cik009 commited on
Commit
7d9a633
1 Parent(s): ec4854c

Upload 43 files

Browse files
.gitattributes CHANGED
@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
+ tokenizer.json filter=lfs diff=lfs merge=lfs -text
mlc-chat-config.json ADDED
@@ -0,0 +1,45 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "model_type": "gemma",
3
+ "quantization": "q4f16_0",
4
+ "model_config": {
5
+ "hidden_size": 2048,
6
+ "hidden_act": "gelu",
7
+ "intermediate_size": 16384,
8
+ "attention_bias": false,
9
+ "num_attention_heads": 8,
10
+ "num_key_value_heads": 1,
11
+ "head_dim": 256,
12
+ "num_hidden_layers": 18,
13
+ "rms_norm_eps": 1e-06,
14
+ "vocab_size": 256000,
15
+ "position_embedding_base": 10000.0,
16
+ "context_window_size": 8192,
17
+ "prefill_chunk_size": 8192,
18
+ "tensor_parallel_shards": 1,
19
+ "max_batch_size": 80
20
+ },
21
+ "vocab_size": 256000,
22
+ "context_window_size": 8192,
23
+ "sliding_window_size": -1,
24
+ "prefill_chunk_size": 8192,
25
+ "attention_sink_size": -1,
26
+ "tensor_parallel_shards": 1,
27
+ "mean_gen_len": 128,
28
+ "max_gen_len": 512,
29
+ "shift_fill_factor": 0.3,
30
+ "temperature": 0.7,
31
+ "presence_penalty": 0.0,
32
+ "frequency_penalty": 0.0,
33
+ "repetition_penalty": 1.0,
34
+ "top_p": 0.95,
35
+ "conv_template": "gemma_instruction",
36
+ "pad_token_id": 0,
37
+ "bos_token_id": 2,
38
+ "eos_token_id": 1,
39
+ "tokenizer_files": [
40
+ "tokenizer.model",
41
+ "tokenizer.json",
42
+ "tokenizer_config.json"
43
+ ],
44
+ "version": "0.1.0"
45
+ }
ndarray-cache.json ADDED
@@ -0,0 +1,2289 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metadata": {
3
+ "ParamSize": 183,
4
+ "ParamBytes": 1409830912.0,
5
+ "BitsPerParam": 4.500347711112945
6
+ },
7
+ "records": [
8
+ {
9
+ "dataPath": "params_shard_0.bin",
10
+ "format": "raw-shard",
11
+ "nbytes": 262144000,
12
+ "records": [
13
+ {
14
+ "name": "model.embed_tokens.q_weight",
15
+ "shape": [
16
+ 256000,
17
+ 256
18
+ ],
19
+ "dtype": "uint32",
20
+ "format": "f32-to-bf16",
21
+ "nbytes": 262144000,
22
+ "byteOffset": 0
23
+ }
24
+ ],
25
+ "md5sum": "76834a66b447ed959a5dba4862338312"
26
+ },
27
+ {
28
+ "dataPath": "params_shard_1.bin",
29
+ "format": "raw-shard",
30
+ "nbytes": 16777216,
31
+ "records": [
32
+ {
33
+ "name": "model.layers.0.mlp.down_proj.q_weight",
34
+ "shape": [
35
+ 2048,
36
+ 2048
37
+ ],
38
+ "dtype": "uint32",
39
+ "format": "f32-to-bf16",
40
+ "nbytes": 16777216,
41
+ "byteOffset": 0
42
+ }
43
+ ],
44
+ "md5sum": "5706629174af79cfeb92439752fdc945"
45
+ },
46
+ {
47
+ "dataPath": "params_shard_2.bin",
48
+ "format": "raw-shard",
49
+ "nbytes": 32772096,
50
+ "records": [
51
+ {
52
+ "name": "model.embed_tokens.q_scale",
53
+ "shape": [
54
+ 256000,
55
+ 64
56
+ ],
57
+ "dtype": "float16",
58
+ "format": "f32-to-bf16",
59
+ "nbytes": 32768000,
60
+ "byteOffset": 0
61
+ },
62
+ {
63
+ "name": "model.layers.0.input_layernorm.weight",
64
+ "shape": [
65
+ 2048
66
+ ],
67
+ "dtype": "float16",
68
+ "format": "f32-to-bf16",
69
+ "nbytes": 4096,
70
+ "byteOffset": 32768000
71
+ }
72
+ ],
73
+ "md5sum": "f379b208c95b8217c994896c7b6f2498"
74
+ },
75
+ {
76
+ "dataPath": "params_shard_3.bin",
77
+ "format": "raw-shard",
78
+ "nbytes": 33554432,
79
+ "records": [
80
+ {
81
+ "name": "model.layers.0.mlp.gate_up_proj.q_weight",
82
+ "shape": [
83
+ 256,
84
+ 32768
85
+ ],
86
+ "dtype": "uint32",
87
+ "format": "f32-to-bf16",
88
+ "nbytes": 33554432,
89
+ "byteOffset": 0
90
+ }
91
+ ],
92
+ "md5sum": "9e422ea69524c5742c9dd90ee1489033"
93
+ },
94
+ {
95
+ "dataPath": "params_shard_4.bin",
96
+ "format": "raw-shard",
97
+ "nbytes": 33554432,
98
+ "records": [
99
+ {
100
+ "name": "model.layers.1.mlp.gate_up_proj.q_weight",
101
+ "shape": [
102
+ 256,
103
+ 32768
104
+ ],
105
+ "dtype": "uint32",
106
+ "format": "f32-to-bf16",
107
+ "nbytes": 33554432,
108
+ "byteOffset": 0
109
+ }
110
+ ],
111
+ "md5sum": "ef73d55a46171c99c30a4b590a9ca5ee"
112
+ },
113
+ {
114
+ "dataPath": "params_shard_5.bin",
115
+ "format": "raw-shard",
116
+ "nbytes": 30482432,
117
+ "records": [
118
+ {
119
+ "name": "model.layers.0.mlp.down_proj.q_scale",
120
+ "shape": [
121
+ 512,
122
+ 2048
123
+ ],
124
+ "dtype": "float16",
125
+ "format": "f32-to-bf16",
126
+ "nbytes": 2097152,
127
+ "byteOffset": 0
128
+ },
129
+ {
130
+ "name": "model.layers.0.mlp.gate_up_proj.q_scale",
131
+ "shape": [
132
+ 64,
133
+ 32768
134
+ ],
135
+ "dtype": "float16",
136
+ "format": "f32-to-bf16",
137
+ "nbytes": 4194304,
138
+ "byteOffset": 2097152
139
+ },
140
+ {
141
+ "name": "model.layers.0.post_attention_layernorm.weight",
142
+ "shape": [
143
+ 2048
144
+ ],
145
+ "dtype": "float16",
146
+ "format": "f32-to-bf16",
147
+ "nbytes": 4096,
148
+ "byteOffset": 6291456
149
+ },
150
+ {
151
+ "name": "model.layers.0.self_attn.qkv_proj.q_weight",
152
+ "shape": [
153
+ 256,
154
+ 2560
155
+ ],
156
+ "dtype": "uint32",
157
+ "format": "f32-to-bf16",
158
+ "nbytes": 2621440,
159
+ "byteOffset": 6295552
160
+ },
161
+ {
162
+ "name": "model.layers.0.self_attn.qkv_proj.q_scale",
163
+ "shape": [
164
+ 64,
165
+ 2560
166
+ ],
167
+ "dtype": "float16",
168
+ "format": "f32-to-bf16",
169
+ "nbytes": 327680,
170
+ "byteOffset": 8916992
171
+ },
172
+ {
173
+ "name": "model.layers.0.self_attn.o_proj.q_weight",
174
+ "shape": [
175
+ 256,
176
+ 2048
177
+ ],
178
+ "dtype": "uint32",
179
+ "format": "f32-to-bf16",
180
+ "nbytes": 2097152,
181
+ "byteOffset": 9244672
182
+ },
183
+ {
184
+ "name": "model.layers.0.self_attn.o_proj.q_scale",
185
+ "shape": [
186
+ 64,
187
+ 2048
188
+ ],
189
+ "dtype": "float16",
190
+ "format": "f32-to-bf16",
191
+ "nbytes": 262144,
192
+ "byteOffset": 11341824
193
+ },
194
+ {
195
+ "name": "model.layers.1.input_layernorm.weight",
196
+ "shape": [
197
+ 2048
198
+ ],
199
+ "dtype": "float16",
200
+ "format": "f32-to-bf16",
201
+ "nbytes": 4096,
202
+ "byteOffset": 11603968
203
+ },
204
+ {
205
+ "name": "model.layers.1.mlp.down_proj.q_weight",
206
+ "shape": [
207
+ 2048,
208
+ 2048
209
+ ],
210
+ "dtype": "uint32",
211
+ "format": "f32-to-bf16",
212
+ "nbytes": 16777216,
213
+ "byteOffset": 11608064
214
+ },
215
+ {
216
+ "name": "model.layers.1.mlp.down_proj.q_scale",
217
+ "shape": [
218
+ 512,
219
+ 2048
220
+ ],
221
+ "dtype": "float16",
222
+ "format": "f32-to-bf16",
223
+ "nbytes": 2097152,
224
+ "byteOffset": 28385280
225
+ }
226
+ ],
227
+ "md5sum": "7c1509918451a49df944329e01e3549c"
228
+ },
229
+ {
230
+ "dataPath": "params_shard_6.bin",
231
+ "format": "raw-shard",
232
+ "nbytes": 33554432,
233
+ "records": [
234
+ {
235
+ "name": "model.layers.10.mlp.gate_up_proj.q_weight",
236
+ "shape": [
237
+ 256,
238
+ 32768
239
+ ],
240
+ "dtype": "uint32",
241
+ "format": "f32-to-bf16",
242
+ "nbytes": 33554432,
243
+ "byteOffset": 0
244
+ }
245
+ ],
246
+ "md5sum": "70c3a4032464feb4037d1cb33950cfea"
247
+ },
248
+ {
249
+ "dataPath": "params_shard_7.bin",
250
+ "format": "raw-shard",
251
+ "nbytes": 32583680,
252
+ "records": [
253
+ {
254
+ "name": "model.layers.1.mlp.gate_up_proj.q_scale",
255
+ "shape": [
256
+ 64,
257
+ 32768
258
+ ],
259
+ "dtype": "float16",
260
+ "format": "f32-to-bf16",
261
+ "nbytes": 4194304,
262
+ "byteOffset": 0
263
+ },
264
+ {
265
+ "name": "model.layers.1.post_attention_layernorm.weight",
266
+ "shape": [
267
+ 2048
268
+ ],
269
+ "dtype": "float16",
270
+ "format": "f32-to-bf16",
271
+ "nbytes": 4096,
272
+ "byteOffset": 4194304
273
+ },
274
+ {
275
+ "name": "model.layers.1.self_attn.qkv_proj.q_weight",
276
+ "shape": [
277
+ 256,
278
+ 2560
279
+ ],
280
+ "dtype": "uint32",
281
+ "format": "f32-to-bf16",
