Hugging Face's logo

NikFooBar
/
TinyLlama-1_1B-Chat-2-q4f16_1

Model card Files Community
Nickinfinity commited on
Commit
0bb0c9d
•
1 Parent(s): aba5c57

Init

Browse files

Signed-off-by: Nickinfinity <nikhilg.me@gmail.com>

Files changed (29) hide show
  1. mlc-chat-config.json +38 -0
  2. ndarray-cache.json +2647 -0
  3. params_shard_0.bin +3 -0
  4. params_shard_1.bin +3 -0
  5. params_shard_10.bin +3 -0
  6. params_shard_11.bin +3 -0
  7. params_shard_12.bin +3 -0
  8. params_shard_13.bin +3 -0
  9. params_shard_14.bin +3 -0
  10. params_shard_15.bin +3 -0
  11. params_shard_16.bin +3 -0
  12. params_shard_17.bin +3 -0
  13. params_shard_18.bin +3 -0
  14. params_shard_19.bin +3 -0
  15. params_shard_2.bin +3 -0
  16. params_shard_20.bin +3 -0
  17. params_shard_21.bin +3 -0
  18. params_shard_22.bin +3 -0
  19. params_shard_23.bin +3 -0
  20. params_shard_24.bin +3 -0
  21. params_shard_25.bin +3 -0
  22. params_shard_3.bin +3 -0
  23. params_shard_4.bin +3 -0
  24. params_shard_5.bin +3 -0
  25. params_shard_6.bin +3 -0
  26. params_shard_7.bin +3 -0
  27. params_shard_8.bin +3 -0
  28. params_shard_9.bin +3 -0
  29. tokenizer.json +0 -0
mlc-chat-config.json ADDED
@@ -0,0 +1,38 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "model_type": "llama",
3
+ "quantization": "q4f16_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": 160984,
11
+ "position_embedding_base": 10000.0,
12
+ "context_window_size": 768,
13
+ "prefill_chunk_size": 768,
14
+ "num_key_value_heads": 4,
15
+ "head_dim": 64,
16
+ "tensor_parallel_shards": 1
17
+ },
18
+ "vocab_size": 160984,
19
+ "context_window_size": 768,
20
+ "sliding_window_size": -1,
21
+ "prefill_chunk_size": 768,
22
+ "attention_sink_size": -1,
23
+ "tensor_parallel_shards": 1,
24
+ "mean_gen_len": 128,
25
+ "max_gen_len": 512,
26
+ "shift_fill_factor": 0.3,
27
+ "temperature": 0.7,
28
+ "repetition_penalty": 1.0,
29
+ "top_p": 0.95,
30
+ "conv_template": "llama-2",
31
+ "pad_token_id": 0,
32
+ "bos_token_id": 1,
33
+ "eos_token_id": 2,
34
+ "tokenizer_files": [
35
+ "tokenizer.json"
36
+ ],
37
+ "version": "0.1.0"
38
+ }
ndarray-cache.json ADDED
@@ -0,0 +1,2647 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metadata": {
3
+ "ParamSize": 225,
4
+ "ParamBytes": 916088832.0,
5
+ "BitsPerParam": 4.500650860708262
6
+ },
7
+ "records": [
8
+ {
9
+ "dataPath": "params_shard_0.bin",
10
+ "format": "raw-shard",
11
+ "nbytes": 164847616,
12
+ "records": [
13
+ {
14
+ "name": "lm_head.q_weight",
15
+ "shape": [
16
+ 160984,
17
+ 256
18
+ ],
19
+ "dtype": "uint32",
20
+ "format": "f32-to-bf16",
21
+ "nbytes": 164847616,
22
+ "byteOffset": 0
23
+ }
24
+ ],
25
+ "md5sum": "66fd10d562b3377b621a1b4bacfbe607"
26
+ },
27
+ {
28
+ "dataPath": "params_shard_1.bin",
29
+ "format": "raw-shard",
30
+ "nbytes": 164847616,
31
+ "records": [
32
+ {
33
+ "name": "model.embed_tokens.q_weight",
34
+ "shape": [
35
+ 160984,
36
+ 256
37
+ ],
38
+ "dtype": "uint32",
39
+ "format": "f32-to-bf16",
40
+ "nbytes": 164847616,
41
+ "byteOffset": 0
42
+ }
43
+ ],
44
+ "md5sum": "83824124dc0f0dd818dcdb63c547b911"
45
+ },
46
+ {
47
+ "dataPath": "params_shard_2.bin",
48
+ "format": "raw-shard",
49
+ "nbytes": 20605952,
50
+ "records": [
51
+ {
52
+ "name": "model.embed_tokens.q_scale",
53
+ "shape": [
54
+ 160984,
55
+ 64
56
+ ],
57
+ "dtype": "float16",
58
+ "format": "f32-to-bf16",
59
+ "nbytes": 20605952,
60
+ "byteOffset": 0
61
+ }
62
+ ],
63
+ "md5sum": "a8d163e70ba7ad6d6bd638dd32f28233"
64
+ },
65
+ {
66
+ "dataPath": "params_shard_3.bin",
67
+ "format": "raw-shard",
68
+ "nbytes": 27098112,
69
+ "records": [
70
+ {
71
+ "name": "lm_head.q_scale",
72
+ "shape": [
73
+ 160984,
74
+ 64
75
+ ],
76
+ "dtype": "float16",
77
+ "format": "f32-to-bf16",
78
+ "nbytes": 20605952,
79
+ "byteOffset": 0
80
+ },
81
+ {
82
+ "name": "model.layers.0.input_layernorm.weight",
83
+ "shape": [
84
+ 2048
85
+ ],
86
+ "dtype": "float16",
87
+ "format": "f32-to-bf16",
88
+ "nbytes": 4096,
89
+ "byteOffset": 20605952
90
+ },
91
+ {
92
+ "name": "model.layers.0.mlp.down_proj.q_weight",
93
+ "shape": [
94
+ 2048,
95
+ 704
96
+ ],
97
+ "dtype": "uint32",
98
+ "format": "f32-to-bf16",
99
+ "nbytes": 5767168,
100
+ "byteOffset": 20610048
101
+ },
102
+ {
103
+ "name": "model.layers.0.mlp.down_proj.q_scale",
104
+ "shape": [
105
+ 2048,
106
+ 176
107
+ ],
108
+ "dtype": "float16",
109
+ "format": "f32-to-bf16",
110
+ "nbytes": 720896,
111
+ "byteOffset": 26377216
112
+ }
113
+ ],
114
+ "md5sum": "638a28a548dee7e9430d91d06e2a62dc"
115
+ },
116
+ {
117
+ "dataPath": "params_shard_4.bin",
118
+ "format": "raw-shard",
119
+ "nbytes": 24780800,
120
+ "records": [
121
+ {
122
+ "name": "model.layers.0.mlp.gate_up_proj.q_weight",
123
+ "shape": [
124
+ 11264,
125
+ 256
126
+ ],
127
+ "dtype": "uint32",
128
+ "format": "f32-to-bf16",
129
+ "nbytes": 11534336,
130
+ "byteOffset": 0
131
+ },
132
+ {
133
+ "name": "model.layers.0.mlp.gate_up_proj.q_scale",
134
+ "shape": [
135
+ 11264,
136
+ 64
137
+ ],
138
+ "dtype": "float16",
139
+ "format": "f32-to-bf16",
140
+ "nbytes": 1441792,
141
+ "byteOffset": 11534336
142
+ },
143
+ {
144
+ "name": "model.layers.0.post_attention_layernorm.weight",
145
+ "shape": [
146
+ 2048
147
+ ],
148
+ "dtype": "float16",
149
+ "format": "f32-to-bf16",
150
+ "nbytes": 4096,
151
+ "byteOffset": 12976128
152
+ },
153
+ {
154
+ "name": "model.layers.0.self_attn.qkv_proj.q_weight",
155
+ "shape": [
156
+ 2560,
157
+ 256
158
+ ],
159
+ "dtype": "uint32",
160
+ "format": "f32-to-bf16",
161
+ "nbytes": 2621440,
162
+ "byteOffset": 12980224
163
+ },
164
+ {
165
+ "name": "model.layers.0.self_attn.qkv_proj.q_scale",
166
+ "shape": [
167
+ 2560,
168
+ 64
169
+ ],
170
+ "dtype": "float16",
171
+ "format": "f32-to-bf16",
172
+ "nbytes": 327680,
173
+ "byteOffset": 15601664
174
+ },
175
+ {
176
+ "name": "model.layers.0.self_attn.o_proj.q_weight",
177
+ "shape": [
178
+ 2048,
179
+ 256
180
+ ],
181
+ "dtype": "uint32",
182
+ "format": "f32-to-bf16",
183
+ "nbytes": 2097152,
184
+ "byteOffset": 15929344
185
+ },
186
+ {
187
+ "name": "model.layers.0.self_attn.o_proj.q_scale",
188
+ "shape": [
189
+ 2048,
190
+ 64
191
+ ],
192
+ "dtype": "float16",
193
+ "format": "f32-to-bf16",
194
+ "nbytes": 262144,
195
+ "byteOffset": 18026496
196
+ },
197
+ {
198
+ "name": "model.layers.1.input_layernorm.weight",
199
+ "shape": [
200
+ 2048
201
+ ],
202
+ "dtype": "float16",
203
+ "format": "f32-to-bf16",
204
+ "nbytes": 4096,
205
+ "byteOffset": 18288640
206
+ },
207
+ {
208
+ "name": "model.layers.1.mlp.down_proj.q_weight",
209
+ "shape": [
210
+ 2048,
211
+ 704
212
+ ],
213
+ "dtype": "uint32",
214
+ "format": "f32-to-bf16",
215
+ "nbytes": 5767168,
216
+ "byteOffset": 18292736
217
+ },
218
+ {
219
+ "name": "model.layers.1.mlp.down_proj.q_scale",
220
+ "shape": [
221
+ 2048,
222
+ 176
223
+ ],
224
+ "dtype": "float16",
225
+ "format": "f32-to-bf16",
226
+ "nbytes": 720896,
227
+ "byteOffset": 24059904
228
+ }
229
+ ],
230
+ "md5sum": "3cf853e6f586b52e2c27b5a760a94bbb"
231
+ },
232
+ {
233
+ "dataPath": "params_shard_5.bin",
234
+ "format": "raw-shard",
235
+ "nbytes": 24780800,
236
+ "records": [
237
+ {
238
+ "name": "model.layers.1.mlp.gate_up_proj.q_weight",
239
+ "shape": [
240
+ 11264,
241
+ 256
242
+ ],
243
+ "dtype": "uint32",
244
+ "format": "f32-to-bf16",
245
+ "nbytes": 11534336,
246
+ "byteOffset": 0
247
+ },
248
+ {
249
+ "name": "model.layers.1.mlp.gate_up_proj.q_scale",
250
+ "shape": [
251
+ 11264,
252
+ 64
253
+ ],
254
+ "dtype": "float16",
255
+ "format": "f32-to-bf16",
256
+ "nbytes": 1441792,
257
+ "byteOffset": 11534336
258
+ },
259
+ {
260
+ "name": "model.layers.1.post_attention_layernorm.weight",
261
+ "shape": [
262
+ 2048
263
+ ],
264
+ "dtype": "float16",
265
+ "format": "f32-to-bf16",
266
+ "nbytes": 4096,
267
+ "byteOffset": 12976128
268
+ },
269
+ {
270
+ "name": "model.layers.1.self_attn.qkv_proj.q_weight",
271
+ "shape": [
272
+ 2560,
273
+ 256
274
+ ],
275
+ "dtype": "uint32",
276
+ "format": "f32-to-bf16",
277
+ "nbytes": 2621440,
278
+ "byteOffset": 12980224
279
+ },
280
+ {
281
