Upload 23 files
Browse files- mlc-chat-config.json +77 -0
- ndarray-cache.json +639 -0
- params_shard_0.bin +3 -0
- params_shard_1.bin +3 -0
- params_shard_10.bin +3 -0
- params_shard_11.bin +3 -0
- params_shard_12.bin +3 -0
- params_shard_13.bin +3 -0
- params_shard_14.bin +3 -0
- params_shard_15.bin +3 -0
- params_shard_16.bin +3 -0
- params_shard_17.bin +3 -0
- params_shard_2.bin +3 -0
- params_shard_3.bin +3 -0
- params_shard_4.bin +3 -0
- params_shard_5.bin +3 -0
- params_shard_6.bin +3 -0
- params_shard_7.bin +3 -0
- params_shard_8.bin +3 -0
- params_shard_9.bin +3 -0
- tokenizer.json +0 -0
- tokenizer.model +3 -0
- tokenizer_config.json +35 -0
mlc-chat-config.json
ADDED
@@ -0,0 +1,77 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"model_type": "llama",
|
3 |
+
"quantization": "q4f16_1",
|
4 |
+
"model_config": {
|
5 |
+
"hidden_size": 4096,
|
6 |
+
"intermediate_size": 11008,
|
7 |
+
"num_attention_heads": 32,
|
8 |
+
"num_hidden_layers": 4,
|
9 |
+
"rms_norm_eps": 1e-05,
|
10 |
+
"vocab_size": 55296,
|
11 |
+
"position_embedding_base": 10000.0,
|
12 |
+
"context_window_size": 4096,
|
13 |
+
"prefill_chunk_size": 2048,
|
14 |
+
"num_key_value_heads": 32,
|
15 |
+
"head_dim": 128,
|
16 |
+
"tensor_parallel_shards": 1,
|
17 |
+
"max_batch_size": 80
|
18 |
+
},
|
19 |
+
"vocab_size": 55296,
|
20 |
+
"context_window_size": 4096,
|
21 |
+
"sliding_window_size": -1,
|
22 |
+
"prefill_chunk_size": 2048,
|
23 |
+
"attention_sink_size": -1,
|
24 |
+
"tensor_parallel_shards": 1,
|
25 |
+
"mean_gen_len": 128,
|
26 |
+
"max_gen_len": 512,
|
27 |
+
"shift_fill_factor": 0.3,
|
28 |
+
"temperature": 0.2,
|
29 |
+
"presence_penalty": 0.0,
|
30 |
+
"frequency_penalty": 0.0,
|
31 |
+
"repetition_penalty": 1.0,
|
32 |
+
"top_p": 0.9,
|
33 |
+
"conv_template": {
|
34 |
+
"name": "llama-2",
|
35 |
+
"system_template": "[INST] <<SYS>>\n{system_message}\n<</SYS>>\n\n",
|
36 |
+
"system_message": "You are a helpful, respectful and honest assistant.",
|
37 |
+
"system_prefix_token_ids": [
|
38 |
+
1
|
39 |
+
],
|
40 |
+
"add_role_after_system_message": false,
|
41 |
+
"roles": {
|
42 |
+
"user": "<s>[INST]",
|
43 |
+
"assistant": "[/INST]",
|
44 |
+
"tool": "[INST]"
|
45 |
+
},
|
46 |
+
"role_templates": {
|
47 |
+
"user": "{user_message}",
|
48 |
+
"assistant": "{assistant_message}",
|
49 |
+
"tool": "{tool_message}"
|
50 |
+
},
|
51 |
+
"messages": [],
|
52 |
+
"seps": [
|
53 |
+
" ",
|
54 |
+
" </s>"
|
55 |
+
],
|
56 |
+
"role_content_sep": " ",
|
57 |
+
"role_empty_sep": " ",
|
58 |
+
"stop_str": [
|
59 |
+
"[INST]"
|
60 |
+
],
|
61 |
+
"stop_token_ids": [
|
62 |
+
2
|
63 |
+
],
|
64 |
+
"function_string": "",
|
65 |
+
"use_function_calling": false
|
66 |
+
},
|
67 |
+
"pad_token_id": 0,
|
68 |
+
"bos_token_id": 1,
|
69 |
+
"eos_token_id": 2,
|
70 |
+
"tokenizer_files": [
|
71 |
+
"tokenizer.model",
|
72 |
+
"tokenizer.json",
|
73 |
+
"tokenizer_config.json"
|
74 |
+
],
|
75 |
+