282
+ "nbytes": 2621440,
283
+ "byteOffset": 4198400
284
+ },
285
+ {
286
+ "name": "model.layers.1.self_attn.qkv_proj.q_scale",
287
+ "shape": [
288
+ 64,
289
+ 2560
290
+ ],
291
+ "dtype": "float16",
292
+ "format": "f32-to-bf16",
293
+ "nbytes": 327680,
294
+ "byteOffset": 6819840
295
+ },
296
+ {
297
+ "name": "model.layers.1.self_attn.o_proj.q_weight",
298
+ "shape": [
299
+ 256,
300
+ 2048
301
+ ],
302
+ "dtype": "uint32",
303
+ "format": "f32-to-bf16",
304
+ "nbytes": 2097152,
305
+ "byteOffset": 7147520
306
+ },
307
+ {
308
+ "name": "model.layers.1.self_attn.o_proj.q_scale",
309
+ "shape": [
310
+ 64,
311
+ 2048
312
+ ],
313
+ "dtype": "float16",
314
+ "format": "f32-to-bf16",
315
+ "nbytes": 262144,
316
+ "byteOffset": 9244672
317
+ },
318
+ {
319
+ "name": "model.layers.10.input_layernorm.weight",
320
+ "shape": [
321
+ 2048
322
+ ],
323
+ "dtype": "float16",
324
+ "format": "f32-to-bf16",
325
+ "nbytes": 4096,
326
+ "byteOffset": 9506816
327
+ },
328
+ {
329
+ "name": "model.layers.10.mlp.down_proj.q_weight",
330
+ "shape": [
331
+ 2048,
332
+ 2048
333
+ ],
334
+ "dtype": "uint32",
335
+ "format": "f32-to-bf16",
336
+ "nbytes": 16777216,
337
+ "byteOffset": 9510912
338
+ },
339
+ {
340
+ "name": "model.layers.10.mlp.down_proj.q_scale",
341
+ "shape": [
342
+ 512,
343
+ 2048
344
+ ],
345
+ "dtype": "float16",
346
+ "format": "f32-to-bf16",
347
+ "nbytes": 2097152,
348
+ "byteOffset": 26288128
349
+ },
350
+ {
351
+ "name": "model.layers.10.mlp.gate_up_proj.q_scale",
352
+ "shape": [
353
+ 64,
354
+ 32768
355
+ ],
356
+ "dtype": "float16",
357
+ "format": "f32-to-bf16",
358
+ "nbytes": 4194304,
359
+ "byteOffset": 28385280
360
+ },
361
+ {
362
+ "name": "model.layers.10.post_attention_layernorm.weight",
363
+ "shape": [
364
+ 2048
365
+ ],
366
+ "dtype": "float16",
367
+ "format": "f32-to-bf16",
368
+ "nbytes": 4096,
369
+ "byteOffset": 32579584
370
+ }
371
+ ],
372
+ "md5sum": "3e80e2b4cd5ea1dde1b463d8682285b4"
373
+ },
374
+ {
375
+ "dataPath": "params_shard_8.bin",
376
+ "format": "raw-shard",
377
+ "nbytes": 33554432,
378
+ "records": [
379
+ {
380
+ "name": "model.layers.11.mlp.gate_up_proj.q_weight",
381
+ "shape": [
382
+ 256,
383
+ 32768
384
+ ],
385
+ "dtype": "uint32",
386
+ "format": "f32-to-bf16",
387
+ "nbytes": 33554432,
388
+ "byteOffset": 0
389
+ }
390
+ ],
391
+ "md5sum": "5b9bc15793441be20aa77d2d4a350f28"
392
+ },
393
+ {
394
+ "dataPath": "params_shard_9.bin",
395
+ "format": "raw-shard",
396
+ "nbytes": 33431552,
397
+ "records": [
398
+ {
399
+ "name": "model.layers.10.self_attn.qkv_proj.q_weight",
400
+ "shape": [
401
+ 256,
402
+ 2560
403
+ ],
404
+ "dtype": "uint32",
405
+ "format": "f32-to-bf16",
406
+ "nbytes": 2621440,
407
+ "byteOffset": 0
408
+ },
409
+ {
410
+ "name": "model.layers.10.self_attn.qkv_proj.q_scale",
411
+ "shape": [
412
+ 64,
413
+ 2560
414
+ ],
415
+ "dtype": "float16",
416
+ "format": "f32-to-bf16",
417
+ "nbytes": 327680,
418
+ "byteOffset": 2621440
419
+ },
420
+ {
421
+ "name": "model.layers.10.self_attn.o_proj.q_weight",
422
+ "shape": [
423
+ 256,
424
+ 2048
425
+ ],
426
+ "dtype": "uint32",
427
+ "format": "f32-to-bf16",
428
+ "nbytes": 2097152,
429
+ "byteOffset": 2949120
430
+ },
431
+ {
432
+ "name": "model.layers.10.self_attn.o_proj.q_scale",
433
+ "shape": [
434
+ 64,
435
+ 2048
436
+ ],
437
+ "dtype": "float16",
438
+ "format": "f32-to-bf16",
439
+ "nbytes": 262144,
440
+ "byteOffset": 5046272
441
+ },
442
+ {
443
+ "name": "model.layers.11.input_layernorm.weight",
444
+ "shape": [
445
+ 2048
446
+ ],
447
+ "dtype": "float16",
448
+ "format": "f32-to-bf16",
449
+ "nbytes": 4096,
450
+ "byteOffset": 5308416
451
+ },
452
+ {
453
+ "name": "model.layers.11.mlp.down_proj.q_weight",
454
+ "shape": [
455
+ 2048,
456
+ 2048
457
+ ],
458
+ "dtype": "uint32",
459
+ "format": "f32-to-bf16",
460
+ "nbytes": 16777216,
461
+ "byteOffset": 5312512
462
+ },
463
+ {
464
+ "name": "model.layers.11.mlp.down_proj.q_scale",
465
+ "shape": [
466
+ 512,
467
+ 2048
468
+ ],
469
+ "dtype": "float16",
470
+ "format": "f32-to-bf16",
471
+ "nbytes": 2097152,
472
+ "byteOffset": 22089728
473
+ },
474
+ {
475
+ "name": "model.layers.11.mlp.gate_up_proj.q_scale",
476
+ "shape": [
477
+ 64,
478
+ 32768
479
+ ],
480
+ "dtype": "float16",
481
+ "format": "f32-to-bf16",
482
+ "nbytes": 4194304,
483
+ "byteOffset": 24186880
484
+ },
485
+ {
486
+ "name": "model.layers.11.post_attention_layernorm.weight",
487
+ "shape": [
488
+ 2048
489
+ ],
490
+ "dtype": "float16",
491
+ "format": "f32-to-bf16",
492
+ "nbytes": 4096,
493
+ "byteOffset": 28381184
494
+ },
495
+ {
496
+ "name": "model.layers.11.self_attn.qkv_proj.q_weight",
497
+ "shape": [
498
+ 256,
499
+ 2560
500
+ ],
501
+ "dtype": "uint32",
502
+ "format": "f32-to-bf16",
503
+ "nbytes": 2621440,
504
+ "byteOffset": 28385280
505
+ },
506
+ {
507
+ "name": "model.layers.11.self_attn.qkv_proj.q_scale",
508
+ "shape": [
509
+ 64,
510
+ 2560
511
+ ],
512
+ "dtype": "float16",
513
+ "format": "f32-to-bf16",
514
+ "nbytes": 327680,
515
+ "byteOffset": 31006720
516
+ },
517
+ {
518
+ "name": "model.layers.11.self_attn.o_proj.q_weight",
519
+ "shape": [
520
+ 256,
521
+ 2048
522
+ ],
523
+ "dtype": "uint32",
524
+ "format": "f32-to-bf16",
525
+ "nbytes": 2097152,
526
+ "byteOffset": 31334400
527
+ }
528
+ ],
529
+ "md5sum": "d181ebf2d388531c80ef81a78648a1a2"
530
+ },
531
+ {
532
+ "dataPath": "params_shard_10.bin",
533
+ "format": "raw-shard",
534
+ "nbytes": 33554432,
535
+ "records": [
536
+ {
537
+ "name": "model.layers.12.mlp.gate_up_proj.q_weight",
538
+ "shape": [
539
+ 256,
540
+ 32768
541
+ ],
542
+ "dtype": "uint32",
543
+ "format": "f32-to-bf16",
544
+ "nbytes": 33554432,
545
+ "byteOffset": 0
546
+ }
547
+ ],
548
+ "md5sum": "c95904949c26b47f9837cef8668ba59f"
549
+ },
550
+ {
551
+ "dataPath": "params_shard_11.bin",
552
+ "format": "raw-shard",
553
+ "nbytes": 16777216,
554
+ "records": [
555
+ {
556
+ "name": "model.layers.13.mlp.down_proj.q_weight",
557
+ "shape": [
558
+ 2048,
559
+ 2048
560
+ ],
561
+ "dtype": "uint32",
562
+ "format": "f32-to-bf16",
563
+ "nbytes": 16777216,
564
+ "byteOffset": 0
565
+ }
566
+ ],
567
+ "md5sum": "42fdfdef261768182fa76f673d3dd20d"
568
+ },
569
+ {
570
+ "dataPath": "params_shard_12.bin",
571
+ "format": "raw-shard",
572
+ "nbytes": 33554432,
573
+ "records": [
574
+ {
575
+ "name": "model.layers.13.mlp.gate_up_proj.q_weight",
576
+ "shape": [
577
+ 256,
578
+ 32768
579
+ ],
580
+ "dtype": "uint32",
581
+ "format": "f32-to-bf16",
582
+ "nbytes": 33554432,
583
+ "byteOffset": 0
584
+ }
585
+ ],
586
+ "md5sum": "129b3a3b19f6ab65fda0eca5138ed013"
587
+ },
588
+ {
589
+ "dataPath": "params_shard_13.bin",
590
+ "format": "raw-shard",
591
+ "nbytes": 30748672,
592
+ "records": [
593
+ {
594
+ "name": "model.layers.11.self_attn.o_proj.q_scale",
595
+ "shape": [
596
+ 64,
597
+ 2048
598
+ ],
599
+ "dtype": "float16",
600
+ "format": "f32-to-bf16",
601
+ "nbytes": 262144,
602
+ "byteOffset": 0
603
+ },
604
+ {
605
+ "name": "model.layers.12.input_layernorm.weight",
606
+ "shape": [
607
+ 2048
608
+ ],
609
+ "dtype": "float16",
610
+ "format": "f32-to-bf16",
611
+ "nbytes": 4096,
612
+ "byteOffset": 262144
613
+ },
614
+ {
615
+ "name": "model.layers.12.mlp.down_proj.q_weight",
616
+ "shape": [
617
+ 2048,
618
+ 2048
619
+ ],
620
+ "dtype": "uint32",
621
+ "format": "f32-to-bf16",
622
+ "nbytes": 16777216,
623
+ "byteOffset": 266240
624
+ },
625
+ {
626
+ "name": "model.layers.12.mlp.down_proj.q_scale",
627
+ "shape": [
628
+ 512,
629
+ 2048
630
+ ],
631
+ "dtype": "float16",
632