+ "name": "model.layers.1.self_attn.qkv_proj.q_scale",
282
+ "shape": [
283
+ 2560,
284
+ 64
285
+ ],
286
+ "dtype": "float16",
287
+ "format": "f32-to-bf16",
288
+ "nbytes": 327680,
289
+ "byteOffset": 15601664
290
+ },
291
+ {
292
+ "name": "model.layers.1.self_attn.o_proj.q_weight",
293
+ "shape": [
294
+ 2048,
295
+ 256
296
+ ],
297
+ "dtype": "uint32",
298
+ "format": "f32-to-bf16",
299
+ "nbytes": 2097152,
300
+ "byteOffset": 15929344
301
+ },
302
+ {
303
+ "name": "model.layers.1.self_attn.o_proj.q_scale",
304
+ "shape": [
305
+ 2048,
306
+ 64
307
+ ],
308
+ "dtype": "float16",
309
+ "format": "f32-to-bf16",
310
+ "nbytes": 262144,
311
+ "byteOffset": 18026496
312
+ },
313
+ {
314
+ "name": "model.layers.10.input_layernorm.weight",
315
+ "shape": [
316
+ 2048
317
+ ],
318
+ "dtype": "float16",
319
+ "format": "f32-to-bf16",
320
+ "nbytes": 4096,
321
+ "byteOffset": 18288640
322
+ },
323
+ {
324
+ "name": "model.layers.10.mlp.down_proj.q_weight",
325
+ "shape": [
326
+ 2048,
327
+ 704
328
+ ],
329
+ "dtype": "uint32",
330
+ "format": "f32-to-bf16",
331
+ "nbytes": 5767168,
332
+ "byteOffset": 18292736
333
+ },
334
+ {
335
+ "name": "model.layers.10.mlp.down_proj.q_scale",
336
+ "shape": [
337
+ 2048,
338
+ 176
339
+ ],
340
+ "dtype": "float16",
341
+ "format": "f32-to-bf16",
342
+ "nbytes": 720896,
343
+ "byteOffset": 24059904
344
+ }
345
+ ],
346
+ "md5sum": "28b16f3a2f3b53d1352916220db52388"
347
+ },
348
+ {
349
+ "dataPath": "params_shard_6.bin",
350
+ "format": "raw-shard",
351
+ "nbytes": 24780800,
352
+ "records": [
353
+ {
354
+ "name": "model.layers.10.mlp.gate_up_proj.q_weight",
355
+ "shape": [
356
+ 11264,
357
+ 256
358
+ ],
359
+ "dtype": "uint32",
360
+ "format": "f32-to-bf16",
361
+ "nbytes": 11534336,
362
+ "byteOffset": 0
363
+ },
364
+ {
365
+ "name": "model.layers.10.mlp.gate_up_proj.q_scale",
366
+ "shape": [
367
+ 11264,
368
+ 64
369
+ ],
370
+ "dtype": "float16",
371
+ "format": "f32-to-bf16",
372
+ "nbytes": 1441792,
373
+ "byteOffset": 11534336
374
+ },
375
+ {
376
+ "name": "model.layers.10.post_attention_layernorm.weight",
377
+ "shape": [
378
+ 2048
379
+ ],
380
+ "dtype": "float16",
381
+ "format": "f32-to-bf16",
382
+ "nbytes": 4096,
383
+ "byteOffset": 12976128
384
+ },
385
+ {
386
+ "name": "model.layers.10.self_attn.qkv_proj.q_weight",
387
+ "shape": [
388
+ 2560,
389
+ 256
390
+ ],
391
+ "dtype": "uint32",
392
+ "format": "f32-to-bf16",
393
+ "nbytes": 2621440,
394
+ "byteOffset": 12980224
395
+ },
396
+ {
397
+ "name": "model.layers.10.self_attn.qkv_proj.q_scale",
398
+ "shape": [
399
+ 2560,
400
+ 64
401
+ ],
402
+ "dtype": "float16",
403
+ "format": "f32-to-bf16",
404
+ "nbytes": 327680,
405
+ "byteOffset": 15601664
406
+ },
407
+ {
408
+ "name": "model.layers.10.self_attn.o_proj.q_weight",
409
+ "shape": [
410
+ 2048,
411
+ 256
412
+ ],
413
+ "dtype": "uint32",
414
+ "format": "f32-to-bf16",
415
+ "nbytes": 2097152,
416
+ "byteOffset": 15929344
417
+ },
418
+ {
419
+ "name": "model.layers.10.self_attn.o_proj.q_scale",
420
+ "shape": [
421
+ 2048,
422
+ 64
423
+ ],
424
+ "dtype": "float16",
425
+ "format": "f32-to-bf16",
426
+ "nbytes": 262144,
427
+ "byteOffset": 18026496
428
+ },
429
+ {
430
+ "name": "model.layers.11.input_layernorm.weight",
431
+ "shape": [
432
+ 2048
433
+ ],
434
+ "dtype": "float16",
435
+ "format": "f32-to-bf16",
436
+ "nbytes": 4096,
437
+ "byteOffset": 18288640
438
+ },
439
+ {
440
+ "name": "model.layers.11.mlp.down_proj.q_weight",
441
+ "shape": [
442
+ 2048,
443
+ 704
444
+ ],
445
+ "dtype": "uint32",
446
+ "format": "f32-to-bf16",
447
+ "nbytes": 5767168,
448
+ "byteOffset": 18292736
449
+ },
450
+ {
451
+ "name": "model.layers.11.mlp.down_proj.q_scale",
452
+ "shape": [
453
+ 2048,
454
+ 176
455
+ ],
456
+ "dtype": "float16",
457
+ "format": "f32-to-bf16",
458
+ "nbytes": 720896,
459
+ "byteOffset": 24059904
460
+ }
461
+ ],
462
+ "md5sum": "73fea510e5ce8153c99850aba8ea2dca"
463
+ },
464
+ {
465
+ "dataPath": "params_shard_7.bin",
466
+ "format": "raw-shard",
467
+ "nbytes": 24780800,
468
+ "records": [
469
+ {
470
+ "name": "model.layers.11.mlp.gate_up_proj.q_weight",
471
+ "shape": [
472
+ 11264,
473
+ 256
474
+ ],
475
+ "dtype": "uint32",
476
+ "format": "f32-to-bf16",
477
+ "nbytes": 11534336,
478
+ "byteOffset": 0
479
+ },
480
+ {
481
+ "name": "model.layers.11.mlp.gate_up_proj.q_scale",
482
+ "shape": [
483
+ 11264,
484
+ 64
485
+ ],
486
+ "dtype": "float16",
487
+ "format": "f32-to-bf16",
488
+ "nbytes": 1441792,
489
+ "byteOffset": 11534336
490
+ },
491
+ {
492
+ "name": "model.layers.11.post_attention_layernorm.weight",
493
+ "shape": [
494
+ 2048
495
+ ],
496
+ "dtype": "float16",
497
+ "format": "f32-to-bf16",
498
+ "nbytes": 4096,
499
+ "byteOffset": 12976128
500
+ },
501
+ {
502
+ "name": "model.layers.11.self_attn.qkv_proj.q_weight",
503
+ "shape": [
504
+ 2560,
505
+ 256
506
+ ],
507
+ "dtype": "uint32",
508
+ "format": "f32-to-bf16",
509
+ "nbytes": 2621440,
510
+ "byteOffset": 12980224
511
+ },
512
+ {
513
+ "name": "model.layers.11.self_attn.qkv_proj.q_scale",
514
+ "shape": [
515
+ 2560,
516
+ 64
517
+ ],
518
+ "dtype": "float16",
519
+ "format": "f32-to-bf16",
520
+ "nbytes": 327680,
521
+ "byteOffset": 15601664
522
+ },
523
+ {
524
+ "name": "model.layers.11.self_attn.o_proj.q_weight",
525
+ "shape": [
526
+ 2048,
527
+ 256
528
+ ],
529
+ "dtype": "uint32",
530
+ "format": "f32-to-bf16",
531
+ "nbytes": 2097152,
532
+ "byteOffset": 15929344
533
+ },
534
+ {
535
+ "name": "model.layers.11.self_attn.o_proj.q_scale",
536
+ "shape": [
537
+ 2048,
538
+ 64
539
+ ],
540
+ "dtype": "float16",
541
+ "format": "f32-to-bf16",
542
+ "nbytes": 262144,
543
+ "byteOffset": 18026496
544
+ },
545
+ {
546
+ "name": "model.layers.12.input_layernorm.weight",
547
+ "shape": [
548
+ 2048
549
+ ],
550
+ "dtype": "float16",
551
+ "format": "f32-to-bf16",
552
+ "nbytes": 4096,
553
+ "byteOffset": 18288640
554
+ },
555
+ {
556
+ "name": "model.layers.12.mlp.down_proj.q_weight",
557
+ "shape": [
558
+ 2048,
559
+ 704
560
+ ],
561
+ "dtype": "uint32",
562
+ "format": "f32-to-bf16",
563
+ "nbytes": 5767168,
564
+ "byteOffset": 18292736
565
+ },
566
+ {
567
+ "name": "model.layers.12.mlp.down_proj.q_scale",
568
+ "shape": [
569
+ 2048,
570
+ 176
571
+ ],
572
+ "dtype": "float16",
573
+ "format": "f32-to-bf16",
574
+ "nbytes": 720896,
575
+ "byteOffset": 24059904
576
+ }
577
+ ],
578
+ "md5sum": "f8fa1159fa72bb6c264da79e2b89924b"
579
+ },
580
+ {
581
+ "dataPath": "params_shard_8.bin",
582
+ "format": "raw-shard",
583
+ "nbytes": 24780800,
584
+ "records": [
585
+ {
586
+ "name": "model.layers.12.mlp.gate_up_proj.q_weight",
587
+ "shape": [
588
+ 11264,
589
+ 256
590
+ ],
591
+ "dtype": "uint32",
592
+ "format": "f32-to-bf16",
593
+ "nbytes": 11534336,
594
+ "byteOffset": 0
595
+ },
596
+ {
597
+ "name": "model.layers.12.mlp.gate_up_proj.q_scale",
598
+ "shape": [
599
+ 11264,
600
+ 64
601
+ ],
602
+ "dtype": "float16",
603
+ "format": "f32-to-bf16",
604
+ "nbytes": 1441792,
605
+ "byteOffset": 11534336
606
+ },
607
+ {
608
+ "name": "model.layers.12.post_attention_layernorm.weight",
609
+ "shape": [
610
+ 2048
611
+ ],
612
+ "dtype": "float16",
613
+ "format": "f32-to-bf16",
614
+ "nbytes": 4096,
615
+ "byteOffset": 12976128
616
+ },
617
+ {
618
+ "name": "model.layers.12.self_attn.qkv_proj.q_weight",
619
+ "shape": [
620
+ 2560,
621
+ 256
622
+ ],
623
+ "dtype": "uint32",
624
+ "format": "f32-to-bf16",
625
+ "nbytes": 2621440,
626
+ "byteOffset": 12980224
627
+ },
628
+ {
629
+ "name": "model.layers.12.self_attn.qkv_proj.q_scale",
630
+ "shape": [
631
+ 2560,
632
+ 64
633
+ ],
634
+ "dtype": "float16",
635
+ "format": "f32-to-bf16",
636
+ "nbytes": 327680,
637
+ "byteOffset": 15601664
638
+ },
639
+ {
640
+ "name": "model.layers.12.self_attn.o_proj.q_weight",
641
+ "shape": [
642
+ 2048,
643
+ 256
644
+ ],
645
+ "dtype": "uint32",
646
+ "format": "f32-to-bf16",
647
+ "nbytes": 2097152,
648
+ "byteOffset": 15929344
649
+ },
650
+ {
651
+ "name": "model.layers.12.self_attn.o_proj.q_scale",
652
+ "shape": [
653
+ 2048,
654
+ 64
655
+ ],
656
+ "dtype": "float16",