"token_table_postproc_method": "byte_fallback",
|
76 |
+
"version": "0.1.0"
|
77 |
+
}
|
ndarray-cache.json
ADDED
@@ -0,0 +1,639 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"metadata": {
|
3 |
+
"ParamSize": 45,
|
4 |
+
"ParamBytes": 710221824.0,
|
5 |
+
"BitsPerParam": 4.500335784941911
|
6 |
+
},
|
7 |
+
"records": [
|
8 |
+
{
|
9 |
+
"dataPath": "params_shard_0.bin",
|
10 |
+
"format": "raw-shard",
|
11 |
+
"nbytes": 113246208,
|
12 |
+
"records": [
|
13 |
+
{
|
14 |
+
"name": "model.embed_tokens.q_weight",
|
15 |
+
"shape": [
|
16 |
+
55296,
|
17 |
+
512
|
18 |
+
],
|
19 |
+
"dtype": "uint32",
|
20 |
+
"format": "f32-to-bf16",
|
21 |
+
"nbytes": 113246208,
|
22 |
+
"byteOffset": 0
|
23 |
+
}
|
24 |
+
],
|
25 |
+
"md5sum": "8a6e3d99368ecc20bfedb66a4753a489"
|
26 |
+
},
|
27 |
+
{
|
28 |
+
"dataPath": "params_shard_1.bin",
|
29 |
+
"format": "raw-shard",
|
30 |
+
"nbytes": 25165824,
|
31 |
+
"records": [
|
32 |
+
{
|
33 |
+
"name": "model.layers.0.self_attn.qkv_proj.q_weight",
|
34 |
+
"shape": [
|
35 |
+
12288,
|
36 |
+
512
|
37 |
+
],
|
38 |
+
"dtype": "uint32",
|
39 |
+
"format": "f32-to-bf16",
|
40 |
+
"nbytes": 25165824,
|
41 |
+
"byteOffset": 0
|
42 |
+
}
|
43 |
+
],
|
44 |
+
"md5sum": "6dc6b5a86964a1b8eaf979f2bdb65c5f"
|
45 |
+
},
|
46 |
+
{
|
47 |
+
"dataPath": "params_shard_2.bin",
|
48 |
+
"format": "raw-shard",
|
49 |
+
"nbytes": 45088768,
|
50 |
+
"records": [
|
51 |
+
{
|
52 |
+
"name": "model.layers.0.mlp.gate_up_proj.q_weight",
|
53 |
+
"shape": [
|
54 |
+
22016,
|
55 |
+
512
|
56 |
+
],
|
57 |
+
"dtype": "uint32",
|
58 |
+
"format": "f32-to-bf16",
|
59 |
+
"nbytes": 45088768,
|
60 |
+
"byteOffset": 0
|
61 |
+
}
|
62 |
+
],
|
63 |
+
"md5sum": "f23c578d5279105a2475431b364d6082"
|
64 |
+
},
|
65 |
+
{
|
66 |
+
"dataPath": "params_shard_3.bin",
|
67 |
+
"format": "raw-shard",
|
68 |
+
"nbytes": 22544384,
|
69 |
+
"records": [
|
70 |
+
{
|
71 |
+
"name": "model.layers.0.mlp.down_proj.q_weight",
|
72 |
+
"shape": [
|
73 |
+
4096,
|
74 |
+
1376
|
75 |
+
],
|
76 |
+
"dtype": "uint32",
|
77 |
+
"format": "f32-to-bf16",
|
78 |
+
"nbytes": 22544384,
|
79 |
+
"byteOffset": 0
|
80 |
+
}
|
81 |
+
],
|
82 |
+
"md5sum": "de3af25993f69eee3d38d24398f34933"
|
83 |
+
},
|
84 |
+
{
|
85 |
+
"dataPath": "params_shard_4.bin",
|
86 |
+
"format": "raw-shard",
|
87 |
+
"nbytes": 32374784,
|
88 |
+
"records": [
|
89 |
+
{
|
90 |
+
"name": "model.embed_tokens.q_scale",
|
91 |
+
"shape": [
|
92 |
+
55296,
|
93 |
+
128
|
94 |
+
],
|
95 |
+
"dtype": "float16",
|
96 |
+
"format": "f32-to-bf16",
|
97 |
+
"nbytes": 14155776,
|
98 |
+
"byteOffset": 0
|
99 |
+
},
|
100 |
+
{
|
101 |
+
"name": "model.layers.0.self_attn.qkv_proj.q_scale",
|
102 |
+
"shape": [
|
103 |
+
12288,
|
104 |
+
128
|
105 |
+
],
|
106 |
+
"dtype": "float16",
|
107 |
+
"format": "f32-to-bf16",
|
108 |
+
"nbytes": 3145728,
|
109 |
+
"byteOffset": 14155776
|
110 |
+
},
|
111 |
+
{
|
112 |
+
"name": "model.layers.0.self_attn.o_proj.q_weight",