+ "format": "f32-to-bf16",
633
+ "nbytes": 2097152,
634
+ "byteOffset": 17043456
635
+ },
636
+ {
637
+ "name": "model.layers.12.mlp.gate_up_proj.q_scale",
638
+ "shape": [
639
+ 64,
640
+ 32768
641
+ ],
642
+ "dtype": "float16",
643
+ "format": "f32-to-bf16",
644
+ "nbytes": 4194304,
645
+ "byteOffset": 19140608
646
+ },
647
+ {
648
+ "name": "model.layers.12.post_attention_layernorm.weight",
649
+ "shape": [
650
+ 2048
651
+ ],
652
+ "dtype": "float16",
653
+ "format": "f32-to-bf16",
654
+ "nbytes": 4096,
655
+ "byteOffset": 23334912
656
+ },
657
+ {
658
+ "name": "model.layers.12.self_attn.qkv_proj.q_weight",
659
+ "shape": [
660
+ 256,
661
+ 2560
662
+ ],
663
+ "dtype": "uint32",
664
+ "format": "f32-to-bf16",
665
+ "nbytes": 2621440,
666
+ "byteOffset": 23339008
667
+ },
668
+ {
669
+ "name": "model.layers.12.self_attn.qkv_proj.q_scale",
670
+ "shape": [
671
+ 64,
672
+ 2560
673
+ ],
674
+ "dtype": "float16",
675
+ "format": "f32-to-bf16",
676
+ "nbytes": 327680,
677
+ "byteOffset": 25960448
678
+ },
679
+ {
680
+ "name": "model.layers.12.self_attn.o_proj.q_weight",
681
+ "shape": [
682
+ 256,
683
+ 2048
684
+ ],
685
+ "dtype": "uint32",
686
+ "format": "f32-to-bf16",
687
+ "nbytes": 2097152,
688
+ "byteOffset": 26288128
689
+ },
690
+ {
691
+ "name": "model.layers.12.self_attn.o_proj.q_scale",
692
+ "shape": [
693
+ 64,
694
+ 2048
695
+ ],
696
+ "dtype": "float16",
697
+ "format": "f32-to-bf16",
698
+ "nbytes": 262144,
699
+ "byteOffset": 28385280
700
+ },
701
+ {
702
+ "name": "model.layers.13.input_layernorm.weight",
703
+ "shape": [
704
+ 2048
705
+ ],
706
+ "dtype": "float16",
707
+ "format": "f32-to-bf16",
708
+ "nbytes": 4096,
709
+ "byteOffset": 28647424
710
+ },
711
+ {
712
+ "name": "model.layers.13.mlp.down_proj.q_scale",
713
+ "shape": [
714
+ 512,
715
+ 2048
716
+ ],
717
+ "dtype": "float16",
718
+ "format": "f32-to-bf16",
719
+ "nbytes": 2097152,
720
+ "byteOffset": 28651520
721
+ }
722
+ ],
723
+ "md5sum": "2cc21823cc7b9346681fffebb1f75ed0"
724
+ },
725
+ {
726
+ "dataPath": "params_shard_14.bin",
727
+ "format": "raw-shard",
728
+ "nbytes": 33554432,
729
+ "records": [
730
+ {
731
+ "name": "model.layers.14.mlp.gate_up_proj.q_weight",
732
+ "shape": [
733
+ 256,
734
+ 32768
735
+ ],
736
+ "dtype": "uint32",
737
+ "format": "f32-to-bf16",
738
+ "nbytes": 33554432,
739
+ "byteOffset": 0
740
+ }
741
+ ],
742
+ "md5sum": "2979af8d22bf6c5a03eeff7925278147"
743
+ },
744
+ {
745
+ "dataPath": "params_shard_15.bin",
746
+ "format": "raw-shard",
747
+ "nbytes": 32583680,
748
+ "records": [
749
+ {
750
+ "name": "model.layers.13.mlp.gate_up_proj.q_scale",
751
+ "shape": [
752
+ 64,
753
+ 32768
754
+ ],
755
+ "dtype": "float16",
756
+ "format": "f32-to-bf16",
757
+ "nbytes": 4194304,
758
+ "byteOffset": 0
759
+ },
760
+ {
761
+ "name": "model.layers.13.post_attention_layernorm.weight",
762
+ "shape": [
763
+ 2048
764
+ ],
765
+ "dtype": "float16",
766
+ "format": "f32-to-bf16",
767
+ "nbytes": 4096,
768
+ "byteOffset": 4194304
769
+ },
770
+ {
771
+ "name": "model.layers.13.self_attn.qkv_proj.q_weight",
772
+ "shape": [
773
+ 256,
774
+ 2560
775
+ ],
776
+ "dtype": "uint32",
777
+ "format": "f32-to-bf16",
778
+ "nbytes": 2621440,
779
+ "byteOffset": 4198400
780
+ },
781
+ {
782
+ "name": "model.layers.13.self_attn.qkv_proj.q_scale",
783
+ "shape": [
784
+ 64,
785
+ 2560
786
+ ],
787
+ "dtype": "float16",
788
+ "format": "f32-to-bf16",
789
+ "nbytes": 327680,
790
+ "byteOffset": 6819840
791
+ },
792
+ {
793
+ "name": "model.layers.13.self_attn.o_proj.q_weight",
794
+ "shape": [
795
+ 256,
796
+ 2048
797
+ ],
798
+ "dtype": "uint32",
799
+ "format": "f32-to-bf16",
800
+ "nbytes": 2097152,
801
+ "byteOffset": 7147520
802
+ },
803
+ {
804
+ "name": "model.layers.13.self_attn.o_proj.q_scale",
805
+ "shape": [
806
+ 64,
807
+ 2048
808
+ ],
809
+ "dtype": "float16",
810
+ "format": "f32-to-bf16",
811
+ "nbytes": 262144,
812
+ "byteOffset": 9244672
813
+ },
814
+ {
815
+ "name": "model.layers.14.input_layernorm.weight",
816
+ "shape": [
817
+ 2048
818
+ ],
819
+ "dtype": "float16",
820
+ "format": "f32-to-bf16",
821
+ "nbytes": 4096,
822
+ "byteOffset": 9506816
823
+ },
824
+ {
825
+ "name": "model.layers.14.mlp.down_proj.q_weight",
826
+ "shape": [
827
+ 2048,
828
+ 2048
829
+ ],
830
+ "dtype": "uint32",
831
+ "format": "f32-to-bf16",
832
+ "nbytes": 16777216,
833
+ "byteOffset": 9510912
834
+ },
835
+ {
836
+ "name": "model.layers.14.mlp.down_proj.q_scale",
837
+ "shape": [
838
+ 512,
839
+ 2048
840
+ ],
841
+ "dtype": "float16",
842
+ "format": "f32-to-bf16",
843
+ "nbytes": 2097152,
844
+ "byteOffset": 26288128
845
+ },
846
+ {
847
+ "name": "model.layers.14.mlp.gate_up_proj.q_scale",
848
+ "shape": [
849
+ 64,
850
+ 32768
851
+ ],
852
+ "dtype": "float16",
853
+ "format": "f32-to-bf16",
854
+ "nbytes": 4194304,
855
+ "byteOffset": 28385280
856
+ },
857
+ {
858
+ "name": "model.layers.14.post_attention_layernorm.weight",
859
+ "shape": [
860
+ 2048
861
+ ],
862
+ "dtype": "float16",
863
+ "format": "f32-to-bf16",
864
+ "nbytes": 4096,
865
+ "byteOffset": 32579584
866
+ }
867
+ ],
868
+ "md5sum": "114e300ea2680084ae0f408a5a04b050"
869
+ },
870
+ {
871
+ "dataPath": "params_shard_16.bin",
872
+ "format": "raw-shard",
873
+ "nbytes": 33554432,
874
+ "records": [
875
+ {
876
+ "name": "model.layers.15.mlp.gate_up_proj.q_weight",
877
+ "shape": [
878
+ 256,
879
+ 32768
880
+ ],
881
+ "dtype": "uint32",
882
+ "format": "f32-to-bf16",
883
+ "nbytes": 33554432,
884
+ "byteOffset": 0
885
+ }
886
+ ],
887
+ "md5sum": "372da8dbb10e570f95d8a1b08993f6e6"
888
+ },
889
+ {
890
+ "dataPath": "params_shard_17.bin",
891
+ "format": "raw-shard",
892
+ "nbytes": 33431552,
893
+ "records": [
894
+ {
895
+ "name": "model.layers.14.self_attn.qkv_proj.q_weight",
896
+ "shape": [
897
+ 256,
898
+ 2560
899
+ ],
900
+ "dtype": "uint32",
901
+ "format": "f32-to-bf16",
902
+ "nbytes": 2621440,
903
+ "byteOffset": 0
904
+ },
905
+ {
906
+ "name": "model.layers.14.self_attn.qkv_proj.q_scale",
907
+ "shape": [
908
+ 64,
909
+ 2560
910
+ ],
911
+ "dtype": "float16",
912
+ "format": "f32-to-bf16",
913
+ "nbytes": 327680,
914
+ "byteOffset": 2621440
915
+ },
916
+ {
917
+ "name": "model.layers.14.self_attn.o_proj.q_weight",
918
+ "shape": [
919
+ 256,
920
+ 2048
921
+ ],
922
+ "dtype": "uint32",
923
+ "format": "f32-to-bf16",
924
+ "nbytes": 2097152,
925
+ "byteOffset": 2949120
926
+ },
927
+ {
928
+ "name": "model.layers.14.self_attn.o_proj.q_scale",
929
+ "shape": [
930
+ 64,
931
+ 2048
932
+ ],
933
+ "dtype": "float16",
934
+ "format": "f32-to-bf16",
935
+ "nbytes": 262144,
936
+ "byteOffset": 5046272
937
+ },
938
+ {
939
+ "name": "model.layers.15.input_layernorm.weight",
940
+ "shape": [
941
+ 2048
942
+ ],
943
+ "dtype": "float16",
944
+ "format": "f32-to-bf16",
945
+ "nbytes": 4096,
946
+ "byteOffset": 5308416
947
+ },
948
+ {
949
+ "name": "model.layers.15.mlp.down_proj.q_weight",
950
+ "shape": [
951
+ 2048,
952
+ 2048
953
+ ],
954
+ "dtype": "uint32",
955
+ "format": "f32-to-bf16",
956
+ "nbytes": 16777216,
957
+ "byteOffset": 5312512
958
+ },
959
+ {
960
+ "name": "model.layers.15.mlp.down_proj.q_scale",
961
+ "shape": [
962
+ 512,
963
+ 2048
964
+ ],
965
+ "dtype": "float16",
966
+ "format": "f32-to-bf16",
967
+ "nbytes": 2097152,
968
+ "byteOffset": 22089728
969
+ },
970
+ {
971
+ "name": "model.layers.15.mlp.gate_up_proj.q_scale",
972
+ "shape": [
973
+ 64,
974
+ 32768
975
+ ],
976
+ "dtype": "float16",
977
+ "format": "f32-to-bf16",
978
+ "nbytes": 4194304,
979
+ "byteOffset": 24186880
980
+ },
981
+ {
982