657
+ "format": "f32-to-bf16",
658
+ "nbytes": 262144,
659
+ "byteOffset": 18026496
660
+ },
661
+ {
662
+ "name": "model.layers.13.input_layernorm.weight",
663
+ "shape": [
664
+ 2048
665
+ ],
666
+ "dtype": "float16",
667
+ "format": "f32-to-bf16",
668
+ "nbytes": 4096,
669
+ "byteOffset": 18288640
670
+ },
671
+ {
672
+ "name": "model.layers.13.mlp.down_proj.q_weight",
673
+ "shape": [
674
+ 2048,
675
+ 704
676
+ ],
677
+ "dtype": "uint32",
678
+ "format": "f32-to-bf16",
679
+ "nbytes": 5767168,
680
+ "byteOffset": 18292736
681
+ },
682
+ {
683
+ "name": "model.layers.13.mlp.down_proj.q_scale",
684
+ "shape": [
685
+ 2048,
686
+ 176
687
+ ],
688
+ "dtype": "float16",
689
+ "format": "f32-to-bf16",
690
+ "nbytes": 720896,
691
+ "byteOffset": 24059904
692
+ }
693
+ ],
694
+ "md5sum": "4567b9396fb721482e3fa6c249c9aa3a"
695
+ },
696
+ {
697
+ "dataPath": "params_shard_9.bin",
698
+ "format": "raw-shard",
699
+ "nbytes": 24780800,
700
+ "records": [
701
+ {
702
+ "name": "model.layers.13.mlp.gate_up_proj.q_weight",
703
+ "shape": [
704
+ 11264,
705
+ 256
706
+ ],
707
+ "dtype": "uint32",
708
+ "format": "f32-to-bf16",
709
+ "nbytes": 11534336,
710
+ "byteOffset": 0
711
+ },
712
+ {
713
+ "name": "model.layers.13.mlp.gate_up_proj.q_scale",
714
+ "shape": [
715
+ 11264,
716
+ 64
717
+ ],
718
+ "dtype": "float16",
719
+ "format": "f32-to-bf16",
720
+ "nbytes": 1441792,
721
+ "byteOffset": 11534336
722
+ },
723
+ {
724
+ "name": "model.layers.13.post_attention_layernorm.weight",
725
+ "shape": [
726
+ 2048
727
+ ],
728
+ "dtype": "float16",
729
+ "format": "f32-to-bf16",
730
+ "nbytes": 4096,
731
+ "byteOffset": 12976128
732
+ },
733
+ {
734
+ "name": "model.layers.13.self_attn.qkv_proj.q_weight",
735
+ "shape": [
736
+ 2560,
737
+ 256
738
+ ],
739
+ "dtype": "uint32",
740
+ "format": "f32-to-bf16",
741
+ "nbytes": 2621440,
742
+ "byteOffset": 12980224
743
+ },
744
+ {
745
+ "name": "model.layers.13.self_attn.qkv_proj.q_scale",
746
+ "shape": [
747
+ 2560,
748
+ 64
749
+ ],
750
+ "dtype": "float16",
751
+ "format": "f32-to-bf16",
752
+ "nbytes": 327680,
753
+ "byteOffset": 15601664
754
+ },
755
+ {
756
+ "name": "model.layers.13.self_attn.o_proj.q_weight",
757
+ "shape": [
758
+ 2048,
759
+ 256
760
+ ],
761
+ "dtype": "uint32",
762
+ "format": "f32-to-bf16",
763
+ "nbytes": 2097152,
764
+ "byteOffset": 15929344
765
+ },
766
+ {
767
+ "name": "model.layers.13.self_attn.o_proj.q_scale",
768
+ "shape": [
769
+ 2048,
770
+ 64
771
+ ],
772
+ "dtype": "float16",
773
+ "format": "f32-to-bf16",
774
+ "nbytes": 262144,
775
+ "byteOffset": 18026496
776
+ },
777
+ {
778
+ "name": "model.layers.14.input_layernorm.weight",
779
+ "shape": [
780
+ 2048
781
+ ],
782
+ "dtype": "float16",
783
+ "format": "f32-to-bf16",
784
+ "nbytes": 4096,
785
+ "byteOffset": 18288640
786
+ },
787
+ {
788
+ "name": "model.layers.14.mlp.down_proj.q_weight",
789
+ "shape": [
790
+ 2048,
791
+ 704
792
+ ],
793
+ "dtype": "uint32",
794
+ "format": "f32-to-bf16",
795
+ "nbytes": 5767168,
796
+ "byteOffset": 18292736
797
+ },
798
+ {
799
+ "name": "model.layers.14.mlp.down_proj.q_scale",
800
+ "shape": [
801
+ 2048,
802
+ 176
803
+ ],
804
+ "dtype": "float16",
805
+ "format": "f32-to-bf16",
806
+ "nbytes": 720896,
807
+ "byteOffset": 24059904
808
+ }
809
+ ],
810
+ "md5sum": "539a5193dfdb13cacdd3c3a4c88dcfbf"
811
+ },
812
+ {
813
+ "dataPath": "params_shard_10.bin",
814
+ "format": "raw-shard",
815
+ "nbytes": 24780800,
816
+ "records": [
817
+ {
818
+ "name": "model.layers.14.mlp.gate_up_proj.q_weight",
819
+ "shape": [
820
+ 11264,
821
+ 256
822
+ ],
823
+ "dtype": "uint32",
824
+ "format": "f32-to-bf16",
825
+ "nbytes": 11534336,
826
+ "byteOffset": 0
827
+ },
828
+ {
829
+ "name": "model.layers.14.mlp.gate_up_proj.q_scale",
830
+ "shape": [
831
+ 11264,
832
+ 64
833
+ ],
834
+ "dtype": "float16",
835
+ "format": "f32-to-bf16",
836
+ "nbytes": 1441792,
837
+ "byteOffset": 11534336
838
+ },
839
+ {
840
+ "name": "model.layers.14.post_attention_layernorm.weight",
841
+ "shape": [
842
+ 2048
843
+ ],
844
+ "dtype": "float16",
845
+ "format": "f32-to-bf16",
846
+ "nbytes": 4096,
847
+ "byteOffset": 12976128
848
+ },
849
+ {
850
+ "name": "model.layers.14.self_attn.qkv_proj.q_weight",
851
+ "shape": [
852
+ 2560,
853
+ 256
854
+ ],
855
+ "dtype": "uint32",
856
+ "format": "f32-to-bf16",
857
+ "nbytes": 2621440,
858
+ "byteOffset": 12980224
859
+ },
860
+ {
861
+ "name": "model.layers.14.self_attn.qkv_proj.q_scale",
862
+ "shape": [
863
+ 2560,
864
+ 64
865
+ ],
866
+ "dtype": "float16",
867
+ "format": "f32-to-bf16",
868
+ "nbytes": 327680,
869
+ "byteOffset": 15601664
870
+ },
871
+ {
872
+ "name": "model.layers.14.self_attn.o_proj.q_weight",
873
+ "shape": [
874
+ 2048,
875
+ 256
876
+ ],
877
+ "dtype": "uint32",
878
+ "format": "f32-to-bf16",
879
+ "nbytes": 2097152,
880
+ "byteOffset": 15929344
881
+ },
882
+ {
883
+ "name": "model.layers.14.self_attn.o_proj.q_scale",
884
+ "shape": [
885
+ 2048,
886
+ 64
887
+ ],
888
+ "dtype": "float16",
889
+ "format": "f32-to-bf16",
890
+ "nbytes": 262144,
891
+ "byteOffset": 18026496
892
+ },
893
+ {
894
+ "name": "model.layers.15.input_layernorm.weight",
895
+ "shape": [
896
+ 2048
897
+ ],
898
+ "dtype": "float16",
899
+ "format": "f32-to-bf16",
900
+ "nbytes": 4096,
901
+ "byteOffset": 18288640
902
+ },
903
+ {
904
+ "name": "model.layers.15.mlp.down_proj.q_weight",
905
+ "shape": [
906
+ 2048,
907
+ 704
908
+ ],
909
+ "dtype": "uint32",
910
+ "format": "f32-to-bf16",
911
+ "nbytes": 5767168,
912
+ "byteOffset": 18292736
913
+ },
914
+ {
915
+ "name": "model.layers.15.mlp.down_proj.q_scale",
916
+ "shape": [
917
+ 2048,
918
+ 176
919
+ ],
920
+ "dtype": "float16",
921
+ "format": "f32-to-bf16",
922
+ "nbytes": 720896,
923
+ "byteOffset": 24059904
924
+ }
925
+ ],
926
+ "md5sum": "0468961fda097b96aaf0e782f93f8b00"
927
+ },
928
+ {
929
+ "dataPath": "params_shard_11.bin",
930
+ "format": "raw-shard",
931
+ "nbytes": 24780800,
932
+ "records": [
933
+ {
934
+ "name": "model.layers.15.mlp.gate_up_proj.q_weight",
935
+ "shape": [
936
+ 11264,
937
+ 256
938
+ ],
939
+ "dtype": "uint32",
940
+ "format": "f32-to-bf16",
941
+ "nbytes": 11534336,
942
+ "byteOffset": 0
943
+ },
944
+ {
945
+ "name": "model.layers.15.mlp.gate_up_proj.q_scale",
946
+ "shape": [
947
+ 11264,
948
+ 64
949
+ ],
950
+ "dtype": "float16",
951
+ "format": "f32-to-bf16",
952
+ "nbytes": 1441792,
953
+ "byteOffset": 11534336
954
+ },
955
+ {
956
+ "name": "model.layers.15.post_attention_layernorm.weight",
957
+ "shape": [
958
+ 2048
959
+ ],
960
+ "dtype": "float16",
961
+ "format": "f32-to-bf16",
962
+ "nbytes": 4096,
963
+ "byteOffset": 12976128
964
+ },
965
+ {
966
+ "name": "model.layers.15.self_attn.qkv_proj.q_weight",
967
+ "shape": [
968
+ 2560,
969
+ 256
970
+ ],
971
+ "dtype": "uint32",
972
+ "format": "f32-to-bf16",
973
+ "nbytes": 2621440,
974
+ "byteOffset": 12980224
975
+ },
976
+ {
977
+ "name": "model.layers.15.self_attn.qkv_proj.q_scale",
978
+ "shape": [
979
+ 2560,
980
+ 64
981
+ ],
982
+ "dtype": "float16",
983
+ "format": "f32-to-bf16",
984
+ "nbytes": 327680,
985
+ "byteOffset": 15601664
986
+ },
987
+ {
988
+ "name": "model.layers.15.self_attn.o_proj.q_weight",
989
+ "shape": [
990
+ 2048,
991
+ 256
992
+ ],
993
+ "dtype": "uint32",
994
+ "format": "f32-to-bf16",
995
+ "nbytes": 2097152,
996
+ "byteOffset": 15929344
997
+ },
998
+ {
999
+ "name": "model.layers.15.self_attn.o_proj.q_scale",
1000
+ "shape": [
1001
+ 2048,
1002
+ 64
1003
+ ],
1004
+ "dtype": "float16",
1005
+ "format": "f32-to-bf16",
1006
+ "nbytes": 262144,
1007
+ "byteOffset": 18026496
1008
+ },
1009
+ {
1010
+ "name": "model.layers.16.input_layernorm.weight",
1011
+ "shape": [
1012
+ 2048
1013
+ ],
1014
+ "dtype": "float16",
1015
+ "format": "f32-to-bf16",
1016
+ "nbytes": 4096,
1017
+ "byteOffset": 18288640
1018
+ },
1019
+ {
1020
+ "name": "model.layers.16.mlp.down_proj.q_weight",
1021
+ "shape": [
1022
+ 2048,
1023
+ 704
1024
+ ],
1025
+ "dtype": "uint32",
1026
+ "format": "f32-to-bf16",
1027
+ "nbytes": 5767168,
1028
+ "byteOffset": 18292736
1029