|
113 |
+
"shape": [
|
114 |
+
4096,
|
115 |
+
512
|
116 |
+
],
|
117 |
+
"dtype": "uint32",
|
118 |
+
"format": "f32-to-bf16",
|
119 |
+
"nbytes": 8388608,
|
120 |
+
"byteOffset": 17301504
|
121 |
+
},
|
122 |
+
{
|
123 |
+
"name": "model.layers.0.self_attn.o_proj.q_scale",
|
124 |
+
"shape": [
|
125 |
+
4096,
|
126 |
+
128
|
127 |
+
],
|
128 |
+
"dtype": "float16",
|
129 |
+
"format": "f32-to-bf16",
|
130 |
+
"nbytes": 1048576,
|
131 |
+
"byteOffset": 25690112
|
132 |
+
},
|
133 |
+
{
|
134 |
+
"name": "model.layers.0.mlp.gate_up_proj.q_scale",
|
135 |
+
"shape": [
|
136 |
+
22016,
|
137 |
+
128
|
138 |
+
],
|
139 |
+
"dtype": "float16",
|
140 |
+
"format": "f32-to-bf16",
|
141 |
+
"nbytes": 5636096,
|
142 |
+
"byteOffset": 26738688
|
143 |
+
}
|
144 |
+
],
|
145 |
+
"md5sum": "4bbe7d5de8b78ecaca0e56086a595d2b"
|
146 |
+
},
|
147 |
+
{
|
148 |
+
"dataPath": "params_shard_5.bin",
|
149 |
+
"format": "raw-shard",
|
150 |
+
"nbytes": 31145984,
|
151 |
+
"records": [
|
152 |
+
{
|
153 |
+
"name": "model.layers.0.mlp.down_proj.q_scale",
|
154 |
+
"shape": [
|
155 |
+
4096,
|
156 |
+
344
|
157 |
+
],
|
158 |
+
"dtype": "float16",
|
159 |
+
"format": "f32-to-bf16",
|
160 |
+
"nbytes": 2818048,
|
161 |
+
"byteOffset": 0
|
162 |
+
},
|
163 |
+
{
|
164 |
+
"name": "model.layers.0.input_layernorm.weight",
|
165 |
+
"shape": [
|
166 |
+
4096
|
167 |
+
],
|
168 |
+
"dtype": "float16",
|
169 |
+
"format": "f32-to-bf16",
|
170 |
+
"nbytes": 8192,
|
171 |
+
"byteOffset": 2818048
|
172 |
+
},
|
173 |
+
{
|
174 |
+
"name": "model.layers.0.post_attention_layernorm.weight",
|
175 |
+
"shape": [
|
176 |
+
4096
|
177 |
+
],
|
178 |
+
"dtype": "float16",
|
179 |
+
"format": "f32-to-bf16",
|
180 |
+
"nbytes": 8192,
|
181 |
+
"byteOffset": 2826240
|
182 |
+
},
|
183 |
+
{
|
184 |
+
"name": "model.layers.1.self_attn.qkv_proj.q_weight",
|
185 |
+
"shape": [
|
186 |
+
12288,
|
187 |
+
512
|
188 |
+
],
|
189 |
+
"dtype": "uint32",
|
190 |
+
"format": "f32-to-bf16",
|
191 |
+
"nbytes": 25165824,
|
192 |
+
"byteOffset": 2834432
|
193 |
+
},
|
194 |
+
{
|
195 |
+
"name": "model.layers.1.self_attn.qkv_proj.q_scale",
|
196 |
+
"shape": [
|
197 |
+
12288,
|
198 |
+
128
|
199 |
+
],
|
200 |
+
"dtype": "float16",
|
201 |
+
"format": "f32-to-bf16",
|
202 |
+
"nbytes": 3145728,
|
203 |
+
"byteOffset": 28000256
|
204 |
+
}
|
205 |
+
],
|
206 |
+
"md5sum": "f188574cefeead2adf6866163c151917"
|
207 |
+
},
|
208 |
+
{
|
209 |
+
"dataPath": "params_shard_6.bin",
|
210 |
+
"format": "raw-shard",
|
211 |
+
"nbytes": 45088768,
|
212 |
+
"records": [
|
213 |
+
{
|
214 |
+
"name": "model.layers.1.mlp.gate_up_proj.q_weight",
|
215 |
+
"shape": [
|
216 |
+
22016,
|
217 |
+
512
|
218 |
+
],
|
219 |
+
"dtype": "uint32",
|
220 |
+
"format": "f32-to-bf16",
|
221 |
+
"nbytes": 45088768,
|
222 |
+
"byteOffset": 0
|
223 |
+
}
|
224 |
+
],
|
225 |
+
"md5sum": "f419bbcd8f931798727fd5963f7b76a9"
|
226 |
+
},
|
227 |
+
{
|
228 |
+
"dataPath": "params_shard_7.bin",
|
229 |
+
"format": "raw-shard",
|
230 |
+
"nbytes": 22544384,
|
231 |
+
"records": [
|
232 |
+
{
|
233 |
+
"name": "model.layers.1.mlp.down_proj.q_weight",
|
234 |
+
"shape": [
|
235 |
+
4096,
|
236 |
+
1376
|
237 |
+
],
|
238 |
+
"dtype": "uint32",
|
239 |
+
"format": "f32-to-bf16",
|
240 |
+
"nbytes": 22544384,
|
241 |
+
"byteOffset": 0
|
242 |
+
}
|
243 |
+
],
|
244 |
+
"md5sum": "1ed071ab5bc7d7ed093e56bdcfc0dfc1"
|
245 |
+
},
|
246 |
+
{
|
247 |
+
"dataPath": "params_shard_8.bin",
|
248 |
+
"format": "raw-shard",
|
249 |
+
"nbytes": 25165824,
|
250 |
+
"records": [
|
251 |
+
{
|
252 |
+
"name": "model.layers.2.self_attn.qkv_proj.q_weight",
|
253 |
+
"shape": [
|
254 |
+
12288,
|
255 |
+
512
|
256 |
+
],
|
257 |
+
"dtype": "uint32",
|
258 |
+
"format": "f32-to-bf16",
|
259 |
+
"nbytes": 25165824,
|
260 |
+
"byteOffset": 0
|
261 |
+
}
|
262 |
+
],
|
263 |
+
"md5sum": "64517f2b3f4c35296dd9711cb6dd8748"
|
264 |
+
},
|
265 |
+
{
|
266 |
+
"dataPath": "params_shard_9.bin",
|
267 |
+
"format": "raw-shard",
|
268 |
+
"nbytes": 45088768,
|
269 |
+
"records": [
|
270 |
+
{
|
271 |
+
"name": "model.layers.2.mlp.gate_up_proj.q_weight",
|
272 |
+
"shape": [
|
273 |
+
22016,
|
274 |
+
512
|
275 |
+
],
|
276 |
+
"dtype": "uint32",
|
277 |
+
"format": "f32-to-bf16",
|
278 |
+
"nbytes": 45088768,
|
279 |
+
"byteOffset": 0
|
280 |
+
}
|
281 |
+
],
|
282 |
+
"md5sum": "f3449ee6e8be73f2f04a55eb1354eed7"
|
283 |
+
},
|
284 |
+
{
|
285 |
+
"dataPath": "params_shard_10.bin",
|
286 |
+
"format": "raw-shard",
|
287 |
+
"nbytes": 30490624,
|
288 |
+
"records": [
|
289 |
+
{
|
290 |
+
"name": "model.layers.1.self_attn.o_proj.q_weight",
|
291 |
+
"shape": [
|
292 |
+
4096,
|
293 |
+
512
|
294 |
+
],
|
295 |
+
"dtype": "uint32",
|
296 |
+
"format": "f32-to-bf16",
|
297 |
+
"nbytes": 8388608,
|
298 |
+
"byteOffset": 0
|
299 |
+
},
|
300 |
+
{
|
301 |
+
"name": "model.layers.1.self_attn.o_proj.q_scale",
|
302 |
+
"shape": [
|
303 |
+
4096,
|
304 |
+
128
|
305 |
+
],
|
306 |
+
"dtype": "float16",
|
307 |
+
"format": "f32-to-bf16",
|
308 |
+
"nbytes": 1048576,
|
309 |
+
"byteOffset": 8388608
|
310 |
+
},
|
311 |
+
{
|
312 |
+
"name": "model.layers.1.mlp.gate_up_proj.q_scale",
|
313 |
+
"shape": [
|
314 |
+
22016,
|
315 |
+
128
|
316 |
+
],
|
317 |
+
"dtype": "float16",
|
318 |
+
"format": "f32-to-bf16",
|
319 |
+
"nbytes": 5636096,
|
320 |
+
"byteOffset": 9437184
|
321 |
+
},
|
322 |
+
{
|
323 |
+
"name": "model.layers.1.mlp.down_proj.q_scale",
|
324 |
+
"shape": [
|
325 |
+
4096,
|
326 |
+
344
|
327 |
+
],
|
328 |
+
"dtype": "float16",
|
329 |
+
"format": "f32-to-bf16",
|
330 |
+
"nbytes": 2818048,
|
331 |
+
"byteOffset": 15073280
|
332 |
+
},
|
333 |
+
{
|
334 |
+
"name": "model.layers.1.input_layernorm.weight",
|
335 |
+
"shape": [
|
336 |
+
4096
|
337 |
+
],
|
338 |
+
"dtype": "float16",
|
339 |
+
"format": "f32-to-bf16",
|
340 |
+
"nbytes": 8192,
|
341 |
+
"byteOffset": 17891328
|
342 |
+
},
|
343 |
+
{
|
344 |
+
"name": "model.layers.1.post_attention_layernorm.weight",
|
345 |
+
"shape": [
|
346 |
+
4096
|
347 |
+
],
|
348 |
+
"dtype": "float16",
|
349 |
+
"format": "f32-to-bf16",
|
350 |
+
"nbytes": 8192,
|
351 |
+
"byteOffset": 17899520
|
352 |
+
},
|
353 |
+
{
|
354 |
+
"name": "model.layers.2.self_attn.qkv_proj.q_scale",
|
355 |
+
"shape": [
|
356 |
+
12288,
|
357 |
+
128
|
358 |
+
],
|
359 |
+
"dtype": "float16",
|
360 |
+
"format": "f32-to-bf16",
|
361 |
+
"nbytes": 3145728,
|
362 |
+
"byteOffset": 17907712
|
363 |
+
},
|
364 |
+
{
|
365 |
+
"name": "model.layers.2.self_attn.o_proj.q_weight",
|
366 |
+
"shape": [
|
367 |
+
4096,
|
368 |
+
512
|
369 |
+
],
|
370 |
+
"dtype": "uint32",
|
371 |
+
"format": "f32-to-bf16",
|
372 |
+
"nbytes": 8388608,
|
373 |
+
"byteOffset": 21053440
|
374 |
+
},
|
375 |
+
{
|
376 |
+
"name": "model.layers.2.self_attn.o_proj.q_scale",
|
377 |
+
"shape": [
|
378 |
+
4096,
|
379 |
+
128
|
380 |
+
],
|
381 |
+
"dtype": "float16",
|
382 |
+
"format": "f32-to-bf16",
|
383 |
+
"nbytes": 1048576,
|
384 |
+
"byteOffset": 29442048
|
385 |
+
}
|
386 |
+
],
|
387 |
+
"md5sum": "5b17842481ec0700776b00984d62fdd5"
|
388 |
+
},
|
389 |
+
{
|
390 |
+
"dataPath": "params_shard_11.bin",
|
391 |
+
"format": "raw-shard",
|
392 |
+
"nbytes": 25165824,
|
393 |
+
"records": [
|
394 |
+
{
|
395 |
+
"name": "model.layers.3.self_attn.qkv_proj.q_weight",
|
396 |
+
"shape": [
|
397 |
+
12288,
|
398 |
+
512
|
399 |
+
],
|
400 |
+
"dtype": "uint32",
|
401 |
+
"format": "f32-to-bf16",
|
402 |
+
"nbytes": 25165824,
|
403 |
+
"byteOffset": 0
|
404 |
+
}
|
405 |
+
],
|
406 |
+
"md5sum": "81dda134f78fe8bc32eb9333ffb489dd"
|
407 |
+
},
|
408 |
+
{
|
409 |
+
"dataPath": "params_shard_12.bin",
|
410 |
+
"format": "raw-shard",
|
411 |
+
"nbytes": 31014912,
|
412 |
+
"records": [
|
413 |
+
{
|
414 |
+
"name": "model.layers.2.mlp.gate_up_proj.q_scale",
|
415 |
+
"shape": [
|
416 |
+
22016,
|
417 |
+
128
|
418 |
+
],
|
419 |
+
"dtype": "float16",
|
420 |
+
"format": "f32-to-bf16",
|
421 |
+
"nbytes": 5636096,
|
422 |
+
"byteOffset": 0
|
423 |
+
},
|
424 |
+
{
|
425 |
+
"name": "model.layers.2.mlp.down_proj.q_weight",
|
426 |
+
"shape": [
|
427 |
+
4096,
|
428 |
+
1376
|
429 |
+
],
|
430 |
+
"dtype": "uint32",
|
431 |
+
"format": "f32-to-bf16",
|
432 |
+
"nbytes": 22544384,
|
433 |
+
"byteOffset": 5636096
|
434 |
+
},
|
435 |
+
{
|
436 |
+
"name": "model.layers.2.mlp.down_proj.q_scale",
|
437 |
+
"shape": [
|
438 |
+
4096,
|
439 |
+
344
|
440 |
+
],
|
441 |
+
"dtype": "float16",
|
442 |
+
"format": "f32-to-bf16",
|
443 |
+
"nbytes": 2818048,
|
444 |
+
"byteOffset": 28180480
|
445 |
+
},
|
446 |
+
{
|
447 |
+
"name": "model.layers.2.input_layernorm.weight",
|
448 |
+
"shape": [
|
449 |
+
4096
|
450 |
+
],
|
451 |
+
"dtype": "float16",
|
452 |
+
"format": "f32-to-bf16",
|
453 |
+
"nbytes": 8192,
|
454 |
+
"byteOffset": 30998528
|
455 |
+
},
|
456 |
+
{
|
457 |
+
"name": "model.layers.2.post_attention_layernorm.weight",
|
458 |
+
"shape": [
|
459 |
+
4096
|
460 |
+
],
|
461 |
+
"dtype": "float16",
|
462 |
+
"format": "f32-to-bf16",
|
463 |
+
"nbytes": 8192,
|
464 |
+
"byteOffset": 31006720
|
465 |
+
}
|
466 |
+
],
|
467 |
+
"md5sum": "59366ee4fb0104a9d93e7a45c7a1a27a"
|
468 |
+
},
|
469 |
+
{
|
470 |
+
"dataPath": "params_shard_13.bin",
|
471 |
+
"format": "raw-shard",
|
472 |
+
"nbytes": 45088768,
|
473 |
+
"records": [
|
474 |
+
{
|
475 |
+
"name": "model.layers.3.mlp.gate_up_proj.q_weight",
|
476 |
+
"shape": [
|
477 |
+
22016,
|
478 |
+
512
|
479 |
+
],
|
480 |
+
"dtype": "uint32",
|
481 |
+
"format": "f32-to-bf16",
|
482 |
+
"nbytes": 45088768,
|
483 |
+
"byteOffset": 0
|
484 |
+
}
|
485 |
+
],
|
486 |
+
"md5sum": "ccb83f6c6dac4773793fdc5c474cc8c2"
|
487 |
+
},
|
488 |
+
{
|
489 |
+
"dataPath": "params_shard_14.bin",
|
490 |
+
"format": "raw-shard",
|
491 |
+
"nbytes": 22544384,
|
492 |
+
"records": [
|
493 |
+
{
|
494 |
+
"name": "model.layers.3.mlp.down_proj.q_weight",
|
495 |
+
"shape": [
|
496 |
+
4096,
|
497 |
+
1376
|
498 |
+
],
|
499 |
+
"dtype": "uint32",
|
500 |
+
"format": "f32-to-bf16",
|
501 |
+
"nbytes": 22544384,
|
502 |
+
"byteOffset": 0
|
503 |
+
}
|
504 |
+
],
|
505 |
+
"md5sum": "88a2bbad50d90a1e27ba7d1490d75aee"
|
506 |
+
},
|
507 |
+
{
|
508 |
+
"dataPath": "params_shard_15.bin",
|
509 |
+
"format": "raw-shard",
|
510 |
+
"nbytes": 113246208,
|
511 |
+
"records": [
|
512 |
+
{
|
513 |
+
"name": "lm_head.q_weight",
|
514 |
+
"shape": [
|
515 |
+
55296,
|
516 |
+
512
|
517 |
+
],
|
518 |
+
"dtype": "uint32",
|
519 |
+
"format": "f32-to-bf16",
|
520 |
+
"nbytes": 113246208,
|
521 |
+
"byteOffset": 0
|
522 |
+
}
|
523 |
+
],
|
524 |
+
"md5sum": "576ee2120b5bcefc80df4416726393ce"
|
525 |
+
},
|
526 |
+
{
|
527 |
+
"dataPath": "params_shard_16.bin",
|
528 |
+
"format": "raw-shard",
|
529 |
+
"nbytes": 21061632,
|
530 |
+
"records": [
|
531 |
+
{
|
532 |
+
"name": "model.layers.3.self_attn.qkv_proj.q_scale",
|
533 |
+
"shape": [
|
534 |
+
12288,
|
535 |
+
128
|
536 |
+
],
|
537 |
+
"dtype": "float16",
|
538 |
+
"format": "f32-to-bf16",
|
539 |
+
"nbytes": 3145728,
|
540 |
+
"byteOffset": 0
|
541 |
+
},
|
542 |
+
{
|
543 |
+
"name": "model.layers.3.self_attn.o_proj.q_weight",
|
544 |
+
"shape": [
|
545 |
+
4096,
|
546 |
+
512
|
547 |
+
],
|
548 |
+
"dtype": "uint32",
|
549 |
+
"format": "f32-to-bf16",
|
550 |
+
"nbytes": 8388608,
|
551 |
+
"byteOffset": 3145728
|
552 |
+
},
|
553 |
+
{
|
554 |
+
"name": "model.layers.3.self_attn.o_proj.q_scale",
|
555 |
+
"shape": [
|
556 |
+
4096,
|
557 |
+
128
|
558 |
+
],
|
559 |
+
"dtype": "float16",
|
560 |
+
"format": "f32-to-bf16",
|
561 |
+
"nbytes": 1048576,
|
562 |
+
"byteOffset": 11534336
|
563 |
+
},
|
564 |
+
{
|
565 |
+
"name": "model.layers.3.mlp.gate_up_proj.q_scale",
|
566 |
+
"shape": [
|
567 |
+
22016,
|
568 |
+
128
|
569 |
+
],
|
570 |
+
"dtype": "float16",
|
571 |
+
"format": "f32-to-bf16",
|
572 |
+
"nbytes": 5636096,
|
573 |
+
"byteOffset": 12582912
|
574 |
+
},
|
575 |
+
{
|
576 |
+
"name": "model.layers.3.mlp.down_proj.q_scale",
|
577 |
+
"shape": [
|
578 |
+
4096,
|
579 |
+
344
|
580 |
+
],
|
581 |
+
"dtype": "float16",
|
582 |
+
"format": "f32-to-bf16",
|
583 |
+
"nbytes": 2818048,
|
584 |
+
"byteOffset": 18219008
|
585 |
+
},
|
586 |
+
{
|
587 |
+
"name": "model.layers.3.input_layernorm.weight",
|
588 |
+
"shape": [
|
589 |
+
4096
|
590 |
+
],
|
591 |
+
"dtype": "float16",
|
592 |
+
"format": "f32-to-bf16",
|
593 |
+
"nbytes": 8192,
|
594 |
+
"byteOffset": 21037056
|
595 |
+
},
|
596 |
+
{
|
597 |
+
"name": "model.layers.3.post_attention_layernorm.weight",
|
598 |
+
"shape": [
|
599 |
+
4096
|
600 |
+
],
|
601 |
+
"dtype": "float16",
|
602 |
+
"format": "f32-to-bf16",
|
603 |
+
"nbytes": 8192,
|
604 |
+
"byteOffset": 21045248
|
605 |
+
},
|
606 |
+
{
|
607 |
+
"name": "model.norm.weight",
|
608 |
+
"shape": [
|
609 |
+
4096
|
610 |
+
],
|
611 |
+
"dtype": "float16",
|
612 |
+
"format": "f32-to-bf16",
|
613 |
+
"nbytes": 8192,
|
614 |
+
"byteOffset": 21053440
|
615 |
+
}
|
616 |
+
],
|
617 |
+
"md5sum": "51bd241eece6fa67f1a4bd0dde4d5fff"
|
618 |
+
},
|
619 |
+
{
|
620 |
+
"dataPath": "params_shard_17.bin",
|
621 |
+
"format": "raw-shard",
|
622 |
+
"nbytes": 14155776,
|
623 |
+
"records": [
|
624 |
+
{
|
625 |
+
"name": "lm_head.q_scale",
|
626 |
+
"shape": [
|
627 |
+
55296,
|
628 |
+
128
|
629 |
+
],
|
630 |
+
"dtype": "float16",
|
631 |
+
"format": "f32-to-bf16",
|
632 |
+
"nbytes": 14155776,
|
633 |
+
"byteOffset": 0
|
634 |
+
}
|
635 |
+
],
|
636 |
+
"md5sum": "dc48e203ccd2de292325004b14806dd3"
|
637 |
+
}
|
638 |
+
]
|
639 |
+
}
|
params_shard_0.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:12a8cc38fcc2589fd889e8ecdd10654c9be0510fbdb83e1c4a233597ffcba4bc
|
3 |
+
size 113246208
|
params_shard_1.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:3b4edc2221b07ebd721d2b27381bcaa7ed8059dd3ac85b5513af97e9ff5595ba
|
3 |
+
size 25165824
|
params_shard_10.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:fd0b95ae347c76c1454a393253c09d30ce1202efd3b2c29c8ca130f2b113aef2
|
3 |
+
size 30490624
|
params_shard_11.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:fd6374d715b5a92a93b1571e6edddf117a0eb24e0eb63ec79932de042800db23
|
3 |
+
size 25165824
|
params_shard_12.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:d455875abb27ff44ed4baef4291be300c6db3f47b5d5d5133f41691b876b7f37
|
3 |
+
size 31014912
|
params_shard_13.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:ba0d5148c6e7d862a123bd400b56b7dc292ec5a8ad3c665ae3a4198027792026
|
3 |
+
size 45088768
|
params_shard_14.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:2d60bb747bee9b5fd0becbf139037d8113f2a2d251b64188f948197121b0d259
|
3 |
+
size 22544384
|
params_shard_15.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:13683d7ea0ecd4e3f3ff72fd219d828e3096b4a57c6b1781ddd03d50b43dd0d7
|
3 |
+
size 113246208
|
params_shard_16.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:ec55c25703b4aec39831fe996222f3fc36b1f9eebc187e4e33c190768c2a2886
|
3 |
+
size 21061632
|
params_shard_17.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:a52192ac0f31f329a761aa7977d0efd8b5587e098a0b51eadcb0ad9c0678dd7f
|
3 |
+
size 14155776
|
params_shard_2.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:4c68eda0b375a7ca086bd9d3f9d3bd0a708e1f47b31ea684a24f1cb9061f4e8c
|
3 |
+
size 45088768
|
params_shard_3.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:43b408a53f6744f067dde7ad1a9387c8bccbda2ee1b91775bac77fc7bd7c7846
|
3 |
+
size 22544384
|
params_shard_4.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f28956782a37313e2fd0cfe5b61716ba1f38368fed767cdf454dee838152f87f
|
3 |
+
size 32374784
|
params_shard_5.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:857752056bd3b48cd88685fa210c284f07ea124b1d561713eaf8b701a6422656
|
3 |
+
size 31145984
|
params_shard_6.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:964e937791d49589b084ef869e9377150ab53cdfea8130c7dd374c02af843058
|
3 |
+
size 45088768
|
params_shard_7.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:e746632367a8469572b3dc9d21834d76a9ee55e422e027a09ebddf1ca4d48a66
|
3 |
+
size 22544384
|
params_shard_8.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:2731b2d5ba1b3d737d21f3bb6f36eeb575cd9e0a190f4381bd06baaa7216ed4a
|
3 |
+
size 25165824
|
params_shard_9.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:aa239d3082116dd776c68f050bb217aac21095da71915522283f0f6b50898400
|
3 |
+
size 45088768
|
tokenizer.json
ADDED
The diff for this file is too large to render.
See raw diff
|
|
tokenizer.model
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:a3b8844863b200dfcca971db228e96ce388290dfcf72c15d7a9d2f604bac787c
|
3 |
+
size 844403
|
tokenizer_config.json
ADDED
@@ -0,0 +1,35 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"add_bos_token": true,
|
3 |
+
"add_eos_token": false,
|
4 |
+
"bos_token": {
|
5 |
+
"__type": "AddedToken",
|
6 |
+
"content": "<s>",
|
7 |
+
"lstrip": false,
|
8 |
+
"normalized": true,
|
9 |
+
"rstrip": false,
|
10 |
+
"single_word": false
|
11 |
+
},
|
12 |
+
"clean_up_tokenization_spaces": false,
|
13 |
+
"eos_token": {
|
14 |
+
"__type": "AddedToken",
|
15 |
+
"content": "</s>",
|
16 |
+
"lstrip": false,
|
17 |
+
"normalized": true,
|
18 |
+
"rstrip": false,
|
19 |
+
"single_word": false
|
20 |
+
},
|
21 |
+
"legacy": true,
|
22 |
+
"model_max_length": 1000000000000000019884624838656,
|
23 |
+
"pad_token": null,
|
24 |
+
"sp_model_kwargs": {},
|
25 |
+
"tokenizer_class": "LlamaTokenizer",
|
26 |
+
"unk_token": {
|
27 |
+
"__type": "AddedToken",
|
28 |
+
"content": "<unk>",
|
29 |
+
"lstrip": false,
|
30 |
+
"normalized": true,
|
31 |
+
"rstrip": false,
|
32 |
+
"single_word": false
|
33 |
+
},
|
34 |
+
"use_fast": false
|
35 |
+
}
|