+ "name": "model.layers.15.post_attention_layernorm.weight",
983
+ "shape": [
984
+ 2048
985
+ ],
986
+ "dtype": "float16",
987
+ "format": "f32-to-bf16",
988
+ "nbytes": 4096,
989
+ "byteOffset": 28381184
990
+ },
991
+ {
992
+ "name": "model.layers.15.self_attn.qkv_proj.q_weight",
993
+ "shape": [
994
+ 256,
995
+ 2560
996
+ ],
997
+ "dtype": "uint32",
998
+ "format": "f32-to-bf16",
999
+ "nbytes": 2621440,
1000
+ "byteOffset": 28385280
1001
+ },
1002
+ {
1003
+ "name": "model.layers.15.self_attn.qkv_proj.q_scale",
1004
+ "shape": [
1005
+ 64,
1006
+ 2560
1007
+ ],
1008
+ "dtype": "float16",
1009
+ "format": "f32-to-bf16",
1010
+ "nbytes": 327680,
1011
+ "byteOffset": 31006720
1012
+ },
1013
+ {
1014
+ "name": "model.layers.15.self_attn.o_proj.q_weight",
1015
+ "shape": [
1016
+ 256,
1017
+ 2048
1018
+ ],
1019
+ "dtype": "uint32",
1020
+ "format": "f32-to-bf16",
1021
+ "nbytes": 2097152,
1022
+ "byteOffset": 31334400
1023
+ }
1024
+ ],
1025
+ "md5sum": "fedcf8c7d3a246baad9dbfac127436ac"
1026
+ },
1027
+ {
1028
+ "dataPath": "params_shard_18.bin",
1029
+ "format": "raw-shard",
1030
+ "nbytes": 33554432,
1031
+ "records": [
1032
+ {
1033
+ "name": "model.layers.16.mlp.gate_up_proj.q_weight",
1034
+ "shape": [
1035
+ 256,
1036
+ 32768
1037
+ ],
1038
+ "dtype": "uint32",
1039
+ "format": "f32-to-bf16",
1040
+ "nbytes": 33554432,
1041
+ "byteOffset": 0
1042
+ }
1043
+ ],
1044
+ "md5sum": "29d50626233e1af9d7c51a32a8cc2106"
1045
+ },
1046
+ {
1047
+ "dataPath": "params_shard_19.bin",
1048
+ "format": "raw-shard",
1049
+ "nbytes": 33554432,
1050
+ "records": [
1051
+ {
1052
+ "name": "model.layers.17.mlp.gate_up_proj.q_weight",
1053
+ "shape": [
1054
+ 256,
1055
+ 32768
1056
+ ],
1057
+ "dtype": "uint32",
1058
+ "format": "f32-to-bf16",
1059
+ "nbytes": 33554432,
1060
+ "byteOffset": 0
1061
+ }
1062
+ ],
1063
+ "md5sum": "4cdbd9a339239468f0d4d6570b66bdd3"
1064
+ },
1065
+ {
1066
+ "dataPath": "params_shard_20.bin",
1067
+ "format": "raw-shard",
1068
+ "nbytes": 32841728,
1069
+ "records": [
1070
+ {
1071
+ "name": "model.layers.15.self_attn.o_proj.q_scale",
1072
+ "shape": [
1073
+ 64,
1074
+ 2048
1075
+ ],
1076
+ "dtype": "float16",
1077
+ "format": "f32-to-bf16",
1078
+ "nbytes": 262144,
1079
+ "byteOffset": 0
1080
+ },
1081
+ {
1082
+ "name": "model.layers.16.input_layernorm.weight",
1083
+ "shape": [
1084
+ 2048
1085
+ ],
1086
+ "dtype": "float16",
1087
+ "format": "f32-to-bf16",
1088
+ "nbytes": 4096,
1089
+ "byteOffset": 262144
1090
+ },
1091
+ {
1092
+ "name": "model.layers.16.mlp.down_proj.q_weight",
1093
+ "shape": [
1094
+ 2048,
1095
+ 2048
1096
+ ],
1097
+ "dtype": "uint32",
1098
+ "format": "f32-to-bf16",
1099
+ "nbytes": 16777216,
1100
+ "byteOffset": 266240
1101
+ },
1102
+ {
1103
+ "name": "model.layers.16.mlp.down_proj.q_scale",
1104
+ "shape": [
1105
+ 512,
1106
+ 2048
1107
+ ],
1108
+ "dtype": "float16",
1109
+ "format": "f32-to-bf16",
1110
+ "nbytes": 2097152,
1111
+ "byteOffset": 17043456
1112
+ },
1113
+ {
1114
+ "name": "model.layers.16.mlp.gate_up_proj.q_scale",
1115
+ "shape": [
1116
+ 64,
1117
+ 32768
1118
+ ],
1119
+ "dtype": "float16",
1120
+ "format": "f32-to-bf16",
1121
+ "nbytes": 4194304,
1122
+ "byteOffset": 19140608
1123
+ },
1124
+ {
1125
+ "name": "model.layers.16.post_attention_layernorm.weight",
1126
+ "shape": [
1127
+ 2048
1128
+ ],
1129
+ "dtype": "float16",
1130
+ "format": "f32-to-bf16",
1131
+ "nbytes": 4096,
1132
+ "byteOffset": 23334912
1133
+ },
1134
+ {
1135
+ "name": "model.layers.16.self_attn.qkv_proj.q_weight",
1136
+ "shape": [
1137
+ 256,
1138
+ 2560
1139
+ ],
1140
+ "dtype": "uint32",
1141
+ "format": "f32-to-bf16",
1142
+ "nbytes": 2621440,
1143
+ "byteOffset": 23339008
1144
+ },
1145
+ {
1146
+ "name": "model.layers.16.self_attn.qkv_proj.q_scale",
1147
+ "shape": [
1148
+ 64,
1149
+ 2560
1150
+ ],
1151
+ "dtype": "float16",
1152
+ "format": "f32-to-bf16",
1153
+ "nbytes": 327680,
1154
+ "byteOffset": 25960448
1155
+ },
1156
+ {
1157
+ "name": "model.layers.16.self_attn.o_proj.q_weight",
1158
+ "shape": [
1159
+ 256,
1160
+ 2048
1161
+ ],
1162
+ "dtype": "uint32",
1163
+ "format": "f32-to-bf16",
1164
+ "nbytes": 2097152,
1165
+ "byteOffset": 26288128
1166
+ },
1167
+ {
1168
+ "name": "model.layers.16.self_attn.o_proj.q_scale",
1169
+ "shape": [
1170
+ 64,
1171
+ 2048
1172
+ ],
1173
+ "dtype": "float16",
1174
+ "format": "f32-to-bf16",
1175
+ "nbytes": 262144,
1176
+ "byteOffset": 28385280
1177
+ },
1178
+ {
1179
+ "name": "model.layers.17.mlp.gate_up_proj.q_scale",
1180
+ "shape": [
1181
+ 64,
1182
+ 32768
1183
+ ],
1184
+ "dtype": "float16",
1185
+ "format": "f32-to-bf16",
1186
+ "nbytes": 4194304,
1187
+ "byteOffset": 28647424
1188
+ }
1189
+ ],
1190
+ "md5sum": "9e4ecdaa6c3be7ca76f538c3f6324e9e"
1191
+ },
1192
+ {
1193
+ "dataPath": "params_shard_21.bin",
1194
+ "format": "raw-shard",
1195
+ "nbytes": 33554432,
1196
+ "records": [
1197
+ {
1198
+ "name": "model.layers.2.mlp.gate_up_proj.q_weight",
1199
+ "shape": [
1200
+ 256,
1201
+ 32768
1202
+ ],
1203
+ "dtype": "uint32",
1204
+ "format": "f32-to-bf16",
1205
+ "nbytes": 33554432,
1206
+ "byteOffset": 0
1207
+ }
1208
+ ],
1209
+ "md5sum": "d84dcee5f133dba81276f791fdf14b1e"
1210
+ },
1211
+ {
1212
+ "dataPath": "params_shard_22.bin",
1213
+ "format": "raw-shard",
1214
+ "nbytes": 33431552,
1215
+ "records": [
1216
+ {
1217
+ "name": "model.layers.17.self_attn.qkv_proj.q_weight",
1218
+ "shape": [
1219
+ 256,
1220
+ 2560
1221
+ ],
1222
+ "dtype": "uint32",
1223
+ "format": "f32-to-bf16",
1224
+ "nbytes": 2621440,
1225
+ "byteOffset": 0
1226
+ },
1227
+ {
1228
+ "name": "model.layers.17.self_attn.qkv_proj.q_scale",
1229
+ "shape": [
1230
+ 64,
1231
+ 2560
1232
+ ],
1233
+ "dtype": "float16",
1234
+ "format": "f32-to-bf16",
1235
+ "nbytes": 327680,
1236
+ "byteOffset": 2621440
1237
+ },
1238
+ {
1239
+ "name": "model.layers.17.self_attn.o_proj.q_weight",
1240
+ "shape": [
1241
+ 256,
1242
+ 2048
1243
+ ],
1244
+ "dtype": "uint32",
1245
+ "format": "f32-to-bf16",
1246
+ "nbytes": 2097152,
1247
+ "byteOffset": 2949120
1248
+ },
1249
+ {
1250
+ "name": "model.layers.17.self_attn.o_proj.q_scale",
1251
+ "shape": [
1252
+ 64,
1253
+ 2048
1254
+ ],
1255
+ "dtype": "float16",
1256
+ "format": "f32-to-bf16",
1257
+ "nbytes": 262144,
1258
+ "byteOffset": 5046272
1259
+ },
1260
+ {
1261
+ "name": "model.layers.2.input_layernorm.weight",
1262
+ "shape": [
1263
+ 2048
1264
+ ],
1265
+ "dtype": "float16",
1266
+ "format": "f32-to-bf16",
1267
+ "nbytes": 4096,
1268
+ "byteOffset": 5308416
1269
+ },
1270
+ {
1271
+ "name": "model.layers.2.mlp.down_proj.q_weight",
1272
+ "shape": [
1273
+ 2048,
1274
+ 2048
1275
+ ],
1276
+ "dtype": "uint32",
1277
+ "format": "f32-to-bf16",
1278
+ "nbytes": 16777216,
1279
+ "byteOffset": 5312512
1280
+ },
1281
+ {
1282
+ "name": "model.layers.2.mlp.down_proj.q_scale",
1283
+ "shape": [
1284
+ 512,
1285
+ 2048
1286
+ ],
1287
+ "dtype": "float16",
1288
+ "format": "f32-to-bf16",
1289
+ "nbytes": 2097152,
1290
+ "byteOffset": 22089728
1291
+ },
1292
+ {
1293
+ "name": "model.layers.2.mlp.gate_up_proj.q_scale",
1294
+ "shape": [
1295
+ 64,
1296
+ 32768
1297
+ ],
1298
+ "dtype": "float16",
1299
+ "format": "f32-to-bf16",
1300
+ "nbytes": 4194304,
1301
+ "byteOffset": 24186880
1302
+ },
1303
+ {
1304
+ "name": "model.layers.2.post_attention_layernorm.weight",
1305
+ "shape": [
1306
+ 2048
1307
+ ],
1308
+ "dtype": "float16",
1309
+ "format": "f32-to-bf16",
1310
+ "nbytes": 4096,
1311
+ "byteOffset": 28381184
1312
+ },
1313
+ {
1314
+ "name": "model.layers.2.self_attn.qkv_proj.q_weight",
1315
+ "shape": [
1316
+ 256,