+ },
1030
+ {
1031
+ "name": "model.layers.16.mlp.down_proj.q_scale",
1032
+ "shape": [
1033
+ 2048,
1034
+ 176
1035
+ ],
1036
+ "dtype": "float16",
1037
+ "format": "f32-to-bf16",
1038
+ "nbytes": 720896,
1039
+ "byteOffset": 24059904
1040
+ }
1041
+ ],
1042
+ "md5sum": "aa19f11c9d0b953ea02da32c9c1c129e"
1043
+ },
1044
+ {
1045
+ "dataPath": "params_shard_12.bin",
1046
+ "format": "raw-shard",
1047
+ "nbytes": 24780800,
1048
+ "records": [
1049
+ {
1050
+ "name": "model.layers.16.mlp.gate_up_proj.q_weight",
1051
+ "shape": [
1052
+ 11264,
1053
+ 256
1054
+ ],
1055
+ "dtype": "uint32",
1056
+ "format": "f32-to-bf16",
1057
+ "nbytes": 11534336,
1058
+ "byteOffset": 0
1059
+ },
1060
+ {
1061
+ "name": "model.layers.16.mlp.gate_up_proj.q_scale",
1062
+ "shape": [
1063
+ 11264,
1064
+ 64
1065
+ ],
1066
+ "dtype": "float16",
1067
+ "format": "f32-to-bf16",
1068
+ "nbytes": 1441792,
1069
+ "byteOffset": 11534336
1070
+ },
1071
+ {
1072
+ "name": "model.layers.16.post_attention_layernorm.weight",
1073
+ "shape": [
1074
+ 2048
1075
+ ],
1076
+ "dtype": "float16",
1077
+ "format": "f32-to-bf16",
1078
+ "nbytes": 4096,
1079
+ "byteOffset": 12976128
1080
+ },
1081
+ {
1082
+ "name": "model.layers.16.self_attn.qkv_proj.q_weight",
1083
+ "shape": [
1084
+ 2560,
1085
+ 256
1086
+ ],
1087
+ "dtype": "uint32",
1088
+ "format": "f32-to-bf16",
1089
+ "nbytes": 2621440,
1090
+ "byteOffset": 12980224
1091
+ },
1092
+ {
1093
+ "name": "model.layers.16.self_attn.qkv_proj.q_scale",
1094
+ "shape": [
1095
+ 2560,
1096
+ 64
1097
+ ],
1098
+ "dtype": "float16",
1099
+ "format": "f32-to-bf16",
1100
+ "nbytes": 327680,
1101
+ "byteOffset": 15601664
1102
+ },
1103
+ {
1104
+ "name": "model.layers.16.self_attn.o_proj.q_weight",
1105
+ "shape": [
1106
+ 2048,
1107
+ 256
1108
+ ],
1109
+ "dtype": "uint32",
1110
+ "format": "f32-to-bf16",
1111
+ "nbytes": 2097152,
1112
+ "byteOffset": 15929344
1113
+ },
1114
+ {
1115
+ "name": "model.layers.16.self_attn.o_proj.q_scale",
1116
+ "shape": [
1117
+ 2048,
1118
+ 64
1119
+ ],
1120
+ "dtype": "float16",
1121
+ "format": "f32-to-bf16",
1122
+ "nbytes": 262144,
1123
+ "byteOffset": 18026496
1124
+ },
1125
+ {
1126
+ "name": "model.layers.17.input_layernorm.weight",
1127
+ "shape": [
1128
+ 2048
1129
+ ],
1130
+ "dtype": "float16",
1131
+ "format": "f32-to-bf16",
1132
+ "nbytes": 4096,
1133
+ "byteOffset": 18288640
1134
+ },
1135
+ {
1136
+ "name": "model.layers.17.mlp.down_proj.q_weight",
1137
+ "shape": [
1138
+ 2048,
1139
+ 704
1140
+ ],
1141
+ "dtype": "uint32",
1142
+ "format": "f32-to-bf16",
1143
+ "nbytes": 5767168,
1144
+ "byteOffset": 18292736
1145
+ },
1146
+ {
1147
+ "name": "model.layers.17.mlp.down_proj.q_scale",
1148
+ "shape": [
1149
+ 2048,
1150
+ 176
1151
+ ],
1152
+ "dtype": "float16",
1153
+ "format": "f32-to-bf16",
1154
+ "nbytes": 720896,
1155
+ "byteOffset": 24059904
1156
+ }
1157
+ ],
1158
+ "md5sum": "0d1433365bb4e76f8212019422e6a437"
1159
+ },
1160
+ {
1161
+ "dataPath": "params_shard_13.bin",
1162
+ "format": "raw-shard",
1163
+ "nbytes": 24780800,
1164
+ "records": [
1165
+ {
1166
+ "name": "model.layers.17.mlp.gate_up_proj.q_weight",
1167
+ "shape": [
1168
+ 11264,
1169
+ 256
1170
+ ],
1171
+ "dtype": "uint32",
1172
+ "format": "f32-to-bf16",
1173
+ "nbytes": 11534336,
1174
+ "byteOffset": 0
1175
+ },
1176
+ {
1177
+ "name": "model.layers.17.mlp.gate_up_proj.q_scale",
1178
+ "shape": [
1179
+ 11264,
1180
+ 64
1181
+ ],
1182
+ "dtype": "float16",
1183
+ "format": "f32-to-bf16",
1184
+ "nbytes": 1441792,
1185
+ "byteOffset": 11534336
1186
+ },
1187
+ {
1188
+ "name": "model.layers.17.post_attention_layernorm.weight",
1189
+ "shape": [
1190
+ 2048
1191
+ ],
1192
+ "dtype": "float16",
1193
+ "format": "f32-to-bf16",
1194
+ "nbytes": 4096,
1195
+ "byteOffset": 12976128
1196
+ },
1197
+ {
1198
+ "name": "model.layers.17.self_attn.qkv_proj.q_weight",
1199
+ "shape": [
1200
+ 2560,
1201
+ 256
1202
+ ],
1203
+ "dtype": "uint32",
1204
+ "format": "f32-to-bf16",
1205
+ "nbytes": 2621440,
1206
+ "byteOffset": 12980224
1207
+ },
1208
+ {
1209
+ "name": "model.layers.17.self_attn.qkv_proj.q_scale",
1210
+ "shape": [
1211
+ 2560,
1212
+ 64
1213
+ ],
1214
+ "dtype": "float16",
1215
+ "format": "f32-to-bf16",
1216
+ "nbytes": 327680,
1217
+ "byteOffset": 15601664
1218
+ },
1219
+ {
1220
+ "name": "model.layers.17.self_attn.o_proj.q_weight",
1221
+ "shape": [
1222
+ 2048,
1223
+ 256
1224
+ ],
1225
+ "dtype": "uint32",
1226
+ "format": "f32-to-bf16",
1227
+ "nbytes": 2097152,
1228
+ "byteOffset": 15929344
1229
+ },
1230
+ {
1231
+ "name": "model.layers.17.self_attn.o_proj.q_scale",
1232
+ "shape": [
1233
+ 2048,
1234
+ 64
1235
+ ],
1236
+ "dtype": "float16",
1237
+ "format": "f32-to-bf16",
1238
+ "nbytes": 262144,
1239
+ "byteOffset": 18026496
1240
+ },
1241
+ {
1242
+ "name": "model.layers.18.input_layernorm.weight",
1243
+ "shape": [
1244
+ 2048
1245
+ ],
1246
+ "dtype": "float16",
1247
+ "format": "f32-to-bf16",
1248
+ "nbytes": 4096,
1249
+ "byteOffset": 18288640
1250
+ },
1251
+ {
1252
+ "name": "model.layers.18.mlp.down_proj.q_weight",
1253
+ "shape": [
1254
+ 2048,
1255
+ 704
1256
+ ],
1257
+ "dtype": "uint32",
1258
+ "format": "f32-to-bf16",
1259
+ "nbytes": 5767168,
1260
+ "byteOffset": 18292736
1261
+ },
1262
+ {
1263
+ "name": "model.layers.18.mlp.down_proj.q_scale",
1264
+ "shape": [
1265
+ 2048,
1266
+ 176
1267
+ ],
1268
+ "dtype": "float16",
1269
+ "format": "f32-to-bf16",
1270
+ "nbytes": 720896,
1271
+ "byteOffset": 24059904
1272
+ }
1273
+ ],
1274
+ "md5sum": "715aec75e345801dfb2b515cda65d382"
1275
+ },
1276
+ {
1277
+ "dataPath": "params_shard_14.bin",
1278
+ "format": "raw-shard",
1279
+ "nbytes": 24780800,
1280
+ "records": [
1281
+ {
1282
+ "name": "model.layers.18.mlp.gate_up_proj.q_weight",
1283
+ "shape": [
1284
+ 11264,
1285
+ 256
1286
+ ],
1287
+ "dtype": "uint32",
1288
+ "format": "f32-to-bf16",
1289
+ "nbytes": 11534336,
1290
+ "byteOffset": 0
1291
+ },
1292
+ {
1293
+ "name": "model.layers.18.mlp.gate_up_proj.q_scale",
1294
+ "shape": [
1295
+ 11264,
1296
+ 64
1297
+ ],
1298
+ "dtype": "float16",
1299
+ "format": "f32-to-bf16",
1300
+ "nbytes": 1441792,
1301
+ "byteOffset": 11534336
1302
+ },
1303
+ {
1304
+ "name": "model.layers.18.post_attention_layernorm.weight",
1305
+ "shape": [
1306
+ 2048
1307
+ ],
1308
+ "dtype": "float16",
1309
+ "format": "f32-to-bf16",
1310
+ "nbytes": 4096,
1311
+ "byteOffset": 12976128
1312
+ },
1313
+ {
1314
+ "name": "model.layers.18.self_attn.qkv_proj.q_weight",
1315
+ "shape": [
1316
+ 2560,
1317
+ 256
1318
+ ],
1319
+ "dtype": "uint32",
1320
+ "format": "f32-to-bf16",
1321
+ "nbytes": 2621440,
1322
+ "byteOffset": 12980224
1323
+ },
1324
+ {
1325
+ "name": "model.layers.18.self_attn.qkv_proj.q_scale",
1326
+ "shape": [
1327
+ 2560,
1328
+ 64
1329
+ ],
1330
+ "dtype": "float16",
1331
+ "format": "f32-to-bf16",
1332
+ "nbytes": 327680,
1333
+ "byteOffset": 15601664
1334
+ },
1335
+ {
1336
+ "name": "model.layers.18.self_attn.o_proj.q_weight",
1337
+ "shape": [
1338
+ 2048,
1339
+ 256
1340
+ ],
1341
+ "dtype": "uint32",
1342
+ "format": "f32-to-bf16",
1343
+ "nbytes": 2097152,
1344
+ "byteOffset": 15929344
1345
+ },
1346
+ {
1347
+ "name": "model.layers.18.self_attn.o_proj.q_scale",
1348
+ "shape": [
1349
+ 2048,
1350
+ 64
1351
+ ],
1352
+ "dtype": "float16",
1353
+ "format": "f32-to-bf16",
1354
+ "nbytes": 262144,
1355
+ "byteOffset": 18026496
1356
+ },
1357
+ {
1358
+ "name": "model.layers.19.input_layernorm.weight",
1359
+ "shape": [
1360
+ 2048
1361
+ ],
1362
+ "dtype": "float16",
1363
+ "format": "f32-to-bf16",
1364
+ "nbytes": 4096,
1365
+ "byteOffset": 18288640
1366
+ },
1367
+ {
1368
+ "name": "model.layers.19.mlp.down_proj.q_weight",
1369
+ "shape": [
1370
+ 2048,
1371
+ 704
1372
+ ],
1373
+ "dtype": "uint32",
1374
+ "format": "f32-to-bf16",
1375
+ "nbytes": 5767168,
1376
+ "byteOffset": 18292736
1377
+ },
1378
+ {
1379
+ "name": "model.layers.19.mlp.down_proj.q_scale",
1380
+ "shape": [
1381
+ 2048,
1382
+ 176
1383
+ ],
1384
+ "dtype": "float16",
1385
+ "format": "f32-to-bf16",
1386
+ "nbytes": 720896,
1387
+ "byteOffset": 24059904
