288
1658
+ },
1659
+ {
1660
+ "name": "model.layers.14.self_attn.qkv_proj.q_weight",
1661
+ "shape": [
1662
+ 2560,
1663
+ 256
1664
+ ],
1665
+ "dtype": "uint32",
1666
+ "format": "f32-to-bf16",
1667
+ "nbytes": 2621440,
1668
+ "byteOffset": 19472384
1669
+ },
1670
+ {
1671
+ "name": "model.layers.14.self_attn.qkv_proj.q_scale",
1672
+ "shape": [
1673
+ 2560,
1674
+ 64
1675
+ ],
1676
+ "dtype": "float32",
1677
+ "format": "f32-to-bf16",
1678
+ "nbytes": 327680,
1679
+ "byteOffset": 22093824
1680
+ },
1681
+ {
1682
+ "name": "model.layers.14.self_attn.o_proj.q_weight",
1683
+ "shape": [
1684
+ 2048,
1685
+ 256
1686
+ ],
1687
+ "dtype": "uint32",
1688
+ "format": "f32-to-bf16",
1689
+ "nbytes": 2097152,
1690
+ "byteOffset": 22421504
1691
+ },
1692
+ {
1693
+ "name": "model.layers.14.self_attn.o_proj.q_scale",
1694
+ "shape": [
1695
+ 2048,
1696
+ 64
1697
+ ],
1698
+ "dtype": "float32",
1699
+ "format": "f32-to-bf16",
1700
+ "nbytes": 262144,
1701
+ "byteOffset": 24518656
1702
+ }
1703
+ ],
1704
+ "md5sum": "97fec4eb81b0c8ba326d6bd7c2d0d01e"
1705
+ },
1706
+ {
1707
+ "dataPath": "params_shard_15.bin",
1708
+ "format": "raw-shard",
1709
+ "nbytes": 24780800,
1710
+ "records": [
1711
+ {
1712
+ "name": "model.layers.14.mlp.gate_up_proj.q_weight",
1713
+ "shape": [
1714
+ 11264,
1715
+ 256
1716
+ ],
1717
+ "dtype": "uint32",
1718
+ "format": "f32-to-bf16",
1719
+ "nbytes": 11534336,
1720
+ "byteOffset": 0
1721
+ },
1722
+ {
1723
+ "name": "model.layers.14.mlp.gate_up_proj.q_scale",
1724
+ "shape": [
1725
+ 11264,
1726
+ 64
1727
+ ],
1728
+ "dtype": "float32",
1729
+ "format": "f32-to-bf16",
1730
+ "nbytes": 1441792,
1731
+ "byteOffset": 11534336
1732
+ },
1733
+ {
1734
+ "name": "model.layers.14.mlp.down_proj.q_weight",
1735
+ "shape": [
1736
+ 2048,
1737
+ 704
1738
+ ],
1739
+ "dtype": "uint32",
1740
+ "format": "f32-to-bf16",
1741
+ "nbytes": 5767168,
1742
+ "byteOffset": 12976128
1743
+ },
1744
+ {
1745
+ "name": "model.layers.14.mlp.down_proj.q_scale",
1746
+ "shape": [
1747
+ 2048,
1748
+ 176
1749
+ ],
1750
+ "dtype": "float32",
1751
+ "format": "f32-to-bf16",
1752
+ "nbytes": 720896,
1753
+ "byteOffset": 18743296
1754
+ },
1755
+ {
1756
+ "name": "model.layers.14.input_layernorm.weight",
1757
+ "shape": [
1758
+ 2048
1759
+ ],
1760
+ "dtype": "float32",
1761
+ "format": "f32-to-bf16",
1762
+ "nbytes": 4096,
1763
+ "byteOffset": 19464192
1764
+ },
1765
+ {
1766
+ "name": "model.layers.14.post_attention_layernorm.weight",
1767
+ "shape": [
1768
+ 2048
1769
+ ],
1770
+ "dtype": "float32",
1771
+ "format": "f32-to-bf16",
1772
+ "nbytes": 4096,
1773
+ "byteOffset": 19468288
1774
+ },
1775
+ {
1776
+ "name": "model.layers.15.self_attn.qkv_proj.q_weight",
1777
+ "shape": [
1778
+ 2560,
1779
+ 256
1780
+ ],
1781
+ "dtype": "uint32",
1782
+ "format": "f32-to-bf16",
1783
+ "nbytes": 2621440,
1784
+ "byteOffset": 19472384
1785
+ },
1786
+ {
1787
+ "name": "model.layers.15.self_attn.qkv_proj.q_scale",
1788
+ "shape": [
1789
+ 2560,
1790
+ 64
1791
+ ],
1792
+ "dtype": "float32",
1793
+ "format": "f32-to-bf16",
1794
+ "nbytes": 327680,
1795
+ "byteOffset": 22093824
1796
+ },
1797
+ {
1798
+ "name": "model.layers.15.self_attn.o_proj.q_weight",
1799
+ "shape": [
1800
+ 2048,
1801
+ 256
1802
+ ],
1803
+ "dtype": "uint32",
1804
+ "format": "f32-to-bf16",
1805