1317
+ 2560
1318
+ ],
1319
+ "dtype": "uint32",
1320
+ "format": "f32-to-bf16",
1321
+ "nbytes": 2621440,
1322
+ "byteOffset": 28385280
1323
+ },
1324
+ {
1325
+ "name": "model.layers.2.self_attn.qkv_proj.q_scale",
1326
+ "shape": [
1327
+ 64,
1328
+ 2560
1329
+ ],
1330
+ "dtype": "float16",
1331
+ "format": "f32-to-bf16",
1332
+ "nbytes": 327680,
1333
+ "byteOffset": 31006720
1334
+ },
1335
+ {
1336
+ "name": "model.layers.2.self_attn.o_proj.q_weight",
1337
+ "shape": [
1338
+ 256,
1339
+ 2048
1340
+ ],
1341
+ "dtype": "uint32",
1342
+ "format": "f32-to-bf16",
1343
+ "nbytes": 2097152,
1344
+ "byteOffset": 31334400
1345
+ }
1346
+ ],
1347
+ "md5sum": "1e0d8ceaa3587d86245358080020e1bc"
1348
+ },
1349
+ {
1350
+ "dataPath": "params_shard_23.bin",
1351
+ "format": "raw-shard",
1352
+ "nbytes": 33554432,
1353
+ "records": [
1354
+ {
1355
+ "name": "model.layers.3.mlp.gate_up_proj.q_weight",
1356
+ "shape": [
1357
+ 256,
1358
+ 32768
1359
+ ],
1360
+ "dtype": "uint32",
1361
+ "format": "f32-to-bf16",
1362
+ "nbytes": 33554432,
1363
+ "byteOffset": 0
1364
+ }
1365
+ ],
1366
+ "md5sum": "8d158b4a153d626ed4a86ec561a58ee0"
1367
+ },
1368
+ {
1369
+ "dataPath": "params_shard_24.bin",
1370
+ "format": "raw-shard",
1371
+ "nbytes": 16777216,
1372
+ "records": [
1373
+ {
1374
+ "name": "model.layers.4.mlp.down_proj.q_weight",
1375
+ "shape": [
1376
+ 2048,
1377
+ 2048
1378
+ ],
1379
+ "dtype": "uint32",
1380
+ "format": "f32-to-bf16",
1381
+ "nbytes": 16777216,
1382
+ "byteOffset": 0
1383
+ }
1384
+ ],
1385
+ "md5sum": "d4da995026501bf0a2378f038c590dc3"
1386
+ },
1387
+ {
1388
+ "dataPath": "params_shard_25.bin",
1389
+ "format": "raw-shard",
1390
+ "nbytes": 33554432,
1391
+ "records": [
1392
+ {
1393
+ "name": "model.layers.4.mlp.gate_up_proj.q_weight",
1394
+ "shape": [
1395
+ 256,
1396
+ 32768
1397
+ ],
1398
+ "dtype": "uint32",
1399
+ "format": "f32-to-bf16",
1400
+ "nbytes": 33554432,
1401
+ "byteOffset": 0
1402
+ }
1403
+ ],
1404
+ "md5sum": "e055e4cf3b03a9789a371e439c6943e2"
1405
+ },
1406
+ {
1407
+ "dataPath": "params_shard_26.bin",
1408
+ "format": "raw-shard",
1409
+ "nbytes": 30748672,
1410
+ "records": [
1411
+ {
1412
+ "name": "model.layers.2.self_attn.o_proj.q_scale",
1413
+ "shape": [
1414
+ 64,
1415
+ 2048
1416
+ ],
1417
+ "dtype": "float16",
1418
+ "format": "f32-to-bf16",
1419
+ "nbytes": 262144,
1420
+ "byteOffset": 0
1421
+ },
1422
+ {
1423
+ "name": "model.layers.3.input_layernorm.weight",
1424
+ "shape": [
1425
+ 2048
1426
+ ],
1427
+ "dtype": "float16",
1428
+ "format": "f32-to-bf16",
1429
+ "nbytes": 4096,
1430
+ "byteOffset": 262144
1431
+ },
1432
+ {
1433
+ "name": "model.layers.3.mlp.down_proj.q_weight",
1434
+ "shape": [
1435
+ 2048,
1436
+ 2048
1437
+ ],
1438
+ "dtype": "uint32",
1439
+ "format": "f32-to-bf16",
1440
+ "nbytes": 16777216,
1441
+ "byteOffset": 266240
1442
+ },
1443
+ {
1444
+ "name": "model.layers.3.mlp.down_proj.q_scale",
1445
+ "shape": [
1446
+ 512,
1447
+ 2048
1448
+ ],
1449
+ "dtype": "float16",
1450
+ "format": "f32-to-bf16",
1451
+ "nbytes": 2097152,
1452
+ "byteOffset": 17043456
1453
+ },
1454
+ {
1455
+ "name": "model.layers.3.mlp.gate_up_proj.q_scale",
1456
+ "shape": [
1457
+ 64,
1458
+ 32768
1459
+ ],
1460
+ "dtype": "float16",
1461
+ "format": "f32-to-bf16",
1462
+ "nbytes": 4194304,
1463
+ "byteOffset": 19140608
1464
+ },
1465
+ {
1466
+ "name": "model.layers.3.post_attention_layernorm.weight",
1467
+ "shape": [
1468
+ 2048
1469
+ ],
1470
+ "dtype": "float16",
1471
+ "format": "f32-to-bf16",
1472
+ "nbytes": 4096,
1473
+ "byteOffset": 23334912
1474
+ },
1475
+ {
1476
+ "name": "model.layers.3.self_attn.qkv_proj.q_weight",
1477
+ "shape": [
1478
+ 256,
1479
+ 2560
1480
+ ],
1481
+ "dtype": "uint32",
1482
+ "format": "f32-to-bf16",
1483
+ "nbytes": 2621440,
1484
+ "byteOffset": 23339008
1485
+ },
1486
+ {
1487
+ "name": "model.layers.3.self_attn.qkv_proj.q_scale",
1488
+ "shape": [
1489
+ 64,
1490
+ 2560
1491
+ ],
1492
+ "dtype": "float16",
1493
+ "format": "f32-to-bf16",
1494
+ "nbytes": 327680,
1495
+ "byteOffset": 25960448
1496
+ },
1497
+ {
1498
+ "name": "model.layers.3.self_attn.o_proj.q_weight",
1499
+ "shape": [
1500
+ 256,
1501
+ 2048
1502
+ ],
1503
+ "dtype": "uint32",
1504
+ "format": "f32-to-bf16",
1505
+ "nbytes": 2097152,
1506
+ "byteOffset": 26288128
1507
+ },
1508
+ {
1509
+ "name": "model.layers.3.self_attn.o_proj.q_scale",
1510
+ "shape": [
1511
+ 64,
1512
+ 2048
1513
+ ],
1514
+ "dtype": "float16",
1515
+ "format": "f32-to-bf16",
1516
+ "nbytes": 262144,
1517
+ "byteOffset": 28385280
1518
+ },
1519
+ {
1520
+ "name": "model.layers.4.input_layernorm.weight",
1521
+ "shape": [
1522
+ 2048
1523
+ ],
1524
+ "dtype": "float16",
1525
+ "format": "f32-to-bf16",
1526
+ "nbytes": 4096,
1527
+ "byteOffset": 28647424
1528
+ },
1529
+ {
1530
+ "name": "model.layers.4.mlp.down_proj.q_scale",
1531
+ "shape": [
1532
+ 512,
1533
+ 2048
1534
+ ],
1535
+ "dtype": "float16",
1536
+ "format": "f32-to-bf16",
1537
+ "nbytes": 2097152,
1538
+ "byteOffset": 28651520
1539
+ }
1540
+ ],
1541
+ "md5sum": "2c4d526eef59fc7ccdbe1f0a140c20f5"
1542
+ },
1543
+ {
1544
+ "dataPath": "params_shard_27.bin",
1545
+ "format": "raw-shard",
1546
+ "nbytes": 33554432,
1547
+ "records": [
1548
+ {
1549
+ "name": "model.layers.5.mlp.gate_up_proj.q_weight",
1550
+ "shape": [
1551
+ 256,
1552
+ 32768
1553
+ ],
1554
+ "dtype": "uint32",
1555
+ "format": "f32-to-bf16",
1556
+ "nbytes": 33554432,
1557
+ "byteOffset": 0
1558
+ }
1559
+ ],
1560
+ "md5sum": "d3026966f6bc4067e43fbd15eedd268b"
1561
+ },
1562
+ {
1563
+ "dataPath": "params_shard_28.bin",
1564
+ "format": "raw-shard",
1565
+ "nbytes": 32583680,
1566
+ "records": [
1567
+ {
1568
+ "name": "model.layers.4.mlp.gate_up_proj.q_scale",
1569
+ "shape": [
1570
+ 64,
1571
+ 32768
1572
+ ],
1573
+ "dtype": "float16",
1574
+ "format": "f32-to-bf16",
1575
+ "nbytes": 4194304,
1576
+ "byteOffset": 0
1577
+ },
1578
+ {
1579
+ "name": "model.layers.4.post_attention_layernorm.weight",
1580
+ "shape": [
1581
+ 2048
1582
+ ],
1583
+ "dtype": "float16",
1584
+ "format": "f32-to-bf16",
1585
+ "nbytes": 4096,
1586
+ "byteOffset": 4194304
1587
+ },
1588
+ {
1589
+ "name": "model.layers.4.self_attn.qkv_proj.q_weight",
1590
+ "shape": [
1591
+ 256,
1592
+ 2560
1593
+ ],
1594
+ "dtype": "uint32",
1595
+ "format": "f32-to-bf16",
1596
+ "nbytes": 2621440,
1597
+ "byteOffset": 4198400
1598
+ },
1599
+ {
1600
+ "name": "model.layers.4.self_attn.qkv_proj.q_scale",
1601
+ "shape": [
1602
+ 64,
1603
+ 2560
1604
+ ],
1605
+ "dtype": "float16",
1606
+ "format": "f32-to-bf16",
1607
+ "nbytes": 327680,
1608
+ "byteOffset": 6819840
1609
+ },
1610
+ {
1611
+ "name": "model.layers.4.self_attn.o_proj.q_weight",
1612
+ "shape": [
1613
+ 256,
1614
+ 2048
1615
+ ],
1616
+ "dtype": "uint32",
1617
+ "format": "f32-to-bf16",
1618
+ "nbytes": 2097152,
1619
+ "byteOffset": 7147520
1620
+ },
1621
+ {
1622
+ "name": "model.layers.4.self_attn.o_proj.q_scale",
1623
+ "shape": [
1624
+ 64,
1625
+ 2048
1626
+ ],
1627
+ "dtype": "float16",
1628
+ "format": "f32-to-bf16",
1629
+ "nbytes": 262144,
1630
+ "byteOffset": 9244672
1631
+ },
1632
+ {
1633
+ "name": "model.layers.5.input_layernorm.weight",
1634
+ "shape": [
1635
+ 2048
1636
+ ],
1637
+ "dtype": "float16",
1638
+ "format": "f32-to-bf16",
1639
+ "nbytes": 4096,
1640
+ "byteOffset": 9506816
1641
+ },
1642
+ {
1643
+ "name": "model.layers.5.mlp.down_proj.q_weight",
1644
+ "shape": [
1645
+ 2048,
1646
+ 2048
1647
+ ],
1648
+ "dtype": "uint32",
1649
+ "format": "f32-to-bf16",
1650
+ "nbytes": 16777216,
1651
+ "byteOffset": 9510912