1388
+ }
1389
+ ],
1390
+ "md5sum": "d5790615c3c2626be51ff0fc3a34fa1a"
1391
+ },
1392
+ {
1393
+ "dataPath": "params_shard_15.bin",
1394
+ "format": "raw-shard",
1395
+ "nbytes": 24780800,
1396
+ "records": [
1397
+ {
1398
+ "name": "model.layers.19.mlp.gate_up_proj.q_weight",
1399
+ "shape": [
1400
+ 11264,
1401
+ 256
1402
+ ],
1403
+ "dtype": "uint32",
1404
+ "format": "f32-to-bf16",
1405
+ "nbytes": 11534336,
1406
+ "byteOffset": 0
1407
+ },
1408
+ {
1409
+ "name": "model.layers.19.mlp.gate_up_proj.q_scale",
1410
+ "shape": [
1411
+ 11264,
1412
+ 64
1413
+ ],
1414
+ "dtype": "float16",
1415
+ "format": "f32-to-bf16",
1416
+ "nbytes": 1441792,
1417
+ "byteOffset": 11534336
1418
+ },
1419
+ {
1420
+ "name": "model.layers.19.post_attention_layernorm.weight",
1421
+ "shape": [
1422
+ 2048
1423
+ ],
1424
+ "dtype": "float16",
1425
+ "format": "f32-to-bf16",
1426
+ "nbytes": 4096,
1427
+ "byteOffset": 12976128
1428
+ },
1429
+ {
1430
+ "name": "model.layers.19.self_attn.qkv_proj.q_weight",
1431
+ "shape": [
1432
+ 2560,
1433
+ 256
1434
+ ],
1435
+ "dtype": "uint32",
1436
+ "format": "f32-to-bf16",
1437
+ "nbytes": 2621440,
1438
+ "byteOffset": 12980224
1439
+ },
1440
+ {
1441
+ "name": "model.layers.19.self_attn.qkv_proj.q_scale",
1442
+ "shape": [
1443
+ 2560,
1444
+ 64
1445
+ ],
1446
+ "dtype": "float16",
1447
+ "format": "f32-to-bf16",
1448
+ "nbytes": 327680,
1449
+ "byteOffset": 15601664
1450
+ },
1451
+ {
1452
+ "name": "model.layers.19.self_attn.o_proj.q_weight",
1453
+ "shape": [
1454
+ 2048,
1455
+ 256
1456
+ ],
1457
+ "dtype": "uint32",
1458
+ "format": "f32-to-bf16",
1459
+ "nbytes": 2097152,
1460
+ "byteOffset": 15929344
1461
+ },
1462
+ {
1463
+ "name": "model.layers.19.self_attn.o_proj.q_scale",
1464
+ "shape": [
1465
+ 2048,
1466
+ 64
1467
+ ],
1468
+ "dtype": "float16",
1469
+ "format": "f32-to-bf16",
1470
+ "nbytes": 262144,
1471
+ "byteOffset": 18026496
1472
+ },
1473
+ {
1474
+ "name": "model.layers.2.input_layernorm.weight",
1475
+ "shape": [
1476
+ 2048
1477
+ ],
1478
+ "dtype": "float16",
1479
+ "format": "f32-to-bf16",
1480
+ "nbytes": 4096,
1481
+ "byteOffset": 18288640
1482
+ },
1483
+ {
1484
+ "name": "model.layers.2.mlp.down_proj.q_weight",
1485
+ "shape": [
1486
+ 2048,
1487
+ 704
1488
+ ],
1489
+ "dtype": "uint32",
1490
+ "format": "f32-to-bf16",
1491
+ "nbytes": 5767168,
1492
+ "byteOffset": 18292736
1493
+ },
1494
+ {
1495
+ "name": "model.layers.2.mlp.down_proj.q_scale",
1496
+ "shape": [
1497
+ 2048,
1498
+ 176
1499
+ ],
1500
+ "dtype": "float16",
1501
+ "format": "f32-to-bf16",
1502
+ "nbytes": 720896,
1503
+ "byteOffset": 24059904
1504
+ }
1505
+ ],
1506
+ "md5sum": "d9bd6d310275218595c1d6484617426a"
1507
+ },
1508
+ {
1509
+ "dataPath": "params_shard_16.bin",
1510
+ "format": "raw-shard",
1511
+ "nbytes": 24780800,
1512
+ "records": [
1513
+ {
1514
+ "name": "model.layers.2.mlp.gate_up_proj.q_weight",
1515
+ "shape": [
1516
+ 11264,
1517
+ 256
1518
+ ],
1519
+ "dtype": "uint32",
1520
+ "format": "f32-to-bf16",
1521
+ "nbytes": 11534336,
1522
+ "byteOffset": 0
1523
+ },
1524
+ {
1525
+ "name": "model.layers.2.mlp.gate_up_proj.q_scale",
1526
+ "shape": [
1527
+ 11264,
1528
+ 64
1529
+ ],
1530
+ "dtype": "float16",
1531
+ "format": "f32-to-bf16",
1532
+ "nbytes": 1441792,
1533
+ "byteOffset": 11534336
1534
+ },
1535
+ {
1536
+ "name": "model.layers.2.post_attention_layernorm.weight",
1537
+ "shape": [
1538
+ 2048
1539
+ ],
1540
+ "dtype": "float16",
1541
+ "format": "f32-to-bf16",
1542
+ "nbytes": 4096,
1543
+ "byteOffset": 12976128
1544
+ },
1545
+ {
1546
+ "name": "model.layers.2.self_attn.qkv_proj.q_weight",
1547
+ "shape": [
1548
+ 2560,
1549
+ 256
1550
+ ],
1551
+ "dtype": "uint32",
1552
+ "format": "f32-to-bf16",
1553
+ "nbytes": 2621440,
1554
+ "byteOffset": 12980224
1555
+ },
1556
+ {
1557
+ "name": "model.layers.2.self_attn.qkv_proj.q_scale",
1558
+ "shape": [
1559
+ 2560,
1560
+ 64
1561
+ ],
1562
+ "dtype": "float16",
1563
+ "format": "f32-to-bf16",
1564
+ "nbytes": 327680,
1565
+ "byteOffset": 15601664
1566
+ },
1567
+ {
1568
+ "name": "model.layers.2.self_attn.o_proj.q_weight",
1569
+ "shape": [
1570
+ 2048,
1571
+ 256
1572
+ ],
1573
+ "dtype": "uint32",
1574
+ "format": "f32-to-bf16",
1575
+ "nbytes": 2097152,
1576
+ "byteOffset": 15929344
1577
+ },
1578
+ {
1579
+ "name": "model.layers.2.self_attn.o_proj.q_scale",
1580
+ "shape": [
1581
+ 2048,
1582
+ 64
1583
+ ],
1584
+ "dtype": "float16",
1585
+ "format": "f32-to-bf16",
1586
+ "nbytes": 262144,
1587
+ "byteOffset": 18026496
1588
+ },
1589
+ {
1590
+ "name": "model.layers.20.input_layernorm.weight",
1591
+ "shape": [
1592
+ 2048
1593
+ ],
1594
+ "dtype": "float16",
1595
+ "format": "f32-to-bf16",
1596
+ "nbytes": 4096,
1597
+ "byteOffset": 18288640
1598
+ },
1599
+ {
1600
+ "name": "model.layers.20.mlp.down_proj.q_weight",
1601
+ "shape": [
1602
+ 2048,
1603
+ 704
1604
+ ],
1605
+ "dtype": "uint32",
1606
+ "format": "f32-to-bf16",
1607
+ "nbytes": 5767168,
1608
+ "byteOffset": 18292736
1609
+ },
1610
+ {
1611
+ "name": "model.layers.20.mlp.down_proj.q_scale",
1612
+ "shape": [
1613
+ 2048,
1614
+ 176
1615
+ ],
1616
+ "dtype": "float16",
1617
+ "format": "f32-to-bf16",
1618
+ "nbytes": 720896,
1619
+ "byteOffset": 24059904
1620
+ }
1621
+ ],
1622
+ "md5sum": "955a56034913a0475558510118fce4a2"
1623
+ },
1624
+ {
1625
+ "dataPath": "params_shard_17.bin",
1626
+ "format": "raw-shard",
1627
+ "nbytes": 24780800,
1628
+ "records": [
1629
+ {
1630
+ "name": "model.layers.20.mlp.gate_up_proj.q_weight",
1631
+ "shape": [
1632
+ 11264,
1633
+ 256
1634
+ ],
1635
+ "dtype": "uint32",
1636
+ "format": "f32-to-bf16",
1637
+ "nbytes": 11534336,
1638
+ "byteOffset": 0
1639
+ },
1640
+ {
1641
+ "name": "model.layers.20.mlp.gate_up_proj.q_scale",
1642
+ "shape": [
1643
+ 11264,
1644
+ 64
1645
+ ],
1646
+ "dtype": "float16",
1647
+ "format": "f32-to-bf16",
1648
+ "nbytes": 1441792,
1649
+ "byteOffset": 11534336
1650
+ },
1651
+ {
1652
+ "name": "model.layers.20.post_attention_layernorm.weight",
1653
+ "shape": [
1654
+ 2048
1655
+ ],
1656
+ "dtype": "float16",
1657
+ "format": "f32-to-bf16",
1658
+ "nbytes": 4096,
1659
+ "byteOffset": 12976128
1660
+ },
1661
+ {
1662
+ "name": "model.layers.20.self_attn.qkv_proj.q_weight",
1663
+ "shape": [
1664
+ 2560,
1665
+ 256
1666
+ ],
1667
+ "dtype": "uint32",
1668
+ "format": "f32-to-bf16",
1669
+ "nbytes": 2621440,
1670
+ "byteOffset": 12980224
1671
+ },
1672
+ {
1673
+ "name": "model.layers.20.self_attn.qkv_proj.q_scale",
1674
+ "shape": [
1675
+ 2560,
1676
+ 64
1677
+ ],
1678
+ "dtype": "float16",
1679
+ "format": "f32-to-bf16",
1680
+ "nbytes": 327680,
1681
+ "byteOffset": 15601664
1682
+ },
1683
+ {
1684
+ "name": "model.layers.20.self_attn.o_proj.q_weight",
1685
+ "shape": [
1686
+ 2048,
1687
+ 256
1688
+ ],
1689
+ "dtype": "uint32",
1690
+ "format": "f32-to-bf16",
1691
+ "nbytes": 2097152,
1692
+ "byteOffset": 15929344
1693
+ },
1694
+ {
1695
+ "name": "model.layers.20.self_attn.o_proj.q_scale",
1696
+ "shape": [
1697
+ 2048,
1698
+ 64
1699
+ ],
1700
+ "dtype": "float16",
1701
+ "format": "f32-to-bf16",
1702
+ "nbytes": 262144,
1703
+ "byteOffset": 18026496
1704
+ },
1705
+ {
1706
+ "name": "model.layers.21.input_layernorm.weight",
1707
+ "shape": [
1708
+ 2048
1709
+ ],
1710
+ "dtype": "float16",
1711
+ "format": "f32-to-bf16",
1712
+ "nbytes": 4096,
1713
+ "byteOffset": 18288640
1714
+ },
1715
+ {
1716
+ "name": "model.layers.21.mlp.down_proj.q_weight",
1717
+ "shape": [
1718
+ 2048,
1719
+ 704
1720
+ ],
1721
+ "dtype": "uint32",
1722
+ "format": "f32-to-bf16",
1723
+ "nbytes": 5767168,
1724
+ "byteOffset": 18292736
1725
+ },
1726
+ {
1727
+ "name": "model.layers.21.mlp.down_proj.q_scale",
1728
+ "shape": [
1729
+ 2048,
1730
+ 176
1731
+ ],
1732
+ "dtype": "float16",
1733
+ "format": "f32-to-bf16",
1734
+ "nbytes": 720896,
1735
+ "byteOffset": 24059904
1736
+ }
1737
+ ],
1738
+ "md5sum": "e264ee20d99d95912760bafd8c76e2d0"
1739
+ },
1740
+ {
1741
+ "dataPath": "params_shard_18.bin",
1742
+ "format": "raw-shard",
1743
+ "nbytes": 24780800,
1744
+ "records": [
1745
+ {
1746