+ "nbytes": 2097152,
1806
+ "byteOffset": 22421504
1807
+ },
1808
+ {
1809
+ "name": "model.layers.15.self_attn.o_proj.q_scale",
1810
+ "shape": [
1811
+ 2048,
1812
+ 64
1813
+ ],
1814
+ "dtype": "float32",
1815
+ "format": "f32-to-bf16",
1816
+ "nbytes": 262144,
1817
+ "byteOffset": 24518656
1818
+ }
1819
+ ],
1820
+ "md5sum": "5942e747ef08c2f1f6aee113d2fdb843"
1821
+ },
1822
+ {
1823
+ "dataPath": "params_shard_16.bin",
1824
+ "format": "raw-shard",
1825
+ "nbytes": 24780800,
1826
+ "records": [
1827
+ {
1828
+ "name": "model.layers.15.mlp.gate_up_proj.q_weight",
1829
+ "shape": [
1830
+ 11264,
1831
+ 256
1832
+ ],
1833
+ "dtype": "uint32",
1834
+ "format": "f32-to-bf16",
1835
+ "nbytes": 11534336,
1836
+ "byteOffset": 0
1837
+ },
1838
+ {
1839
+ "name": "model.layers.15.mlp.gate_up_proj.q_scale",
1840
+ "shape": [
1841
+ 11264,
1842
+ 64
1843
+ ],
1844
+ "dtype": "float32",
1845
+ "format": "f32-to-bf16",
1846
+ "nbytes": 1441792,
1847
+ "byteOffset": 11534336
1848
+ },
1849
+ {
1850
+ "name": "model.layers.15.mlp.down_proj.q_weight",
1851
+ "shape": [
1852
+ 2048,
1853
+ 704
1854
+ ],
1855
+ "dtype": "uint32",
1856
+ "format": "f32-to-bf16",
1857
+ "nbytes": 5767168,
1858
+ "byteOffset": 12976128
1859
+ },
1860
+ {
1861
+ "name": "model.layers.15.mlp.down_proj.q_scale",
1862
+ "shape": [
1863
+ 2048,
1864
+ 176
1865
+ ],
1866
+ "dtype": "float32",
1867
+ "format": "f32-to-bf16",
1868
+ "nbytes": 720896,
1869
+ "byteOffset": 18743296
1870
+ },
1871
+ {
1872
+ "name": "model.layers.15.input_layernorm.weight",
1873
+ "shape": [
1874
+ 2048
1875
+ ],
1876
+ "dtype": "float32",
1877
+ "format": "f32-to-bf16",
1878
+ "nbytes": 4096,
1879
+ "byteOffset": 19464192
1880
+ },
1881
+ {
1882
+ "name": "model.layers.15.post_attention_layernorm.weight",
1883
+ "shape": [
1884
+ 2048
1885
+ ],
1886
+ "dtype": "float32",
1887
+ "format": "f32-to-bf16",
1888
+ "nbytes": 4096,
1889
+ "byteOffset": 19468288
1890
+ },
1891
+ {
1892
+ "name": "model.layers.16.self_attn.qkv_proj.q_weight",
1893
+ "shape": [
1894
+ 2560,
1895
+ 256
1896
+ ],
1897
+ "dtype": "uint32",
1898
+ "format": "f32-to-bf16",
1899
+ "nbytes": 2621440,
1900
+ "byteOffset": 19472384
1901
+ },
1902
+ {
1903
+ "name": "model.layers.16.self_attn.qkv_proj.q_scale",
1904
+ "shape": [
1905
+ 2560,
1906
+ 64
1907
+ ],
1908
+ "dtype": "float32",
1909
+ "format": "f32-to-bf16",
1910
+ "nbytes": 327680,
1911
+ "byteOffset": 22093824
1912
+ },
1913
+ {
1914
+ "name": "model.layers.16.self_attn.o_proj.q_weight",
1915
+ "shape": [
1916
+ 2048,
1917
+ 256
1918
+ ],
1919
+ "dtype": "uint32",
1920
+ "format": "f32-to-bf16",
1921
+ "nbytes": 2097152,
1922
+ "byteOffset": 22421504
1923
+ },
1924
+ {
1925
+ "name": "model.layers.16.self_attn.o_proj.q_scale",
1926
+ "shape": [
1927
+ 2048,
1928
+ 64
1929
+ ],
1930
+ "dtype": "float32",
1931
+ "format": "f32-to-bf16",
1932
+ "nbytes": 262144,
1933
+ "byteOffset": 24518656
1934
+ }
1935
+ ],
1936
+ "md5sum": "fa1b0776d20499dcf35ca524e65addcd"
1937
+ },
1938
+ {
1939
+ "dataPath": "params_shard_17.bin",
1940
+ "format": "raw-shard",
1941
+ "nbytes": 24780800,
1942
+ "records": [
1943
+ {
1944
+ "name": "model.layers.16.mlp.gate_up_proj.q_weight",
1945
+ "shape": [
1946
+ 11264,
1947
+ 256
1948
+ ],
1949
+ "dtype": "uint32",
1950
+ "format": "f32-to-bf16",
1951
+ "nbytes": 11534336,
1952
+ "byteOffset": 0
1953
+ },
1954
+ {
1955
+ "name": "model.layers.16.mlp.gate_up_proj.q_scale",
1956
+ "shape": [
1957
+ 11264,
1958
+ 64
1959
+ ],
1960
+ "dtype": "float32",
1961
+ "format": "f32-to-bf16",
1962
+ "nbytes": 1441792,
1963
+ "byteOffset": 11534336
1964
+ },
1965
+ {
1966
+ "name": "model.layers.16.mlp.down_proj.q_weight",
1967
+ "shape": [
1968
+ 2048,
1969
+ 704
1970
+ ],
1971
+ "dtype": "uint32",
1972
+ "format": "f32-to-bf16",
1973
+ "nbytes": 5767168,
1974
+ "byteOffset": 12976128
1975
+ },
1976
+ {
1977
+ "name": "model.layers.16.mlp.down_proj.q_scale",
1978
+ "shape": [
1979
+ 2048,
1980
+ 176
1981
+ ],
1982
+ "dtype": "float32",
1983
+ "format": "f32-to-bf16",
1984
+ "nbytes": 720896,
1985
+ "byteOffset": 18743296
1986
+ },
1987
+ {
1988
+ "name": "model.layers.16.input_layernorm.weight",
1989
+ "shape": [
1990
+ 2048
1991
+ ],
1992
+ "dtype": "float32",
1993
+ "format": "f32-to-bf16",
1994
+ "nbytes": 4096,
1995
+ "byteOffset": 19464192
1996
+ },
1997
+ {
1998
+ "name": "model.layers.16.post_attention_layernorm.weight",
1999
+ "shape": [
2000
+ 2048
2001
+ ],
2002
+ "dtype": "float32",
2003
+ "format": "f32-to-bf16",
2004
+ "nbytes": 4096,
2005
+ "byteOffset": 19468288
2006
+ },
2007
+ {
2008
+ "name": "model.layers.17.self_attn.qkv_proj.q_weight",
2009
+ "shape": [
2010
+ 2560,
2011
+ 256
2012
+ ],
2013
+ "dtype": "uint32",
2014
+ "format": "f32-to-bf16",
2015
+ "nbytes": 2621440,
2016
+ "byteOffset": 19472384
2017
+ },
2018
+ {
2019
+ "name": "model.layers.17.self_attn.qkv_proj.q_scale",
2020
+ "shape": [
2021
+ 2560,
2022
+ 64
2023
+ ],
2024
+ "dtype": "float32",
2025
+ "format": "f32-to-bf16",
2026
+ "nbytes": 327680,
2027
+ "byteOffset": 22093824
2028
+ },
2029
+ {
2030
+ "name": "model.layers.17.self_attn.o_proj.q_weight",
2031
+ "shape": [
2032
+ 2048,
2033
+ 256
2034
+ ],
2035
+ "dtype": "uint32",
2036
+ "format": "f32-to-bf16",
2037
+ "nbytes": 2097152,
2038
+ "byteOffset": 22421504
2039
+ },
2040
+ {
2041
+ "name": "model.layers.17.self_attn.o_proj.q_scale",
2042
+ "shape": [
2043
+ 2048,
2044
+ 64
2045
+ ],
2046
+ "dtype": "float32",
2047
+ "format": "f32-to-bf16",
2048
+ "nbytes": 262144,
2049
+ "byteOffset": 24518656
2050
+ }
2051
+ ],
2052
+ "md5sum": "4eec190a3558aad16647bedadf63ad68"
2053
+ },
2054
+ {
2055
+ "dataPath": "params_shard_18.bin",
2056
+ "format": "raw-shard",
2057
+ "nbytes": 24780800,
2058
+ "records": [
2059
+ {
2060
+ "name": "model.layers.17.mlp.gate_up_proj.q_weight",
2061
+ "shape": [
2062
+ 11264,
2063
+ 256
2064
+ ],
2065
+ "dtype": "uint32",
2066
+ "format": "f32-to-bf16",
2067
+ "nbytes": 11534336,
2068
+ "byteOffset": 0
2069
+ },
2070
+ {
2071
+ "name": "model.layers.17.mlp.gate_up_proj.q_scale",
2072
+ "shape": [
2073
+ 11264,
2074
+ 64
2075
+ ],
2076
+ "dtype": "float32",
2077
+ "format": "f32-to-bf16",
2078
+ "nbytes": 1441792,
2079
+ "byteOffset": 11534336
2080
+ },
2081
+ {
2082
+ "name": "model.layers.17.mlp.down_proj.q_weight",
2083
+ "shape": [
2084
+ 2048,
2085
+ 704
2086
+ ],
2087
+ "dtype": "uint32",
2088
+ "format": "f32-to-bf16",
2089
+ "nbytes": 5767168,
2090
+ "byteOffset": 12976128
2091
+ },
2092
+ {
2093
+ "name": "model.layers.17.mlp.down_proj.q_scale",
2094
+ "shape": [
2095
+ 2048,
2096
+ 176
2097
+ ],
2098
+ "dtype": "float32",
2099
+ "format": "f32-to-bf16",
2100
+ "nbytes": 720896,
2101
+ "byteOffset": 18743296
2102
+ },
2103
+ {
2104
+ "name": "model.layers.17.input_layernorm.weight",
2105
+ "shape": [
2106
+ 2048
2107
+ ],
2108
+ "dtype": "float32",
2109
+ "format": "f32-to-bf16",
2110
+ "nbytes": 4096,
2111
+ "byteOffset": 19464192
2112
+ },
2113
+ {
2114
+ "name": "model.layers.17.post_attention_layernorm.weight",
2115
+ "shape": [
2116
+ 2048
2117
+ ],
2118
+ "dtype": "float32",
2119
+ "format": "f32-to-bf16",
2120
+ "nbytes": 4096,
2121
+ "byteOffset": 19468288
2122
+ },
2123
+ {
2124
+ "name": "model.layers.18.self_attn.qkv_proj.q_weight",
2125
+ "shape": [
2126
+ 2560,
2127
+ 256
2128
+ ],
2129
+ "dtype": "uint32",
2130
+ "format": "f32-to-bf16",
2131
+ "nbytes": 2621440,
2132
+ "byteOffset": 19472384
2133
+ },
2134
+ {
2135
+ "name": "model.layers.18.self_attn.qkv_proj.q_scale",
2136
+ "shape": [
2137
+ 2560,
2138
+ 64
2139
+ ],
2140
+ "dtype": "float32",
2141
+ "format": "f32-to-bf16",
2142
+ "nbytes": 327680,
2143
+ "byteOffset": 22093824
2144
+ },
2145
+ {
2146
+ "name": "model.layers.18.self_attn.o_proj.q_weight",
2147
+ "shape": [
2148
+ 2048,
2149
+ 256
2150
+ ],
2151
+ "dtype": "uint32",
2152
+ "format": "f32-to-bf16",
2153
+ "nbytes": 2097152,
2154
+ "byteOffset": 22421504
2155
+ },
2156
+ {
2157
+ "name": "model.layers.18.self_attn.o_proj.q_scale",
2158
+ "shape": [
2159
+ 2048,
2160
+ 64
2161
+ ],
2162
+ "dtype": "float32",
2163
+ "format": "f32-to-bf16",
2164
+ "nbytes": 262144,
2165
+ "byteOffset": 24518656
2166
+ }
2167
+ ],
2168
+ "md5sum": "d34c9a3f1d0cb3900907226eca95af8d"
2169
+ },
2170
+ {
2171
+ "dataPath": "params_shard_19.bin",
2172
+ "format": "raw-shard",
2173
+ "nbytes": 24780800,
2174
+ "records": [
2175
+ {
2176
+ "name": "model.layers.18.mlp.gate_up_proj.q_weight",
2177
+ "shape": [
2178
+ 11264,
2179
+ 256
2180
+ ],
2181
+ "dtype": "uint32",
2182
+ "format": "f32-to-bf16",
2183
+ "nbytes": 11534336,
2184
+ "byteOffset": 0
2185
+ },
2186
+ {
2187
+ "name": "model.layers.18.mlp.gate_up_proj.q_scale",
2188
+ "shape": [
2189
+ 11264,
2190
+ 64
2191
+ ],
2192
+ "dtype": "float32",
2193
+ "format": "f32-to-bf16",
2194
+ "nbytes": 1441792,
2195
+ "byteOffset": 11534336
2196
+ },
2197
+ {
2198
+ "name": "model.layers.18.mlp.down_proj.q_weight",
2199
+ "shape": [
2200
+ 2048,
2201
+ 704
2202
+ ],
2203
+ "dtype": "uint32",
2204
+ "format": "f32-to-bf16",
2205
+ "nbytes": 5767168,
2206
+ "byteOffset": 12976128
2207
+ },
2208
+ {
2209
+ "name": "model.layers.18.mlp.down_proj.q_scale",
2210
+ "shape": [
2211
+ 2048,
2212
+ 176
2213
+ ],
2214
+ "dtype": "float32",
2215
+ "format": "f32-to-bf16",
2216
+ "nbytes": 720896,
2217
+ "byteOffset": 18743296
2218
+ },
2219
+ {
2220
+ "name": "model.layers.18.input_layernorm.weight",
2221
+ "shape": [
2222
+ 2048
2223
+ ],
2224
+ "dtype": "float32",
2225
+ "format": "f32-to-bf16",
2226
+ "nbytes": 4096,
2227
+ "byteOffset": 19464192
2228
+ },
2229
+ {
2230
+ "name": "model.layers.18.post_attention_layernorm.weight",
2231
+ "shape": [
2232
+ 2048
2233
+ ],
2234
+ "dtype": "float32",
2235
+ "format": "f32-to-bf16",
2236
+ "nbytes": 4096,
2237
+ "byteOffset": 19468288
2238
+ },
2239
+ {
2240
+ "name": "model.layers.19.self_attn.qkv_proj.q_weight",
2241
+ "shape": [
2242
+ 2560,
2243
+ 256
2244
+ ],
2245
+ "dtype": "uint32",
2246
+ "format": "f32-to-bf16",
2247
+ "nbytes": 2621440,
2248
+ "byteOffset": 19472384
2249
+ },
2250
+ {
2251
+ "name": "model.layers.19.self_attn.qkv_proj.q_scale",
2252
+ "shape": [
2253
+ 2560,
2254
+ 64
2255
+ ],
2256
+ "dtype": "float32",
2257
+ "format": "f32-to-bf16",
2258
+ "nbytes": 327680,
2259
+ "byteOffset": 22093824
2260
+ },
2261
+ {
2262
+ "name": "model.layers.19.self_attn.o_proj.q_weight",
2263
+ "shape": [
2264
+ 2048,
2265
+ 256
2266
+ ],
2267
+ "dtype": "uint32",
2268
+ "format": "f32-to-bf16",
2269
+ "nbytes": 2097152,
2270
+ "byteOffset": 22421504
2271
+ },
2272
+ {
2273
+ "name": "model.layers.19.self_attn.o_proj.q_scale",
2274
+ "shape": [
2275
+ 2048,
2276
+ 64
2277
+ ],
2278
+ "dtype": "float32",
2279
+ "format": "f32-to-bf16",
2280
+ "nbytes": 262144,
2281
+ "byteOffset": 24518656
2282
+ }
2283
+ ],
2284
+ "md5sum": "829dd454ac1242bc973177bc2248a1f8"
2285
+ },
2286
+ {
2287
+ "dataPath": "params_shard_20.bin",
2288
+ "format": "raw-shard",
2289
+ "nbytes": 24780800,
2290
+ "records": [
2291
+ {
2292
+ "name": "model.layers.19.mlp.gate_up_proj.q_weight",
2293
+ "shape": [
2294
+ 11264,
2295
+ 256
2296
+ ],
2297
+ "dtype": "uint32",
2298
+ "format": "f32-to-bf16",
2299
+ "nbytes": 11534336,
2300
+ "byteOffset": 0
2301
+ },
2302
+ {
2303
+ "name": "model.layers.19.mlp.gate_up_proj.q_scale",
2304
+ "shape": [
2305
+ 11264,
2306
+ 64
2307
+ ],
2308
+ "dtype": "float32",
2309
+ "format": "f32-to-bf16",
2310
+ "nbytes": 1441792,
2311
+ "byteOffset": 11534336
2312
+ },
2313
+ {
2314
+ "name": "model.layers.19.mlp.down_proj.q_weight",
2315
+ "shape": [
2316
+ 2048,
2317
+ 704
2318
+ ],
2319
+ "dtype": "uint32",
2320
+ "format": "f32-to-bf16",
2321
+ "nbytes": 5767168,
2322
+ "byteOffset": 12976128
2323
+ },
2324
+ {
2325
+ "name": "model.layers.19.mlp.down_proj.q_scale",
2326
+ "shape": [
2327
+ 2048,
2328
+ 176
2329
+ ],
2330
+ "dtype": "float32",
2331
+ "format": "f32-to-bf16",
2332
+ "nbytes": 720896,
2333
+ "byteOffset": 18743296
2334
+ },
2335
+ {
2336
+ "name": "model.layers.19.input_layernorm.weight",
2337
+ "shape": [
2338
+ 2048
2339
+ ],
2340
+ "dtype": "float32",
2341
+ "format": "f32-to-bf16",
2342
+ "nbytes": 4096,
2343
+ "byteOffset": 19464192
2344
+ },
2345
+ {
2346
+ "name": "model.layers.19.post_attention_layernorm.weight",
2347
+ "shape": [
2348
+ 2048
2349
+ ],
2350
+ "dtype": "float32",
2351
+ "format": "f32-to-bf16",
2352
+ "nbytes": 4096,
2353
+ "byteOffset": 19468288
2354
+ },
2355
+ {
2356
+ "name": "model.layers.20.self_attn.qkv_proj.q_weight",
2357
+ "shape": [
2358
+ 2560,
2359
+ 256
2360
+ ],
2361
+ "dtype": "uint32",
2362
+ "format": "f32-to-bf16",
2363
+ "nbytes": 2621440,
2364
+ "byteOffset": 19472384
2365
+ },
2366
+ {
2367
+ "name": "model.layers.20.self_attn.qkv_proj.q_scale",
2368
+ "shape": [
2369
+ 2560,
2370
+ 64
2371
+ ],
2372
+ "dtype": "float32",
2373
+ "format": "f32-to-bf16",
2374
+ "nbytes": 327680,
2375
+ "byteOffset": 22093824
2376
+ },
2377
+ {
2378
+ "name": "model.layers.20.self_attn.o_proj.q_weight",
2379
+ "shape": [
2380
+ 2048,
2381
+ 256
2382
+ ],
2383
+ "dtype": "uint32",
2384
+ "format": "f32-to-bf16",
2385
+ "nbytes": 2097152,
2386
+ "byteOffset": 22421504
2387
+ },
2388
+ {
2389
+ "name": "model.layers.20.self_attn.o_proj.q_scale",
2390
+ "shape": [
2391
+ 2048,
2392
+ 64
2393
+ ],
2394
+ "dtype": "float32",
2395
+ "format": "f32-to-bf16",
2396
+ "nbytes": 262144,
2397
+ "byteOffset": 24518656
2398
+ }
2399
+ ],
2400
+ "md5sum": "4b6e5593f9e640846c38651fd89f279c"
2401
+ },
2402
+ {
2403
+ "dataPath": "params_shard_21.bin",
2404
+ "format": "raw-shard",
2405
+ "nbytes": 24780800,
2406
+ "records": [
2407
+ {
2408
+ "name": "model.layers.20.mlp.gate_up_proj.q_weight",
2409
+ "shape": [
2410
+ 11264,
2411
+ 256
2412
+ ],
2413
+ "dtype": "uint32",
2414
+ "format": "f32-to-bf16",
2415
+ "nbytes": 11534336,
2416
+ "byteOffset": 0
2417
+ },
2418
+ {
2419
+ "name": "model.layers.20.mlp.gate_up_proj.q_scale",
2420
+ "shape": [
2421
+ 11264,
2422
+ 64
2423
+ ],
2424
+ "dtype": "float32",
2425
+ "format": "f32-to-bf16",
2426
+ "nbytes": 1441792,
2427
+ "byteOffset": 11534336
2428
+ },
2429
+ {
2430
+ "name": "model.layers.20.mlp.down_proj.q_weight",
2431
+ "shape": [
2432
+ 2048,
2433
+ 704
2434
+ ],
2435
+ "dtype": "uint32",
2436
+ "format": "f32-to-bf16",
2437
+ "nbytes": 5767168,
2438
+ "byteOffset": 12976128
2439
+ },
2440
+ {
2441
+ "name": "model.layers.20.mlp.down_proj.q_scale",
2442
+ "shape": [
2443
+ 2048,
2444
+ 176
2445
+ ],
2446
+ "dtype": "float32",
2447
+ "format": "f32-to-bf16",
2448
+ "nbytes": 720896,
2449
+ "byteOffset": 18743296
2450
+ },
2451
+ {
2452
+ "name": "model.layers.20.input_layernorm.weight",
2453
+ "shape": [
2454
+ 2048
2455
+ ],
2456
+ "dtype": "float32",
2457
+ "format": "f32-to-bf16",
2458
+ "nbytes": 4096,
2459
+ "byteOffset": 19464192
2460
+ },
2461
+ {
2462
+ "name": "model.layers.20.post_attention_layernorm.weight",
2463
+ "shape": [
2464
+ 2048
2465
+ ],
2466
+ "dtype": "float32",
2467
+ "format": "f32-to-bf16",
2468
+ "nbytes": 4096,
2469
+ "byteOffset": 19468288
2470
+ },
2471
+ {
2472
+ "name": "model.layers.21.self_attn.qkv_proj.q_weight",
2473
+ "shape": [
2474
+ 2560,
2475
+ 256
2476
+ ],
2477
+ "dtype": "uint32",
2478
+ "format": "f32-to-bf16",
2479
+ "nbytes": 2621440,
2480
+ "byteOffset": 19472384
2481
+ },
2482
+ {
2483
+ "name": "model.layers.21.self_attn.qkv_proj.q_scale",
2484
+ "shape": [
2485
+ 2560,
2486
+ 64
2487
+ ],
2488
+ "dtype": "float32",
2489
+ "format": "f32-to-bf16",
2490
+ "nbytes": 327680,
2491
+ "byteOffset": 22093824
2492
+ },
2493
+ {
2494
+ "name": "model.layers.21.self_attn.o_proj.q_weight",
2495
+ "shape": [
2496
+ 2048,
2497
+ 256
2498
+ ],
2499
+ "dtype": "uint32",
2500
+ "format": "f32-to-bf16",
2501
+ "nbytes": 2097152,
2502
+ "byteOffset": 22421504
2503
+ },
2504
+ {
2505
+ "name": "model.layers.21.self_attn.o_proj.q_scale",
2506
+ "shape": [
2507
+ 2048,
2508
+ 64
2509
+ ],
2510
+ "dtype": "float32",
2511
+ "format": "f32-to-bf16",
2512
+ "nbytes": 262144,
2513
+ "byteOffset": 24518656
2514
+ }
2515
+ ],
2516
+ "md5sum": "a74f48bc17bc757ea9d4c75002229241"
2517
+ },
2518
+ {
2519
+ "dataPath": "params_shard_22.bin",
2520
+ "format": "raw-shard",
2521
+ "nbytes": 32771072,
2522
+ "records": [
2523
+ {
2524
+ "name": "lm_head.q_weight",
2525
+ "shape": [
2526
+ 32003,
2527
+ 256
2528
+ ],
2529
+ "dtype": "uint32",
2530
+ "format": "f32-to-bf16",
2531
+ "nbytes": 32771072,
2532
+ "byteOffset": 0
2533
+ }
2534
+ ],
2535
+ "md5sum": "9a1c53a4e8ad27cf8d5fb1fcc5c333c2"
2536
+ },
2537
+ {
2538
+ "dataPath": "params_shard_23.bin",
2539
+ "format": "raw-shard",
2540
+ "nbytes": 23572864,
2541
+ "records": [
2542
+ {
2543
+ "name": "model.layers.21.mlp.gate_up_proj.q_weight",
2544
+ "shape": [
2545
+ 11264,
2546
+ 256
2547
+ ],
2548
+ "dtype": "uint32",
2549
+ "format": "f32-to-bf16",
2550
+ "nbytes": 11534336,
2551
+ "byteOffset": 0
2552
+ },
2553
+ {
2554
+ "name": "model.layers.21.mlp.gate_up_proj.q_scale",
2555
+ "shape": [
2556
+ 11264,
2557
+ 64
2558
+ ],
2559
+ "dtype": "float32",
2560
+ "format": "f32-to-bf16",
2561
+ "nbytes": 1441792,
2562
+ "byteOffset": 11534336
2563
+ },
2564
+ {
2565
+ "name": "model.layers.21.mlp.down_proj.q_weight",
2566
+ "shape": [
2567
+ 2048,
2568
+ 704
2569
+ ],
2570
+ "dtype": "uint32",
2571
+ "format": "f32-to-bf16",
2572
+ "nbytes": 5767168,
2573
+ "byteOffset": 12976128
2574
+ },
2575
+ {
2576
+ "name": "model.layers.21.mlp.down_proj.q_scale",
2577
+ "shape": [
2578
+ 2048,
2579
+ 176
2580
+ ],
2581
+ "dtype": "float32",
2582
+ "format": "f32-to-bf16",
2583
+ "nbytes": 720896,
2584
+ "byteOffset": 18743296
2585
+ },
2586
+ {
2587
+ "name": "model.layers.21.input_layernorm.weight",
2588
+ "shape": [
2589
+ 2048
2590
+ ],
2591
+ "dtype": "float32",
2592
+ "format": "f32-to-bf16",
2593
+ "nbytes": 4096,
2594
+ "byteOffset": 19464192
2595
+ },
2596
+ {
2597
+ "name": "model.layers.21.post_attention_layernorm.weight",
2598
+ "shape": [
2599
+ 2048
2600
+ ],
2601
+ "dtype": "float32",
2602
+ "format": "f32-to-bf16",
2603
+ "nbytes": 4096,
2604
+ "byteOffset": 19468288
2605
+ },
2606
+ {
2607
+ "name": "model.norm.weight",
2608
+ "shape": [
2609
+ 2048
2610
+ ],
2611
+ "dtype": "float32",
2612
+ "format": "f32-to-bf16",
2613
+ "nbytes": 4096,
2614
+ "byteOffset": 19472384
2615
+ },
2616
+ {
2617
+ "name": "lm_head.q_scale",
2618
+ "shape": [
2619
+ 32003,
2620
+ 64
2621
+ ],
2622
+ "dtype": "float32",
2623
+ "format": "f32-to-bf16",
2624
+ "nbytes": 4096384,
2625
+ "byteOffset": 19476480
2626
+ }
2627
+ ],
2628
+ "md5sum": "cc2bb39dd7372753ef316485aa555550"
2629
+ }
2630
+ ]
2631
+ }
params_shard_0.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:615bdcb9fc67660c60189c3c4d74d7372b81a125ab6f706dd046e9a8134edcd4
3
+ size 32771072
params_shard_1.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bcdcae6b14b5ae653796a29f606012afe85c427e29020a2ae17c72547082a970
3
+ size 31826304
params_shard_10.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f00453444ae47a2058e3ed9269461dea2a1622be5d1e59995329b00df4240cc1
3
+ size 24780800
params_shard_11.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7d2ed2a71ccae19c871176ab107b1f88f54857f987add6dde9317950030b0add
3
+ size 24780800
params_shard_12.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4efad76847e2bcb6a625edb9b600e2c22fb2d7ea1c0a47ab234b0529ccb14630
3
+ size 24780800
params_shard_13.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a4a14922b33ee8809150dc8a90fd1766d107ca943bfe89f59e54bb723c73cac6
3
+ size 24780800
params_shard_14.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a9b5427f34ef39204c0128dade004397cfa25d1b5a3f9782923bf5f964275f67
3
+ size 24780800
params_shard_15.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6e505496928c62945e4bade677cfb7c028b314167900ef5533dded986a3b656f
3
+ size 24780800
params_shard_16.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a9a0e6a8cf847128eae7758316fc41568ef6ae2c3ff51c8d45d11927b0d323ca
3
+ size 24780800
params_shard_17.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4a4004ff98c6dc5970f0accf4f465e0f1f0d34c918413edc497f58c6f0d964cd
3
+ size 24780800
params_shard_18.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:edb26028cb6fb23b50156d9c154e3ad81d719687badd9b4282035cb3f8e259a6
3
+ size 24780800
params_shard_19.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:12d71eea6b4be098bf2f9db3cc20a859d50e80769d7cfc88195b3708e6ef2949
3
+ size 24780800
params_shard_2.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:44419fe02124d541c1b07b0b8297fff313c81a848db233a223c60f357f7e1915
3
+ size 27140096
params_shard_20.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0bcdecfe536da578fc389adc39475ae850510f5be8aba30ee6b2a3dfa19dae85
3
+ size 24780800
params_shard_21.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:03a8f21ab4e2093a0c24515f34c41adb7c6935b649a6db0b204f34afbf225562
3
+ size 24780800
params_shard_22.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7f41b369bff60ccb0fd4ae5ec2d1108ba673b86657e7f3da5eef8e8d48913952
3
+ size 32771072
params_shard_23.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c5bb41dbc9d6289bd05af2c775d1650a2698ab484c9ceb845d36f16a6a318b82
3
+ size 23572864
params_shard_3.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f31c47632045d87fdfa7e5d65e6c81a6a210e3c522b9fb158442603561b1a984
3
+ size 24780800
params_shard_4.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:682d55552c1f4806cd9ab3fb39cc758cd783f54d29e211213702b08df3691d35
3
+ size 24780800
params_shard_5.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7898470c4ac36a285f4534a0d24f65cab2214191c4194dcd0a7caeb27ace1443
3
+ size 24780800
params_shard_6.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:672b1b6b65c1ce8b9ffb9fcaee85bd6633542a42376e095aeea4062ca1e828a6
3
+ size 24780800
params_shard_7.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8655741aa8547955e2783cf8e59fffb57c2151c3561fd07832f218dd27b6e072
3
+ size 24780800
params_shard_8.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5083560c8499d69802f3cc8e40585f3fcc4c29dbbae7eb6497094156ba4d1fd5
3
+ size 24780800
params_shard_9.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e727b4a88123ce57af8fb8a060679229d703946fe73652249aa93c9bfb5f62bf
3
+ size 24780800
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347
3
+ size 499723
tokenizer_config.json ADDED
@@ -0,0 +1,34 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "__type": "AddedToken",
4
+ "content": "<s>",
5
+ "lstrip": false,
6
+ "normalized": false,
7
+ "rstrip": false,
8
+ "single_word": false
9
+ },
10
+ "clean_up_tokenization_spaces": false,
11
+ "eos_token": {
12
+ "__type": "AddedToken",
13
+ "content": "</s>",
14
+ "lstrip": false,
15
+ "normalized": false,
16
+ "rstrip": false,
17
+ "single_word": false
18
+ },
19
+ "legacy": false,
20
+ "model_max_length": 1000000000000000019884624838656,
21
+ "pad_token": null,
22
+ "padding_side": "right",
23
+ "sp_model_kwargs": {},
24
+ "tokenizer_class": "LlamaTokenizer",
25
+ "unk_token": {
26
+ "__type": "AddedToken",
27
+ "content": "<unk>",
28
+ "lstrip": false,
29
+ "normalized": false,
30
+ "rstrip": false,
31
+ "single_word": false
32
+ },
33
+ "use_default_system_prompt": true
34
+ }