1652
+ },
1653
+ {
1654
+ "name": "model.layers.5.mlp.down_proj.q_scale",
1655
+ "shape": [
1656
+ 512,
1657
+ 2048
1658
+ ],
1659
+ "dtype": "float16",
1660
+ "format": "f32-to-bf16",
1661
+ "nbytes": 2097152,
1662
+ "byteOffset": 26288128
1663
+ },
1664
+ {
1665
+ "name": "model.layers.5.mlp.gate_up_proj.q_scale",
1666
+ "shape": [
1667
+ 64,
1668
+ 32768
1669
+ ],
1670
+ "dtype": "float16",
1671
+ "format": "f32-to-bf16",
1672
+ "nbytes": 4194304,
1673
+ "byteOffset": 28385280
1674
+ },
1675
+ {
1676
+ "name": "model.layers.5.post_attention_layernorm.weight",
1677
+ "shape": [
1678
+ 2048
1679
+ ],
1680
+ "dtype": "float16",
1681
+ "format": "f32-to-bf16",
1682
+ "nbytes": 4096,
1683
+ "byteOffset": 32579584
1684
+ }
1685
+ ],
1686
+ "md5sum": "53e520cb7fa7abe8ee3bc5baef193f28"
1687
+ },
1688
+ {
1689
+ "dataPath": "params_shard_29.bin",
1690
+ "format": "raw-shard",
1691
+ "nbytes": 33554432,
1692
+ "records": [
1693
+ {
1694
+ "name": "model.layers.6.mlp.gate_up_proj.q_weight",
1695
+ "shape": [
1696
+ 256,
1697
+ 32768
1698
+ ],
1699
+ "dtype": "uint32",
1700
+ "format": "f32-to-bf16",
1701
+ "nbytes": 33554432,
1702
+ "byteOffset": 0
1703
+ }
1704
+ ],
1705
+ "md5sum": "eb9b82a97a37adf5d7a4a3c30dd90a3e"
1706
+ },
1707
+ {
1708
+ "dataPath": "params_shard_30.bin",
1709
+ "format": "raw-shard",
1710
+ "nbytes": 33431552,
1711
+ "records": [
1712
+ {
1713
+ "name": "model.layers.5.self_attn.qkv_proj.q_weight",
1714
+ "shape": [
1715
+ 256,
1716
+ 2560
1717
+ ],
1718
+ "dtype": "uint32",
1719
+ "format": "f32-to-bf16",
1720
+ "nbytes": 2621440,
1721
+ "byteOffset": 0
1722
+ },
1723
+ {
1724
+ "name": "model.layers.5.self_attn.qkv_proj.q_scale",
1725
+ "shape": [
1726
+ 64,
1727
+ 2560
1728
+ ],
1729
+ "dtype": "float16",
1730
+ "format": "f32-to-bf16",
1731
+ "nbytes": 327680,
1732
+ "byteOffset": 2621440
1733
+ },
1734
+ {
1735
+ "name": "model.layers.5.self_attn.o_proj.q_weight",
1736
+ "shape": [
1737
+ 256,
1738
+ 2048
1739
+ ],
1740
+ "dtype": "uint32",
1741
+ "format": "f32-to-bf16",
1742
+ "nbytes": 2097152,
1743
+ "byteOffset": 2949120
1744
+ },
1745
+ {
1746
+ "name": "model.layers.5.self_attn.o_proj.q_scale",
1747
+ "shape": [
1748
+ 64,
1749
+ 2048
1750
+ ],
1751
+ "dtype": "float16",
1752
+ "format": "f32-to-bf16",
1753
+ "nbytes": 262144,
1754
+ "byteOffset": 5046272
1755
+ },
1756
+ {
1757
+ "name": "model.layers.6.input_layernorm.weight",
1758
+ "shape": [
1759
+ 2048
1760
+ ],
1761
+ "dtype": "float16",
1762
+ "format": "f32-to-bf16",
1763
+ "nbytes": 4096,
1764
+ "byteOffset": 5308416
1765
+ },
1766
+ {
1767
+ "name": "model.layers.6.mlp.down_proj.q_weight",
1768
+ "shape": [
1769
+ 2048,
1770
+ 2048
1771
+ ],
1772
+ "dtype": "uint32",
1773
+ "format": "f32-to-bf16",
1774
+ "nbytes": 16777216,
1775
+ "byteOffset": 5312512
1776
+ },
1777
+ {
1778
+ "name": "model.layers.6.mlp.down_proj.q_scale",
1779
+ "shape": [
1780
+ 512,
1781
+ 2048
1782
+ ],
1783
+ "dtype": "float16",
1784
+ "format": "f32-to-bf16",
1785
+ "nbytes": 2097152,
1786
+ "byteOffset": 22089728
1787
+ },
1788
+ {
1789
+ "name": "model.layers.6.mlp.gate_up_proj.q_scale",
1790
+ "shape": [
1791
+ 64,
1792
+ 32768
1793
+ ],
1794
+ "dtype": "float16",
1795
+ "format": "f32-to-bf16",
1796
+ "nbytes": 4194304,
1797
+ "byteOffset": 24186880
1798
+ },
1799
+ {
1800
+ "name": "model.layers.6.post_attention_layernorm.weight",
1801
+ "shape": [
1802
+ 2048
1803
+ ],
1804
+ "dtype": "float16",
1805
+ "format": "f32-to-bf16",
1806
+ "nbytes": 4096,
1807
+ "byteOffset": 28381184
1808
+ },
1809
+ {
1810
+ "name": "model.layers.6.self_attn.qkv_proj.q_weight",
1811
+ "shape": [
1812
+ 256,
1813
+ 2560
1814
+ ],
1815
+ "dtype": "uint32",
1816
+ "format": "f32-to-bf16",
1817
+ "nbytes": 2621440,
1818
+ "byteOffset": 28385280
1819
+ },
1820
+ {
1821
+ "name": "model.layers.6.self_attn.qkv_proj.q_scale",
1822
+ "shape": [
1823
+ 64,
1824
+ 2560
1825
+ ],
1826
+ "dtype": "float16",
1827
+ "format": "f32-to-bf16",
1828
+ "nbytes": 327680,
1829
+ "byteOffset": 31006720
1830
+ },
1831
+ {
1832
+ "name": "model.layers.6.self_attn.o_proj.q_weight",
1833
+ "shape": [
1834
+ 256,
1835
+ 2048
1836
+ ],
1837
+ "dtype": "uint32",
1838
+ "format": "f32-to-bf16",
1839
+ "nbytes": 2097152,
1840
+ "byteOffset": 31334400
1841
+ }
1842
+ ],
1843
+ "md5sum": "0964c066b2dba3c9bd0c13fc81aa847d"
1844
+ },
1845
+ {
1846
+ "dataPath": "params_shard_31.bin",
1847
+ "format": "raw-shard",
1848
+ "nbytes": 33554432,
1849
+ "records": [
1850
+ {
1851
+ "name": "model.layers.7.mlp.gate_up_proj.q_weight",
1852
+ "shape": [
1853
+ 256,
1854
+ 32768
1855
+ ],
1856
+ "dtype": "uint32",
1857
+ "format": "f32-to-bf16",
1858
+ "nbytes": 33554432,
1859
+ "byteOffset": 0
1860
+ }
1861
+ ],
1862
+ "md5sum": "48bf6868723196a709f18404bf8e49ff"
1863
+ },
1864
+ {
1865
+ "dataPath": "params_shard_32.bin",
1866
+ "format": "raw-shard",
1867
+ "nbytes": 16777216,
1868
+ "records": [
1869
+ {
1870
+ "name": "model.layers.8.mlp.down_proj.q_weight",
1871
+ "shape": [
1872
+ 2048,
1873
+ 2048
1874
+ ],
1875
+ "dtype": "uint32",
1876
+ "format": "f32-to-bf16",
1877
+ "nbytes": 16777216,
1878
+ "byteOffset": 0
1879
+ }
1880
+ ],
1881
+ "md5sum": "5900b1181607a4004ef494a36b319f56"
1882
+ },
1883
+ {
1884
+ "dataPath": "params_shard_33.bin",
1885
+ "format": "raw-shard",
1886
+ "nbytes": 33554432,
1887
+ "records": [
1888
+ {
1889
+ "name": "model.layers.8.mlp.gate_up_proj.q_weight",
1890
+ "shape": [
1891
+ 256,
1892
+ 32768
1893
+ ],
1894
+ "dtype": "uint32",
1895
+ "format": "f32-to-bf16",
1896
+ "nbytes": 33554432,
1897
+ "byteOffset": 0
1898
+ }
1899
+ ],
1900
+ "md5sum": "6e7cd5014d8ee5a1ca5eb73a9b58666c"
1901
+ },
1902
+ {
1903
+ "dataPath": "params_shard_34.bin",
1904
+ "format": "raw-shard",
1905
+ "nbytes": 30748672,
1906
+ "records": [
1907
+ {
1908
+ "name": "model.layers.6.self_attn.o_proj.q_scale",
1909
+ "shape": [
1910
+ 64,
1911
+ 2048
1912
+ ],
1913
+ "dtype": "float16",
1914
+ "format": "f32-to-bf16",
1915
+ "nbytes": 262144,
1916
+ "byteOffset": 0
1917
+ },
1918
+ {
1919
+ "name": "model.layers.7.input_layernorm.weight",
1920
+ "shape": [
1921
+ 2048
1922
+ ],
1923
+ "dtype": "float16",
1924
+ "format": "f32-to-bf16",
1925
+ "nbytes": 4096,
1926
+ "byteOffset": 262144
1927
+ },
1928
+ {
1929
+ "name": "model.layers.7.mlp.down_proj.q_weight",
1930
+ "shape": [
1931
+ 2048,
1932
+ 2048
1933
+ ],
1934
+ "dtype": "uint32",
1935
+ "format": "f32-to-bf16",
1936
+ "nbytes": 16777216,
1937
+ "byteOffset": 266240
1938
+ },
1939
+ {
1940
+ "name": "model.layers.7.mlp.down_proj.q_scale",
1941
+ "shape": [
1942
+ 512,
1943
+ 2048
1944
+ ],
1945
+ "dtype": "float16",
1946
+ "format": "f32-to-bf16",
1947
+ "nbytes": 2097152,
1948
+ "byteOffset": 17043456
1949
+ },
1950
+ {
1951
+ "name": "model.layers.7.mlp.gate_up_proj.q_scale",
1952
+ "shape": [
1953
+ 64,
1954
+ 32768
1955
+ ],
1956
+ "dtype": "float16",
1957
+ "format": "f32-to-bf16",
1958
+ "nbytes": 4194304,
1959
+ "byteOffset": 19140608
1960
+ },
1961
+ {
1962
+ "name": "model.layers.7.post_attention_layernorm.weight",
1963
+ "shape": [
1964
+ 2048
1965
+ ],
1966
+ "dtype": "float16",
1967
+ "format": "f32-to-bf16",
1968
+ "nbytes": 4096,
1969
+ "byteOffset": 23334912
1970
+ },
1971
+ {
1972
+ "name": "model.layers.7.self_attn.qkv_proj.q_weight",
1973
+ "shape": [
1974
+ 256,
1975
+ 2560
1976
+ ],
1977
+ "dtype": "uint32",
1978
+ "format": "f32-to-bf16",
1979
+ "nbytes": 2621440,
1980
+ "byteOffset": 23339008
1981
+ },
1982
+ {
1983
+ "name": "model.layers.7.self_attn.qkv_proj.q_scale",
1984
+ "shape": [
1985
+ 64,
1986
+ 2560
1987