+ "name": "model.layers.21.mlp.gate_up_proj.q_weight",
1747
+ "shape": [
1748
+ 11264,
1749
+ 256
1750
+ ],
1751
+ "dtype": "uint32",
1752
+ "format": "f32-to-bf16",
1753
+ "nbytes": 11534336,
1754
+ "byteOffset": 0
1755
+ },
1756
+ {
1757
+ "name": "model.layers.21.mlp.gate_up_proj.q_scale",
1758
+ "shape": [
1759
+ 11264,
1760
+ 64
1761
+ ],
1762
+ "dtype": "float16",
1763
+ "format": "f32-to-bf16",
1764
+ "nbytes": 1441792,
1765
+ "byteOffset": 11534336
1766
+ },
1767
+ {
1768
+ "name": "model.layers.21.post_attention_layernorm.weight",
1769
+ "shape": [
1770
+ 2048
1771
+ ],
1772
+ "dtype": "float16",
1773
+ "format": "f32-to-bf16",
1774
+ "nbytes": 4096,
1775
+ "byteOffset": 12976128
1776
+ },
1777
+ {
1778
+ "name": "model.layers.21.self_attn.qkv_proj.q_weight",
1779
+ "shape": [
1780
+ 2560,
1781
+ 256
1782
+ ],
1783
+ "dtype": "uint32",
1784
+ "format": "f32-to-bf16",
1785
+ "nbytes": 2621440,
1786
+ "byteOffset": 12980224
1787
+ },
1788
+ {
1789
+ "name": "model.layers.21.self_attn.qkv_proj.q_scale",
1790
+ "shape": [
1791
+ 2560,
1792
+ 64
1793
+ ],
1794
+ "dtype": "float16",
1795
+ "format": "f32-to-bf16",
1796
+ "nbytes": 327680,
1797
+ "byteOffset": 15601664
1798
+ },
1799
+ {
1800
+ "name": "model.layers.21.self_attn.o_proj.q_weight",
1801
+ "shape": [
1802
+ 2048,
1803
+ 256
1804
+ ],
1805
+ "dtype": "uint32",
1806
+ "format": "f32-to-bf16",
1807
+ "nbytes": 2097152,
1808
+ "byteOffset": 15929344
1809
+ },
1810
+ {
1811
+ "name": "model.layers.21.self_attn.o_proj.q_scale",
1812
+ "shape": [
1813
+ 2048,
1814
+ 64
1815
+ ],
1816
+ "dtype": "float16",
1817
+ "format": "f32-to-bf16",
1818
+ "nbytes": 262144,
1819
+ "byteOffset": 18026496
1820
+ },
1821
+ {
1822
+ "name": "model.layers.3.input_layernorm.weight",
1823
+ "shape": [
1824
+ 2048
1825
+ ],
1826
+ "dtype": "float16",
1827
+ "format": "f32-to-bf16",
1828
+ "nbytes": 4096,
1829
+ "byteOffset": 18288640
1830
+ },
1831
+ {
1832
+ "name": "model.layers.3.mlp.down_proj.q_weight",
1833
+ "shape": [
1834
+ 2048,
1835
+ 704
1836
+ ],
1837
+ "dtype": "uint32",
1838
+ "format": "f32-to-bf16",
1839
+ "nbytes": 5767168,
1840
+ "byteOffset": 18292736
1841
+ },
1842
+ {
1843
+ "name": "model.layers.3.mlp.down_proj.q_scale",
1844
+ "shape": [
1845
+ 2048,
1846
+ 176
1847
+ ],
1848
+ "dtype": "float16",
1849
+ "format": "f32-to-bf16",
1850
+ "nbytes": 720896,
1851
+ "byteOffset": 24059904
1852
+ }
1853
+ ],
1854
+ "md5sum": "6df816583a637670927e9adc7ee55b56"
1855
+ },
1856
+ {
1857
+ "dataPath": "params_shard_19.bin",
1858
+ "format": "raw-shard",
1859
+ "nbytes": 24780800,
1860
+ "records": [
1861
+ {
1862
+ "name": "model.layers.3.mlp.gate_up_proj.q_weight",
1863
+ "shape": [
1864
+ 11264,
1865
+ 256
1866
+ ],
1867
+ "dtype": "uint32",
1868
+ "format": "f32-to-bf16",
1869
+ "nbytes": 11534336,
1870
+ "byteOffset": 0
1871
+ },
1872
+ {
1873
+ "name": "model.layers.3.mlp.gate_up_proj.q_scale",
1874
+ "shape": [
1875
+ 11264,
1876
+ 64
1877
+ ],
1878
+ "dtype": "float16",
1879
+ "format": "f32-to-bf16",
1880
+ "nbytes": 1441792,
1881
+ "byteOffset": 11534336
1882
+ },
1883
+ {
1884
+ "name": "model.layers.3.post_attention_layernorm.weight",
1885
+ "shape": [
1886
+ 2048
1887
+ ],
1888
+ "dtype": "float16",
1889
+ "format": "f32-to-bf16",
1890
+ "nbytes": 4096,
1891
+ "byteOffset": 12976128
1892
+ },
1893
+ {
1894
+ "name": "model.layers.3.self_attn.qkv_proj.q_weight",
1895
+ "shape": [
1896
+ 2560,
1897
+ 256
1898
+ ],
1899
+ "dtype": "uint32",
1900
+ "format": "f32-to-bf16",
1901
+ "nbytes": 2621440,
1902
+ "byteOffset": 12980224
1903
+ },
1904
+ {
1905
+ "name": "model.layers.3.self_attn.qkv_proj.q_scale",
1906
+ "shape": [
1907
+ 2560,
1908
+ 64
1909
+ ],
1910
+ "dtype": "float16",
1911
+ "format": "f32-to-bf16",
1912
+ "nbytes": 327680,
1913
+ "byteOffset": 15601664
1914
+ },
1915
+ {
1916
+ "name": "model.layers.3.self_attn.o_proj.q_weight",
1917
+ "shape": [
1918
+ 2048,
1919
+ 256
1920
+ ],
1921
+ "dtype": "uint32",
1922
+ "format": "f32-to-bf16",
1923
+ "nbytes": 2097152,
1924
+ "byteOffset": 15929344
1925
+ },
1926
+ {
1927
+ "name": "model.layers.3.self_attn.o_proj.q_scale",
1928
+ "shape": [
1929
+ 2048,
1930
+ 64
1931
+ ],
1932
+ "dtype": "float16",
1933
+ "format": "f32-to-bf16",
1934
+ "nbytes": 262144,
1935
+ "byteOffset": 18026496
1936
+ },
1937
+ {
1938
+ "name": "model.layers.4.input_layernorm.weight",
1939
+ "shape": [
1940
+ 2048
1941
+ ],
1942
+ "dtype": "float16",
1943
+ "format": "f32-to-bf16",
1944
+ "nbytes": 4096,
1945
+ "byteOffset": 18288640
1946
+ },
1947
+ {
1948
+ "name": "model.layers.4.mlp.down_proj.q_weight",
1949
+ "shape": [
1950
+ 2048,
1951
+ 704
1952
+ ],
1953
+ "dtype": "uint32",
1954
+ "format": "f32-to-bf16",
1955
+ "nbytes": 5767168,
1956
+ "byteOffset": 18292736
1957
+ },
1958
+ {
1959
+ "name": "model.layers.4.mlp.down_proj.q_scale",
1960
+ "shape": [
1961
+ 2048,
1962
+ 176
1963
+ ],
1964
+ "dtype": "float16",
1965
+ "format": "f32-to-bf16",
1966
+ "nbytes": 720896,
1967
+ "byteOffset": 24059904
1968
+ }
1969
+ ],
1970
+ "md5sum": "20d577ebe3dff0dbe6b1f32e87604d59"
1971
+ },
1972
+ {
1973
+ "dataPath": "params_shard_20.bin",
1974
+ "format": "raw-shard",
1975
+ "nbytes": 24780800,
1976
+ "records": [
1977
+ {
1978
+ "name": "model.layers.4.mlp.gate_up_proj.q_weight",
1979
+ "shape": [
1980
+ 11264,
1981
+ 256
1982
+ ],
1983
+ "dtype": "uint32",
1984
+ "format": "f32-to-bf16",
1985
+ "nbytes": 11534336,
1986
+ "byteOffset": 0
1987
+ },
1988
+ {
1989
+ "name": "model.layers.4.mlp.gate_up_proj.q_scale",
1990
+ "shape": [
1991
+ 11264,
1992
+ 64
1993
+ ],
1994
+ "dtype": "float16",
1995
+ "format": "f32-to-bf16",
1996
+ "nbytes": 1441792,
1997
+ "byteOffset": 11534336
1998
+ },
1999
+ {
2000
+ "name": "model.layers.4.post_attention_layernorm.weight",
2001
+ "shape": [
2002
+ 2048
2003
+ ],
2004
+ "dtype": "float16",
2005
+ "format": "f32-to-bf16",
2006
+ "nbytes": 4096,
2007
+ "byteOffset": 12976128
2008
+ },
2009
+ {
2010
+ "name": "model.layers.4.self_attn.qkv_proj.q_weight",
2011
+ "shape": [
2012
+ 2560,
2013
+ 256
2014
+ ],
2015
+ "dtype": "uint32",
2016
+ "format": "f32-to-bf16",
2017
+ "nbytes": 2621440,
2018
+ "byteOffset": 12980224
2019
+ },
2020
+ {
2021
+ "name": "model.layers.4.self_attn.qkv_proj.q_scale",
2022
+ "shape": [
2023
+ 2560,
2024
+ 64
2025
+ ],
2026
+ "dtype": "float16",
2027
+ "format": "f32-to-bf16",
2028
+ "nbytes": 327680,
2029
+ "byteOffset": 15601664
2030
+ },
2031
+ {
2032
+ "name": "model.layers.4.self_attn.o_proj.q_weight",
2033
+ "shape": [
2034
+ 2048,
2035
+ 256
2036
+ ],
2037
+ "dtype": "uint32",
2038
+ "format": "f32-to-bf16",
2039
+ "nbytes": 2097152,
2040
+ "byteOffset": 15929344
2041
+ },
2042
+ {
2043
+ "name": "model.layers.4.self_attn.o_proj.q_scale",
2044
+ "shape": [
2045
+ 2048,
2046
+ 64
2047
+ ],
2048
+ "dtype": "float16",
2049
+ "format": "f32-to-bf16",
2050
+ "nbytes": 262144,
2051
+ "byteOffset": 18026496
2052
+ },
2053
+ {
2054
+ "name": "model.layers.5.input_layernorm.weight",
2055
+ "shape": [
2056
+ 2048
2057
+ ],
2058
+ "dtype": "float16",
2059
+ "format": "f32-to-bf16",
2060
+ "nbytes": 4096,
2061
+ "byteOffset": 18288640
2062
+ },
2063
+ {
2064
+ "name": "model.layers.5.mlp.down_proj.q_weight",
2065
+ "shape": [
2066
+ 2048,
2067
+ 704
2068
+ ],
2069
+ "dtype": "uint32",
2070
+ "format": "f32-to-bf16",
2071
+ "nbytes": 5767168,
2072
+ "byteOffset": 18292736
2073
+ },
2074
+ {
2075
+ "name": "model.layers.5.mlp.down_proj.q_scale",
2076
+ "shape": [
2077
+ 2048,
2078
+ 176
2079
+ ],
2080
+ "dtype": "float16",
2081
+ "format": "f32-to-bf16",
2082
+ "nbytes": 720896,
2083
+ "byteOffset": 24059904
2084
+ }
2085
+ ],
2086
+ "md5sum": "740f9acb4525b2383170be87bb22e7a0"
2087
+ },
2088
+ {
2089
+ "dataPath": "params_shard_21.bin",
2090
+ "format": "raw-shard",
2091
+ "nbytes": 24780800,
2092
+ "records": [
2093
+ {
2094
+ "name": "model.layers.5.mlp.gate_up_proj.q_weight",
2095
+ "shape": [
2096
+ 11264,
2097
+ 256
2098
+ ],
2099
+ "dtype": "uint32",
2100
+ "format": "f32-to-bf16",
2101
+ "nbytes": 11534336,
2102
+ "byteOffset": 0
2103
+ },
2104
+ {
2105
+ "name": "model.layers.5.mlp.gate_up_proj.q_scale",
2106
+ "shape": [
2107
+ 11264,
2108
+ 64
2109
+ ],
2110
+ "dtype": "float16",
2111
+ "format": "f32-to-bf16",
2112
+ "nbytes": 1441792,
2113
+ "byteOffset": 11534336
2114
+ },
2115
+ {
2116
+ "name": "model.layers.5.post_attention_layernorm.weight",
2117
+ "shape": [
2118
+ 2048
2119
+ ],
2120
+ "dtype": "float16",
2121
+ "format": "f32-to-bf16",
2122
+ "nbytes": 4096,
2123
+ "byteOffset": 12976128
2124
+ },
2125
+ {
2126
+ "name": "model.layers.5.self_attn.qkv_proj.q_weight",
2127
+ "shape": [
2128
+ 2560,
2129
+ 256
2130
+ ],
2131
+ "dtype": "uint32",
2132
+ "format": "f32-to-bf16",
2133
+ "nbytes": 2621440,
2134
+ "byteOffset": 12980224
2135
+ },
2136
+ {
2137
+ "name": "model.layers.5.self_attn.qkv_proj.q_scale",
2138
+ "shape": [
2139
+ 2560,
2140
+ 64
2141
+ ],
2142
+ "dtype": "float16",
2143
+ "format": "f32-to-bf16",
2144
+ "nbytes": 327680,
2145
+ "byteOffset": 15601664
2146
+ },
2147
+ {
2148
+ "name": "model.layers.5.self_attn.o_proj.q_weight",
2149
+ "shape": [
2150
+ 2048,
2151
+ 256
2152
+ ],
2153
+ "dtype": "uint32",
2154
+ "format": "f32-to-bf16",
2155
+ "nbytes": 2097152,
2156
+ "byteOffset": 15929344
2157
+ },
2158
+ {
2159
+ "name": "model.layers.5.self_attn.o_proj.q_scale",
2160
+ "shape": [
2161
+ 2048,
2162
+ 64
2163
+ ],
2164
+ "dtype": "float16",
2165
+ "format": "f32-to-bf16",
2166
+ "nbytes": 262144,
2167
+ "byteOffset": 18026496
2168
+ },
2169
+ {
2170
+ "name": "model.layers.6.input_layernorm.weight",
2171
+ "shape": [
2172
+ 2048
2173
+ ],
2174
+ "dtype": "float16",
2175
+ "format": "f32-to-bf16",
2176
+ "nbytes": 4096,
2177
+ "byteOffset": 18288640
2178
+ },
2179
+ {
2180
+ "name": "model.layers.6.mlp.down_proj.q_weight",
2181
+ "shape": [
2182
+ 2048,
2183
+ 704
2184
+ ],
2185
+ "dtype": "uint32",
2186
+ "format": "f32-to-bf16",
2187
+ "nbytes": 5767168,
2188
+ "byteOffset": 18292736
2189
+ },
2190
+ {
2191
+ "name": "model.layers.6.mlp.down_proj.q_scale",
2192
+ "shape": [
2193
+ 2048,
2194
+ 176
2195
+ ],
2196
+ "dtype": "float16",
2197
+ "format": "f32-to-bf16",
2198
+ "nbytes": 720896,
2199
+ "byteOffset": 24059904
2200
+ }
2201
+ ],
2202
+ "md5sum": "eccaea4f41ba41cb9a61adb8d30778f2"
2203
+ },
2204
+ {
2205
+ "dataPath": "params_shard_22.bin",
2206
+ "format": "raw-shard",
2207
+ "nbytes": 24780800,
2208
+ "records": [
2209
+ {
2210
+ "name": "model.layers.6.mlp.gate_up_proj.q_weight",
2211
+ "shape": [
2212
+ 11264,
2213
+ 256
2214
+ ],
2215
+ "dtype": "uint32",
2216
+ "format": "f32-to-bf16",
2217
+ "nbytes": 11534336,
2218
+ "byteOffset": 0
2219
+ },
2220
+ {
2221
+ "name": "model.layers.6.mlp.gate_up_proj.q_scale",
2222
+ "shape": [
2223
+ 11264,
2224
+ 64
2225
+ ],
2226
+ "dtype": "float16",
2227
+ "format": "f32-to-bf16",
2228
+ "nbytes": 1441792,
2229
+ "byteOffset": 11534336
2230
+ },
2231
+ {
2232
+ "name": "model.layers.6.post_attention_layernorm.weight",
2233
+ "shape": [
2234
+ 2048
2235
+ ],
2236
+ "dtype": "float16",
2237
+ "format": "f32-to-bf16",
2238
+ "nbytes": 4096,
2239
+ "byteOffset": 12976128
2240
+ },
2241
+ {
2242
+ "name": "model.layers.6.self_attn.qkv_proj.q_weight",
2243
+ "shape": [
2244
+ 2560,
2245
+ 256
2246
+ ],
2247
+ "dtype": "uint32",
2248
+ "format": "f32-to-bf16",
2249
+ "nbytes": 2621440,
2250
+ "byteOffset": 12980224
2251
+ },
2252
+ {
2253
+ "name": "model.layers.6.self_attn.qkv_proj.q_scale",
2254
+ "shape": [
2255
+ 2560,
2256
+ 64
2257
+ ],
2258
+ "dtype": "float16",
2259
+ "format": "f32-to-bf16",
2260
+ "nbytes": 327680,
2261
+ "byteOffset": 15601664
2262
+ },
2263
+ {
2264
+ "name": "model.layers.6.self_attn.o_proj.q_weight",
2265
+ "shape": [
2266
+ 2048,
2267
+ 256
2268
+ ],
2269
+ "dtype": "uint32",
2270
+ "format": "f32-to-bf16",
2271
+ "nbytes": 2097152,
2272
+ "byteOffset": 15929344
2273
+ },
2274
+ {
2275
+ "name": "model.layers.6.self_attn.o_proj.q_scale",
2276
+ "shape": [
2277
+ 2048,
2278
+ 64
2279
+ ],
2280
+ "dtype": "float16",
2281
+ "format": "f32-to-bf16",
2282
+ "nbytes": 262144,
2283
+ "byteOffset": 18026496
2284
+ },
2285
+ {
2286
+ "name": "model.layers.7.input_layernorm.weight",
2287
+ "shape": [
2288
+ 2048
2289
+ ],
2290
+ "dtype": "float16",
2291
+ "format": "f32-to-bf16",
2292
+ "nbytes": 4096,
2293
+ "byteOffset": 18288640
2294
+ },
2295
+ {
2296
+ "name": "model.layers.7.mlp.down_proj.q_weight",
2297
+ "shape": [
2298
+ 2048,
2299
+ 704
2300
+ ],
2301
+ "dtype": "uint32",
2302
+ "format": "f32-to-bf16",
2303
+ "nbytes": 5767168,
2304
+ "byteOffset": 18292736
2305
+ },
2306
+ {
2307
+ "name": "model.layers.7.mlp.down_proj.q_scale",
2308
+ "shape": [
2309
+ 2048,
2310
+ 176
2311
+ ],
2312
+ "dtype": "float16",
2313
+ "format": "f32-to-bf16",
2314
+ "nbytes": 720896,
2315
+ "byteOffset": 24059904
2316
+ }
2317
+ ],
2318
+ "md5sum": "13ae56bbebe69cf1bf4b269193bf55b1"
2319
+ },
2320
+ {
2321
+ "dataPath": "params_shard_23.bin",
2322
+ "format": "raw-shard",
2323
+ "nbytes": 24780800,
2324
+ "records": [
2325
+ {
2326
+ "name": "model.layers.7.mlp.gate_up_proj.q_weight",
2327
+ "shape": [
2328
+ 11264,
2329
+ 256
2330
+ ],
2331
+ "dtype": "uint32",
2332
+ "format": "f32-to-bf16",
2333
+ "nbytes": 11534336,
2334
+ "byteOffset": 0
2335
+ },
2336
+ {
2337
+ "name": "model.layers.7.mlp.gate_up_proj.q_scale",
2338
+ "shape": [
2339
+ 11264,
2340
+ 64
2341
+ ],
2342
+ "dtype": "float16",
2343
+ "format": "f32-to-bf16",
2344
+ "nbytes": 1441792,
2345
+ "byteOffset": 11534336
2346
+ },
2347
+ {
2348
+ "name": "model.layers.7.post_attention_layernorm.weight",
2349
+ "shape": [
2350
+ 2048
2351
+ ],
2352
+ "dtype": "float16",
2353
+ "format": "f32-to-bf16",
2354
+ "nbytes": 4096,
2355
+ "byteOffset": 12976128
2356
+ },
2357
+ {
2358
+ "name": "model.layers.7.self_attn.qkv_proj.q_weight",
2359
+ "shape": [
2360
+ 2560,
2361
+ 256
2362
+ ],
2363
+ "dtype": "uint32",
2364
+ "format": "f32-to-bf16",
2365
+ "nbytes": 2621440,
2366
+ "byteOffset": 12980224
2367
+ },
2368
+ {
2369
+ "name": "model.layers.7.self_attn.qkv_proj.q_scale",
2370
+ "shape": [
2371
+ 2560,
2372
+ 64
2373
+ ],
2374
+ "dtype": "float16",
2375
+ "format": "f32-to-bf16",
2376
+ "nbytes": 327680,
2377
+ "byteOffset": 15601664
2378
+ },
2379
+ {
2380
+ "name": "model.layers.7.self_attn.o_proj.q_weight",
2381
+ "shape": [
2382
+ 2048,
2383
+ 256
2384
+ ],
2385
+ "dtype": "uint32",
2386
+ "format": "f32-to-bf16",
2387
+ "nbytes": 2097152,
2388
+ "byteOffset": 15929344
2389
+ },
2390
+ {
2391
+ "name": "model.layers.7.self_attn.o_proj.q_scale",
2392
+ "shape": [
2393
+ 2048,
2394
+ 64
2395
+ ],
2396
+ "dtype": "float16",
2397
+ "format": "f32-to-bf16",
2398
+ "nbytes": 262144,
2399
+ "byteOffset": 18026496
2400
+ },
2401
+ {
2402
+ "name": "model.layers.8.input_layernorm.weight",
2403
+ "shape": [
2404
+ 2048
2405
+ ],
2406
+ "dtype": "float16",
2407
+ "format": "f32-to-bf16",
2408
+ "nbytes": 4096,
2409
+ "byteOffset": 18288640
2410
+ },
2411
+ {
2412
+ "name": "model.layers.8.mlp.down_proj.q_weight",
2413
+ "shape": [
2414
+ 2048,
2415
+ 704
2416
+ ],
2417
+ "dtype": "uint32",
2418
+ "format": "f32-to-bf16",
2419
+ "nbytes": 5767168,
2420
+ "byteOffset": 18292736
2421
+ },
2422
+ {
2423
+ "name": "model.layers.8.mlp.down_proj.q_scale",
2424
+ "shape": [
2425
+ 2048,
2426
+ 176
2427
+ ],
2428
+ "dtype": "float16",
2429
+ "format": "f32-to-bf16",
2430
+ "nbytes": 720896,
2431
+ "byteOffset": 24059904
2432
+ }
2433
+ ],
2434
+ "md5sum": "251e4c557bc60412e9b0180f8a246945"
2435
+ },
2436
+ {
2437
+ "dataPath": "params_shard_24.bin",
2438
+ "format": "raw-shard",
2439
+ "nbytes": 24780800,
2440
+ "records": [
2441
+ {
2442
+ "name": "model.layers.8.mlp.gate_up_proj.q_weight",
2443
+ "shape": [
2444
+ 11264,
2445
+ 256
2446
+ ],
2447
+ "dtype": "uint32",
2448
+ "format": "f32-to-bf16",
2449
+ "nbytes": 11534336,
2450
+ "byteOffset": 0
2451
+ },
2452
+ {
2453
+ "name": "model.layers.8.mlp.gate_up_proj.q_scale",
2454
+ "shape": [
2455
+ 11264,
2456
+ 64
2457
+ ],
2458
+ "dtype": "float16",
2459
+ "format": "f32-to-bf16",
2460
+ "nbytes": 1441792,
2461
+ "byteOffset": 11534336
2462
+ },
2463
+ {
2464
+ "name": "model.layers.8.post_attention_layernorm.weight",