+ ],
1988
+ "dtype": "float16",
1989
+ "format": "f32-to-bf16",
1990
+ "nbytes": 327680,
1991
+ "byteOffset": 25960448
1992
+ },
1993
+ {
1994
+ "name": "model.layers.7.self_attn.o_proj.q_weight",
1995
+ "shape": [
1996
+ 256,
1997
+ 2048
1998
+ ],
1999
+ "dtype": "uint32",
2000
+ "format": "f32-to-bf16",
2001
+ "nbytes": 2097152,
2002
+ "byteOffset": 26288128
2003
+ },
2004
+ {
2005
+ "name": "model.layers.7.self_attn.o_proj.q_scale",
2006
+ "shape": [
2007
+ 64,
2008
+ 2048
2009
+ ],
2010
+ "dtype": "float16",
2011
+ "format": "f32-to-bf16",
2012
+ "nbytes": 262144,
2013
+ "byteOffset": 28385280
2014
+ },
2015
+ {
2016
+ "name": "model.layers.8.input_layernorm.weight",
2017
+ "shape": [
2018
+ 2048
2019
+ ],
2020
+ "dtype": "float16",
2021
+ "format": "f32-to-bf16",
2022
+ "nbytes": 4096,
2023
+ "byteOffset": 28647424
2024
+ },
2025
+ {
2026
+ "name": "model.layers.8.mlp.down_proj.q_scale",
2027
+ "shape": [
2028
+ 512,
2029
+ 2048
2030
+ ],
2031
+ "dtype": "float16",
2032
+ "format": "f32-to-bf16",
2033
+ "nbytes": 2097152,
2034
+ "byteOffset": 28651520
2035
+ }
2036
+ ],
2037
+ "md5sum": "a83ac151c42f554cea6c6bb94ba0f1c1"
2038
+ },
2039
+ {
2040
+ "dataPath": "params_shard_35.bin",
2041
+ "format": "raw-shard",
2042
+ "nbytes": 33554432,
2043
+ "records": [
2044
+ {
2045
+ "name": "model.layers.9.mlp.gate_up_proj.q_weight",
2046
+ "shape": [
2047
+ 256,
2048
+ 32768
2049
+ ],
2050
+ "dtype": "uint32",
2051
+ "format": "f32-to-bf16",
2052
+ "nbytes": 33554432,
2053
+ "byteOffset": 0
2054
+ }
2055
+ ],
2056
+ "md5sum": "451f725d76b25f93509e230024934709"
2057
+ },
2058
+ {
2059
+ "dataPath": "params_shard_36.bin",
2060
+ "format": "raw-shard",
2061
+ "nbytes": 32583680,
2062
+ "records": [
2063
+ {
2064
+ "name": "model.layers.8.mlp.gate_up_proj.q_scale",
2065
+ "shape": [
2066
+ 64,
2067
+ 32768
2068
+ ],
2069
+ "dtype": "float16",
2070
+ "format": "f32-to-bf16",
2071
+ "nbytes": 4194304,
2072
+ "byteOffset": 0
2073
+ },
2074
+ {
2075
+ "name": "model.layers.8.post_attention_layernorm.weight",
2076
+ "shape": [
2077
+ 2048
2078
+ ],
2079
+ "dtype": "float16",
2080
+ "format": "f32-to-bf16",
2081
+ "nbytes": 4096,
2082
+ "byteOffset": 4194304
2083
+ },
2084
+ {
2085
+ "name": "model.layers.8.self_attn.qkv_proj.q_weight",
2086
+ "shape": [
2087
+ 256,
2088
+ 2560
2089
+ ],
2090
+ "dtype": "uint32",
2091
+ "format": "f32-to-bf16",
2092
+ "nbytes": 2621440,
2093
+ "byteOffset": 4198400
2094
+ },
2095
+ {
2096
+ "name": "model.layers.8.self_attn.qkv_proj.q_scale",
2097
+ "shape": [
2098
+ 64,
2099
+ 2560
2100
+ ],
2101
+ "dtype": "float16",
2102
+ "format": "f32-to-bf16",
2103
+ "nbytes": 327680,
2104
+ "byteOffset": 6819840
2105
+ },
2106
+ {
2107
+ "name": "model.layers.8.self_attn.o_proj.q_weight",
2108
+ "shape": [
2109
+ 256,
2110
+ 2048
2111
+ ],
2112
+ "dtype": "uint32",
2113
+ "format": "f32-to-bf16",
2114
+ "nbytes": 2097152,
2115
+ "byteOffset": 7147520
2116
+ },
2117
+ {
2118
+ "name": "model.layers.8.self_attn.o_proj.q_scale",
2119
+ "shape": [
2120
+ 64,
2121
+ 2048
2122
+ ],
2123
+ "dtype": "float16",
2124
+ "format": "f32-to-bf16",
2125
+ "nbytes": 262144,
2126
+ "byteOffset": 9244672
2127
+ },
2128
+ {
2129
+ "name": "model.layers.9.input_layernorm.weight",
2130
+ "shape": [
2131
+ 2048
2132
+ ],
2133
+ "dtype": "float16",
2134
+ "format": "f32-to-bf16",
2135
+ "nbytes": 4096,
2136
+ "byteOffset": 9506816
2137
+ },
2138
+ {
2139
+ "name": "model.layers.9.mlp.down_proj.q_weight",
2140
+ "shape": [
2141
+ 2048,
2142
+ 2048
2143
+ ],
2144
+ "dtype": "uint32",
2145
+ "format": "f32-to-bf16",
2146
+ "nbytes": 16777216,
2147
+ "byteOffset": 9510912
2148
+ },
2149
+ {
2150
+ "name": "model.layers.9.mlp.down_proj.q_scale",
2151
+ "shape": [
2152
+ 512,
2153
+ 2048
2154
+ ],
2155
+ "dtype": "float16",
2156
+ "format": "f32-to-bf16",
2157
+ "nbytes": 2097152,
2158
+ "byteOffset": 26288128
2159
+ },
2160
+ {
2161
+ "name": "model.layers.9.mlp.gate_up_proj.q_scale",
2162
+ "shape": [
2163
+ 64,
2164
+ 32768
2165
+ ],
2166
+ "dtype": "float16",
2167
+ "format": "f32-to-bf16",
2168
+ "nbytes": 4194304,
2169
+ "byteOffset": 28385280
2170
+ },
2171
+ {
2172
+ "name": "model.layers.9.post_attention_layernorm.weight",
2173
+ "shape": [
2174
+ 2048
2175
+ ],
2176
+ "dtype": "float16",
2177
+ "format": "f32-to-bf16",
2178
+ "nbytes": 4096,
2179
+ "byteOffset": 32579584
2180
+ }
2181
+ ],
2182
+ "md5sum": "d2405822bb1ce59620572f678f0ed8fc"
2183
+ },
2184
+ {
2185
+ "dataPath": "params_shard_37.bin",
2186
+ "format": "raw-shard",
2187
+ "nbytes": 24195072,
2188
+ "records": [
2189
+ {
2190
+ "name": "model.layers.9.self_attn.qkv_proj.q_weight",
2191
+ "shape": [
2192
+ 256,
2193
+ 2560
2194
+ ],
2195
+ "dtype": "uint32",
2196
+ "format": "f32-to-bf16",
2197
+ "nbytes": 2621440,
2198
+ "byteOffset": 0
2199
+ },
2200
+ {
2201
+ "name": "model.layers.9.self_attn.qkv_proj.q_scale",
2202
+ "shape": [
2203
+ 64,
2204
+ 2560
2205
+ ],
2206
+ "dtype": "float16",
2207
+ "format": "f32-to-bf16",
2208
+ "nbytes": 327680,
2209
+ "byteOffset": 2621440
2210
+ },
2211
+ {
2212
+ "name": "model.layers.9.self_attn.o_proj.q_weight",
2213
+ "shape": [
2214
+ 256,
2215
+ 2048
2216
+ ],
2217
+ "dtype": "uint32",
2218
+ "format": "f32-to-bf16",
2219
+ "nbytes": 2097152,
2220
+ "byteOffset": 2949120
2221
+ },
2222
+ {
2223
+ "name": "model.layers.9.self_attn.o_proj.q_scale",
2224
+ "shape": [
2225
+ 64,
2226
+ 2048
2227
+ ],
2228
+ "dtype": "float16",
2229
+ "format": "f32-to-bf16",
2230
+ "nbytes": 262144,
2231
+ "byteOffset": 5046272
2232
+ },
2233
+ {
2234
+ "name": "model.layers.17.input_layernorm.weight",
2235
+ "shape": [
2236
+ 2048
2237
+ ],
2238
+ "dtype": "float16",
2239
+ "format": "f32-to-bf16",
2240
+ "nbytes": 4096,
2241
+ "byteOffset": 5308416
2242
+ },
2243
+ {
2244
+ "name": "model.layers.17.mlp.down_proj.q_weight",
2245
+ "shape": [
2246
+ 2048,
2247
+ 2048
2248
+ ],
2249
+ "dtype": "uint32",
2250
+ "format": "f32-to-bf16",
2251
+ "nbytes": 16777216,
2252
+ "byteOffset": 5312512
2253
+ },
2254
+ {
2255
+ "name": "model.layers.17.mlp.down_proj.q_scale",
2256
+ "shape": [
2257
+ 512,
2258
+ 2048
2259
+ ],
2260
+ "dtype": "float16",
2261
+ "format": "f32-to-bf16",
2262
+ "nbytes": 2097152,
2263
+ "byteOffset": 22089728
2264
+ },
2265
+ {
2266
+ "name": "model.layers.17.post_attention_layernorm.weight",
2267
+ "shape": [
2268
+ 2048
2269
+ ],
2270
+ "dtype": "float16",
2271
+ "format": "f32-to-bf16",
2272
+ "nbytes": 4096,
2273
+ "byteOffset": 24186880
2274
+ },
2275
+ {
2276
+ "name": "model.norm.weight",
2277
+ "shape": [
2278
+ 2048
2279
+ ],
2280
+ "dtype": "float16",
2281
+ "format": "f32-to-bf16",
2282
+ "nbytes": 4096,
2283
+ "byteOffset": 24190976
2284
+ }
2285
+ ],
2286
+ "md5sum": "bdd5652139689bc22a90a07b6387219f"
2287
+ }
2288
+ ]
2289
+ }
params_shard_0.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8937c008e82cdeba90a21bf01b3780156f2145a67838864f2466a9fec95ae76f
3
+ size 262144000
params_shard_1.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a9ccd87d2b83c0ce697c4f7bba8286d2db9f7fe9b1864a9e6165db5cdc330b92
3
+ size 16777216
params_shard_10.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:21f46fe5098405b57f126e5a457aaaacd23af97877631821655276c36fa8a292
3
+ size 33554432
params_shard_11.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ecf0750294defb12c2c3d4025d6f71daab0829fb39c1932dfff785f2a5bdf004
3
+ size 16777216
params_shard_12.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:02632b368989df2d4c145d8b692ce109161c45b7c9c17ae6b7499166be6a02e1
3
+ size 33554432
params_shard_13.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5fbe9e69859e5dd75fa3d975aa38651a362a74a4de5ff1644b6ba7f935598aea
3
+ size 30748672
params_shard_14.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5146c60d4304ed275185c1d4e418d22d4aaabbb1fe7756a1273f87c23a7b5c90
3
+ size 33554432
params_shard_15.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:41782249cfdf8386d94a7fad0e9c467d37e0adf80e1c5ce9ff62d6e200d090bb
3
+ size 32583680
params_shard_16.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a71fca298b1076e00a26c0b85ea0027f3f6165dd65235b59eb2d65582067a348
3
+ size 33554432
params_shard_17.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b887621f38228224ac4a6d7dcdd0938b8fd405975824f6f7af6189e732117f36
3
+ size 33431552
params_shard_18.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:62bf85737b3b7e88a44836ce82d675fbc2e5e5d7bf7f4b8de6d88d279630fe9a
3
+ size 33554432
params_shard_19.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5821eaf7f6942f6d2288d751cbf4e8194e9ff3333ac5dc9f7c97a8741e263cc1
3
+ size 33554432
params_shard_2.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:83b4f244d8f401048227888caf493195cc9e20e33742df2583e8f87329100a62
3
+ size 32772096
params_shard_20.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6a7059e760a30507a3d1764248b45ca8ab5ce774d663f849e21905584df409be
3
+ size 32841728
params_shard_21.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9a4621f75ad6cfc436ac95f2a3eaccc6614b6c08a1a3b129f1c7ff8984a58d80
3
+ size 33554432
params_shard_22.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fe3571496eed93133bb2b187f539cea20c3ca14f004e175f0ac380284d9cb33a
3
+ size 33431552
params_shard_23.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8c573b306d39657e8987d84716dcd6a6c8804addab3ef7e5aeb9eb879d6fe357
3
+ size 33554432
params_shard_24.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ec41371da19f961bb1e9f1e7732e2fa0ed1cf8e1c6f5c8aa460b50ed0c013bb3
3
+ size 16777216
params_shard_25.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6de9d2a6fe3017cb73927b5b35b1fd3820df20cfef8726e5afcf6e24bcac735a
3
+ size 33554432
params_shard_26.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f52e7d9a44b1394e4368cfc746701179e4721b19e8cbb468fbda225e1e6aacf9
3
+ size 30748672
params_shard_27.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:90fc7efab00cbd73ff6a52d1315128da6dcc3330a223ef7c69944b96d2c52c34
3
+ size 33554432
params_shard_28.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0676ddfa6ac93f343ce1232cac163b4f23dc62d7852b5264db75fdd1696148d0
3
+ size 32583680
params_shard_29.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:79c0b4bf314f4c8485785e2b9b92a381fcfe9a30e533a3f1dd77d410453a0d0b
3
+ size 33554432
params_shard_3.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b88371548a14a3dcba54e4de638638ea4588c27e7ed8310c25f5df5e8e092a32
3
+ size 33554432
params_shard_30.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b8571c1571c31e17d5310977f906d8753c45f555577a60be76f46be462e6be31
3
+ size 33431552
params_shard_31.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d8760a557b68807413a78e84efa5bf53204af82063b3913a50ab97512fa81b5d
3
+ size 33554432
params_shard_32.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:be30bb701cab0f043fe126a439b3f380d701bb1c36a8b78421fabf2b1fab5195
3
+ size 16777216
params_shard_33.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:89876eb1efa8970577e77f6fb520e8716ac3cf01e33647c03f3d456adf4457ed
3
+ size 33554432
params_shard_34.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:75008d03fa365abed351b80bb7ccb740cafa8d281c07312df40d513c23c7751a
3
+ size 30748672
params_shard_35.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:587937630e2cd13d46da6fe7206fb4e962dee3407f8e83cd029100432c377e82
3
+ size 33554432
params_shard_36.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5234151fb3f5d29ccbf3e4124fecf744c836abba226120dd707d677b33103b0b
3
+ size 32583680
params_shard_37.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:98de598b49476a42f40ba9436b8a9287918a3c565049fcde640e83e1218090f4
3
+ size 24195072
params_shard_4.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5ebed935463a659d2d5b94581c5d2d62954d50aab498bc88a1eed04dd8cc6db7
3
+ size 33554432
params_shard_5.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:388a897a0a26389c7b22b1237672cbf0a983cb3bbd3d396a61483c03fcc908ba
3
+ size 30482432
params_shard_6.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c7a4bad3c6940299d4ecd3abddd5610c5468bfc42eadbee5ae9f8885f45f00f2
3
+ size 33554432
params_shard_7.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6af103f83a9022dba39bea7625200f6a5719c3ce77d9615b4931fa5ca664cc93
3
+ size 32583680
params_shard_8.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8bb44970573979ce270eb221cbd2bc18405e5473ef15687d78e26bc67be67533
3
+ size 33554432
params_shard_9.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d0aebfcd6523af90572f70745928e6e1c54ff747f295db034c76ff00401a3e0d
3
+ size 33431552
tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:05e97791a5e007260de1db7e1692e53150e08cea481e2bf25435553380c147ee
3
+ size 17477929
tokenizer.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:61a7b147390c64585d6c3543dd6fc636906c9af3865a5548f27f31aee1d4c8e2
3
+ size 4241003
tokenizer_config.json ADDED
@@ -0,0 +1,70 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_bos_token": true,
3
+ "add_eos_token": false,
4
+ "added_tokens_decoder": {
5
+ "0": {
6
+ "content": "<pad>",
7
+ "lstrip": false,
8
+ "normalized": false,
9
+ "rstrip": false,
10
+ "single_word": false,
11
+ "special": true
12
+ },
13
+ "1": {
14
+ "content": "<eos>",
15
+ "lstrip": false,
16
+ "normalized": false,
17
+ "rstrip": false,
18
+ "single_word": false,
19
+ "special": true
20
+ },
21
+ "2": {
22
+ "content": "<bos>",
23
+ "lstrip": false,
24
+ "normalized": false,
25
+ "rstrip": false,
26
+ "single_word": false,
27
+ "special": true
28
+ },
29
+ "3": {
30
+ "content": "<unk>",
31
+ "lstrip": false,
32
+ "normalized": false,
33
+ "rstrip": false,
34
+ "single_word": false,
35
+ "special": true
36
+ },
37
+ "106": {
38
+ "content": "<start_of_turn>",
39
+ "lstrip": false,
40
+ "normalized": false,
41
+ "rstrip": false,
42
+ "single_word": false,
43
+ "special": true
44
+ },
45
+ "107": {
46
+ "content": "<end_of_turn>",
47
+ "lstrip": false,
48
+ "normalized": false,
49
+ "rstrip": false,
50
+ "single_word": false,
51
+ "special": true
52
+ }
53
+ },
54
+ "additional_special_tokens": [
55
+ "<start_of_turn>",
56
+ "<end_of_turn>"
57
+ ],
58
+ "bos_token": "<bos>",
59
+ "chat_template": "{{ bos_token }}{% if messages[0]['role'] == 'system' %}{{ raise_exception('System role not supported') }}{% endif %}{% for message in messages %}{% if (message['role'] == 'user') != (loop.index0 % 2 == 0) %}{{ raise_exception('Conversation roles must alternate user/assistant/user/assistant/...') }}{% endif %}{% if (message['role'] == 'assistant') %}{% set role = 'model' %}{% else %}{% set role = message['role'] %}{% endif %}{{ '<start_of_turn>' + role + '\n' + message['content'] | trim + '<end_of_turn>\n' }}{% endfor %}{% if add_generation_prompt %}{{'<start_of_turn>model\n'}}{% endif %}",
60
+ "clean_up_tokenization_spaces": false,
61
+ "eos_token": "<eos>",
62
+ "legacy": null,
63
+ "model_max_length": 1000000000000000019884624838656,
64
+ "pad_token": "<pad>",
65
+ "sp_model_kwargs": {},
66
+ "spaces_between_special_tokens": false,
67
+ "tokenizer_class": "GemmaTokenizer",
68
+ "unk_token": "<unk>",
69
+ "use_default_system_prompt": false
70
+ }