2465
+ "shape": [
2466
+ 2048
2467
+ ],
2468
+ "dtype": "float16",
2469
+ "format": "f32-to-bf16",
2470
+ "nbytes": 4096,
2471
+ "byteOffset": 12976128
2472
+ },
2473
+ {
2474
+ "name": "model.layers.8.self_attn.qkv_proj.q_weight",
2475
+ "shape": [
2476
+ 2560,
2477
+ 256
2478
+ ],
2479
+ "dtype": "uint32",
2480
+ "format": "f32-to-bf16",
2481
+ "nbytes": 2621440,
2482
+ "byteOffset": 12980224
2483
+ },
2484
+ {
2485
+ "name": "model.layers.8.self_attn.qkv_proj.q_scale",
2486
+ "shape": [
2487
+ 2560,
2488
+ 64
2489
+ ],
2490
+ "dtype": "float16",
2491
+ "format": "f32-to-bf16",
2492
+ "nbytes": 327680,
2493
+ "byteOffset": 15601664
2494
+ },
2495
+ {
2496
+ "name": "model.layers.8.self_attn.o_proj.q_weight",
2497
+ "shape": [
2498
+ 2048,
2499
+ 256
2500
+ ],
2501
+ "dtype": "uint32",
2502
+ "format": "f32-to-bf16",
2503
+ "nbytes": 2097152,
2504
+ "byteOffset": 15929344
2505
+ },
2506
+ {
2507
+ "name": "model.layers.8.self_attn.o_proj.q_scale",
2508
+ "shape": [
2509
+ 2048,
2510
+ 64
2511
+ ],
2512
+ "dtype": "float16",
2513
+ "format": "f32-to-bf16",
2514
+ "nbytes": 262144,
2515
+ "byteOffset": 18026496
2516
+ },
2517
+ {
2518
+ "name": "model.layers.9.input_layernorm.weight",
2519
+ "shape": [
2520
+ 2048
2521
+ ],
2522
+ "dtype": "float16",
2523
+ "format": "f32-to-bf16",
2524
+ "nbytes": 4096,
2525
+ "byteOffset": 18288640
2526
+ },
2527
+ {
2528
+ "name": "model.layers.9.mlp.down_proj.q_weight",
2529
+ "shape": [
2530
+ 2048,
2531
+ 704
2532
+ ],
2533
+ "dtype": "uint32",
2534
+ "format": "f32-to-bf16",
2535
+ "nbytes": 5767168,
2536
+ "byteOffset": 18292736
2537
+ },
2538
+ {
2539
+ "name": "model.layers.9.mlp.down_proj.q_scale",
2540
+ "shape": [
2541
+ 2048,
2542
+ 176
2543
+ ],
2544
+ "dtype": "float16",
2545
+ "format": "f32-to-bf16",
2546
+ "nbytes": 720896,
2547
+ "byteOffset": 24059904
2548
+ }
2549
+ ],
2550
+ "md5sum": "3f8563933d4248222bea59edc6f84444"
2551
+ },
2552
+ {
2553
+ "dataPath": "params_shard_25.bin",
2554
+ "format": "raw-shard",
2555
+ "nbytes": 18292736,
2556
+ "records": [
2557
+ {
2558
+ "name": "model.layers.9.mlp.gate_up_proj.q_weight",
2559
+ "shape": [
2560
+ 11264,
2561
+ 256
2562
+ ],
2563
+ "dtype": "uint32",
2564
+ "format": "f32-to-bf16",
2565
+ "nbytes": 11534336,
2566
+ "byteOffset": 0
2567
+ },
2568
+ {
2569
+ "name": "model.layers.9.mlp.gate_up_proj.q_scale",
2570
+ "shape": [
2571
+ 11264,
2572
+ 64
2573
+ ],
2574
+ "dtype": "float16",
2575
+ "format": "f32-to-bf16",
2576
+ "nbytes": 1441792,
2577
+ "byteOffset": 11534336
2578
+ },
2579
+ {
2580
+ "name": "model.layers.9.post_attention_layernorm.weight",
2581
+ "shape": [
2582
+ 2048
2583
+ ],
2584
+ "dtype": "float16",
2585
+ "format": "f32-to-bf16",
2586
+ "nbytes": 4096,
2587
+ "byteOffset": 12976128
2588
+ },
2589
+ {
2590
+ "name": "model.layers.9.self_attn.qkv_proj.q_weight",
2591
+ "shape": [
2592
+ 2560,
2593
+ 256
2594
+ ],
2595
+ "dtype": "uint32",
2596
+ "format": "f32-to-bf16",
2597
+ "nbytes": 2621440,
2598
+ "byteOffset": 12980224
2599
+ },
2600
+ {
2601
+ "name": "model.layers.9.self_attn.qkv_proj.q_scale",
2602
+ "shape": [
2603
+ 2560,
2604
+ 64
2605
+ ],
2606
+ "dtype": "float16",
2607
+ "format": "f32-to-bf16",
2608
+ "nbytes": 327680,
2609
+ "byteOffset": 15601664
2610
+ },
2611
+ {
2612
+ "name": "model.layers.9.self_attn.o_proj.q_weight",
2613
+ "shape": [
2614
+ 2048,
2615
+ 256
2616
+ ],
2617
+ "dtype": "uint32",
2618
+ "format": "f32-to-bf16",
2619
+ "nbytes": 2097152,
2620
+ "byteOffset": 15929344
2621
+ },
2622
+ {
2623
+ "name": "model.layers.9.self_attn.o_proj.q_scale",
2624
+ "shape": [
2625
+ 2048,
2626
+ 64
2627
+ ],
2628
+ "dtype": "float16",
2629
+ "format": "f32-to-bf16",
2630
+ "nbytes": 262144,
2631
+ "byteOffset": 18026496
2632
+ },
2633
+ {
2634
+ "name": "model.norm.weight",
2635
+ "shape": [
2636
+ 2048
2637
+ ],
2638
+ "dtype": "float16",
2639
+ "format": "f32-to-bf16",
2640
+ "nbytes": 4096,
2641
+ "byteOffset": 18288640
2642
+ }
2643
+ ],
2644
+ "md5sum": "415e1da793d4f918f3ffadcacad262bd"
2645
+ }
2646
+ ]
2647
+ }
params_shard_0.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0c57991ac672187e1478ff9749e61404d965ea2e2fd2dfe0c27db3f66fbcf88b
3
+ size 164847616
params_shard_1.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b7c4785bf0e3353bfd17619c100d1c32da565922abb811165ad7bc7618ac8c70
3
+ size 164847616
params_shard_10.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:60e4d11fcd1c0319bd44c417b24b7a2feefa7d2f48271ae84cce6ce059c9d190
3
+ size 24780800
params_shard_11.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d29080dcc924a0ce62520f043d8b46787c2485f065a071582990976e14454157
3
+ size 24780800
params_shard_12.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c8723d18161519a6b8879301d4571b2549f4a58fe24de9b5b8fd979b7f3b5ece
3
+ size 24780800
params_shard_13.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b91bfe928340fd9e1f3dc7cd292e697848bafaf3421406ee975a1af078841485
3
+ size 24780800
params_shard_14.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a9d73d4731268fdb07d2c2e86ba488696bc003cb5cf8c70f3bc3ff4fc059a2c3
3
+ size 24780800
params_shard_15.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:185360161280603bbf9caee4b198008998504b37c64e3435058224c0afab094d
3
+ size 24780800
params_shard_16.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:36feb3b6257881383f2a9d98d456764b06fad43c34581cf7a66a776e746aa862
3
+ size 24780800
params_shard_17.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6789ba827ee9a52e64e066e2b8c5822baa5b878c4be2d5c9c7c05ac56738e6eb
3
+ size 24780800
params_shard_18.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d463b5e3fa5f0b8439b9ac4e0b7dc3e166ae14c405d7070d0ba133717bb02c9c
3
+ size 24780800
params_shard_19.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:39f75ba56a0db8533c5076b66689807cb3e556005daff1208441cbc162c20a34
3
+ size 24780800
params_shard_2.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c9fa96cbedf01883ad078629a98339058133bc2dc00939449ea10e32d898e98e
3
+ size 20605952
params_shard_20.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4f06f14f741a91a9608353acd468a3006aa86bd2b6ad159df8335adbff47d15b
3
+ size 24780800
params_shard_21.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6dcfd4286ad463982b4beabaccb25190a13c7a03149b0bef815137c23be22706
3
+ size 24780800
params_shard_22.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c97913bbe9f9368f0d90f58dbf996da4b66f99c0a799f92fc25853476efc12e1
3
+ size 24780800
params_shard_23.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:22dc99cfd6f5c8e3ebd2addd01ba82e9d0b0bca00ba509fb67b7db40ff76cab9
3
+ size 24780800
params_shard_24.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8dc92486aa3069454d3aae93229705e343fc056c0ba34d892559ee260a95c627
3
+ size 24780800
params_shard_25.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1caab58200ca5faf183f334bef44fd4292331ae135734025c97b5cd028fbfbbc
3
+ size 18292736
params_shard_3.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d279c49ba84f7ed4341ec9c626121cf81206d0c511a00970b7bf3ef5eee0b46b
3
+ size 27098112
params_shard_4.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2d2448998b7124def847e812bbb27c20a81140681b79d264d5a50ffdaf0ac6c2
3
+ size 24780800
params_shard_5.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1ab100d117d3cbec9049c76caf7c2aa3571d4247290583e1a0e43561eb4be0a1
3
+ size 24780800
params_shard_6.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:660f3bf74e23ad8c99371ba2961ed132490a6a02242c97d0e814d7187917fd9f
3
+ size 24780800
params_shard_7.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d183e88d42d5a95227faa926ddb5c0b26fbdcb2eaf6d69a1800d901fbf215eba
3
+ size 24780800
params_shard_8.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0fd998838d2fcd6eabb044f11b9975a789ef1a7b107be1fd25a5af0ecb82c6e6
3
+ size 24780800
params_shard_9.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:de7c96b1626ecca87e5f6bbfe806403afc8e0ce2f4a328ad1345013feddd7e39
3
+ size 24780800
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff