Hugging Face's logo

Qwen
/
Qwen1.5-MoE-A2.7B-Chat-GPTQ-Int4

Text Generation
Transformers
Safetensors
English
qwen2_moe
chat
conversational
Inference Endpoints
4-bit precision
gptq
Model card Files Community
2
bzheng commited on
Commit
d1e19c3
•
1 Parent(s): 207fa13

update models & configs.

Browse files
Files changed (10) hide show
  1. config.json +259 -0
  2. generation_config.json +14 -0
  3. merges.txt +0 -0
  4. model-00001-of-00003.safetensors +3 -0
  5. model-00002-of-00003.safetensors +3 -0
  6. model-00003-of-00003.safetensors +3 -0
  7. model.safetensors.index.json +0 -0
  8. tokenizer.json +0 -0
  9. tokenizer_config.json +41 -0
  10. vocab.json +0 -0
config.json ADDED
@@ -0,0 +1,259 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "Qwen2MoeForCausalLM"
4
+ ],
5
+ "attention_dropout": 0.0,
6
+ "bos_token_id": 151643,
7
+ "eos_token_id": 151643,
8
+ "expert_interval": 1,
9
+ "hidden_act": "silu",
10
+ "hidden_size": 2048,
11
+ "initializer_range": 0.02,
12
+ "intermediate_size": 5632,
13
+ "max_position_embeddings": 32768,
14
+ "max_window_layers": 21,
15
+ "model_type": "qwen2_moe",
16
+ "moe_intermediate_size": 1408,
17
+ "norm_topk_prob": false,
18
+ "num_attention_heads": 16,
19
+ "num_experts": 60,
20
+ "num_experts_per_tok": 4,
21
+ "num_hidden_layers": 24,
22
+ "num_key_value_heads": 16,
23
+ "output_router_logits": false,
24
+ "quantization_config": {
25
+ "batch_size": 1,
26
+ "bits": 4,
27
+ "block_name_to_quantize": null,
28
+ "cache_block_outputs": true,
29
+ "damp_percent": 0.01,
30
+ "dataset": null,
31
+ "desc_act": false,
32
+ "exllama_config": {
33
+ "version": 1
34
+ },
35
+ "group_size": 128,
36
+ "max_input_length": null,
37
+ "model_seqlen": null,
38
+ "module_name_preceding_first_block": null,
39
+ "modules_in_block_to_quantize": [
40
+ [
41
+ "self_attn.k_proj",
42
+ "self_attn.v_proj",
43
+ "self_attn.q_proj"
44
+ ],
45
+ [
46
+ "self_attn.o_proj"
47
+ ],
48
+ [
49
+ "mlp.shared_expert.up_proj",
50
+ "mlp.shared_expert.gate_proj"
51
+ ],
52
+ [
53
+ "mlp.shared_expert.down_proj"
54
+ ],
55
+ [
56
+ "mlp.experts.0.up_proj",
57
+ "mlp.experts.1.up_proj",
58
+ "mlp.experts.2.up_proj",
59
+ "mlp.experts.3.up_proj",
60
+ "mlp.experts.4.up_proj",
61
+ "mlp.experts.5.up_proj",
62
+ "mlp.experts.6.up_proj",
63
+ "mlp.experts.7.up_proj",
64
+ "mlp.experts.8.up_proj",
65
+ "mlp.experts.9.up_proj",
66
+ "mlp.experts.10.up_proj",
67
+ "mlp.experts.11.up_proj",
68
+ "mlp.experts.12.up_proj",
69
+ "mlp.experts.13.up_proj",
70
+ "mlp.experts.14.up_proj",
71
+ "mlp.experts.15.up_proj",
72
+ "mlp.experts.16.up_proj",
73
+ "mlp.experts.17.up_proj",
74
+ "mlp.experts.18.up_proj",
75
+ "mlp.experts.19.up_proj",
76
+ "mlp.experts.20.up_proj",
77
+ "mlp.experts.21.up_proj",
78
+ "mlp.experts.22.up_proj",
79
+ "mlp.experts.23.up_proj",
80
+ "mlp.experts.24.up_proj",
81
+ "mlp.experts.25.up_proj",
82
+ "mlp.experts.26.up_proj",
83
+ "mlp.experts.27.up_proj",
84
+ "mlp.experts.28.up_proj",
85
+ "mlp.experts.29.up_proj",
86
+ "mlp.experts.30.up_proj",
87
+ "mlp.experts.31.up_proj",
88
+ "mlp.experts.32.up_proj",
89
+ "mlp.experts.33.up_proj",
90
+ "mlp.experts.34.up_proj",
91
+ "mlp.experts.35.up_proj",
92
+ "mlp.experts.36.up_proj",
93
+ "mlp.experts.37.up_proj",
94
+ "mlp.experts.38.up_proj",
95
+ "mlp.experts.39.up_proj",
96
+ "mlp.experts.40.up_proj",
97
+ "mlp.experts.41.up_proj",
98
+ "mlp.experts.42.up_proj",
99
+ "mlp.experts.43.up_proj",
100
+ "mlp.experts.44.up_proj",
101
+ "mlp.experts.45.up_proj",
102
+ "mlp.experts.46.up_proj",
103
+ "mlp.experts.47.up_proj",
104
+ "mlp.experts.48.up_proj",
105
+ "mlp.experts.49.up_proj",
106
+ "mlp.experts.50.up_proj",
107
+ "mlp.experts.51.up_proj",
108
+ "mlp.experts.52.up_proj",
109
+ "mlp.experts.53.up_proj",
110
+ "mlp.experts.54.up_proj",
111
+ "mlp.experts.55.up_proj",
112
+ "mlp.experts.56.up_proj",
113
+ "mlp.experts.57.up_proj",
114
+ "mlp.experts.58.up_proj",
115
+ "mlp.experts.59.up_proj",
116
+ "mlp.experts.0.gate_proj",
117
+ "mlp.experts.1.gate_proj",
118
+ "mlp.experts.2.gate_proj",
119
+ "mlp.experts.3.gate_proj",
120
+ "mlp.experts.4.gate_proj",
121
+ "mlp.experts.5.gate_proj",
122
+ "mlp.experts.6.gate_proj",
123
+ "mlp.experts.7.gate_proj",
124
+ "mlp.experts.8.gate_proj",
125
+ "mlp.experts.9.gate_proj",
126
+ "mlp.experts.10.gate_proj",
127
+ "mlp.experts.11.gate_proj",
128
+ "mlp.experts.12.gate_proj",
129
+ "mlp.experts.13.gate_proj",
130
+ "mlp.experts.14.gate_proj",
131
+ "mlp.experts.15.gate_proj",
132
+ "mlp.experts.16.gate_proj",
133
+ "mlp.experts.17.gate_proj",
134
+ "mlp.experts.18.gate_proj",
135
+ "mlp.experts.19.gate_proj",
136
+ "mlp.experts.20.gate_proj",
137
+ "mlp.experts.21.gate_proj",
138
+ "mlp.experts.22.gate_proj",
139
+ "mlp.experts.23.gate_proj",
140
+ "mlp.experts.24.gate_proj",
141
+ "mlp.experts.25.gate_proj",
142
+ "mlp.experts.26.gate_proj",
143
+ "mlp.experts.27.gate_proj",
144
+ "mlp.experts.28.gate_proj",
145
+ "mlp.experts.29.gate_proj",
146
+ "mlp.experts.30.gate_proj",
147
+ "mlp.experts.31.gate_proj",
148
+ "mlp.experts.32.gate_proj",
149
+ "mlp.experts.33.gate_proj",
150
+ "mlp.experts.34.gate_proj",
151
+ "mlp.experts.35.gate_proj",
152
+ "mlp.experts.36.gate_proj",
153
+ "mlp.experts.37.gate_proj",
154
+ "mlp.experts.38.gate_proj",
155
+ "mlp.experts.39.gate_proj",
156
+ "mlp.experts.40.gate_proj",
157
+ "mlp.experts.41.gate_proj",
158
+ "mlp.experts.42.gate_proj",
159
+ "mlp.experts.43.gate_proj",
160
+ "mlp.experts.44.gate_proj",
161
+ "mlp.experts.45.gate_proj",
162
+ "mlp.experts.46.gate_proj",
163
+ "mlp.experts.47.gate_proj",
164
+ "mlp.experts.48.gate_proj",
165
+ "mlp.experts.49.gate_proj",
166
+ "mlp.experts.50.gate_proj",
167
+ "mlp.experts.51.gate_proj",
168
+ "mlp.experts.52.gate_proj",
169
+ "mlp.experts.53.gate_proj",
170
+ "mlp.experts.54.gate_proj",
171
+ "mlp.experts.55.gate_proj",
172
+ "mlp.experts.56.gate_proj",
173
+ "mlp.experts.57.gate_proj",
174
+ "mlp.experts.58.gate_proj",
175
+ "mlp.experts.59.gate_proj"
176
+ ],
177
+ [
178
+ "mlp.experts.0.down_proj",
179
+ "mlp.experts.1.down_proj",
180
+ "mlp.experts.2.down_proj",
181
+ "mlp.experts.3.down_proj",
182
+ "mlp.experts.4.down_proj",
183
+ "mlp.experts.5.down_proj",
184
+ "mlp.experts.6.down_proj",
185
+ "mlp.experts.7.down_proj",
186
+ "mlp.experts.8.down_proj",
187
+ "mlp.experts.9.down_proj",
188
+ "mlp.experts.10.down_proj",
189
+ "mlp.experts.11.down_proj",
190
+ "mlp.experts.12.down_proj",
191
+ "mlp.experts.13.down_proj",
192
+ "mlp.experts.14.down_proj",
193
+ "mlp.experts.15.down_proj",
194
+ "mlp.experts.16.down_proj",
195
+ "mlp.experts.17.down_proj",
196
+ "mlp.experts.18.down_proj",
197
+ "mlp.experts.19.down_proj",
198
+ "mlp.experts.20.down_proj",
199
+ "mlp.experts.21.down_proj",
200
+ "mlp.experts.22.down_proj",
201
+ "mlp.experts.23.down_proj",
202
+ "mlp.experts.24.down_proj",
203
+ "mlp.experts.25.down_proj",
204
+ "mlp.experts.26.down_proj",
205
+ "mlp.experts.27.down_proj",
206
+ "mlp.experts.28.down_proj",
207
+ "mlp.experts.29.down_proj",
208
+ "mlp.experts.30.down_proj",
209
+ "mlp.experts.31.down_proj",
210
+ "mlp.experts.32.down_proj",
211
+ "mlp.experts.33.down_proj",
212
+ "mlp.experts.34.down_proj",
213
+ "mlp.experts.35.down_proj",
214
+ "mlp.experts.36.down_proj",
215
+ "mlp.experts.37.down_proj",
216
+ "mlp.experts.38.down_proj",
217
+ "mlp.experts.39.down_proj",
218
+ "mlp.experts.40.down_proj",
219
+ "mlp.experts.41.down_proj",
220
+ "mlp.experts.42.down_proj",
221
+ "mlp.experts.43.down_proj",
222
+ "mlp.experts.44.down_proj",
223
+ "mlp.experts.45.down_proj",
224
+ "mlp.experts.46.down_proj",
225
+ "mlp.experts.47.down_proj",
226
+ "mlp.experts.48.down_proj",
227
+ "mlp.experts.49.down_proj",
228
+ "mlp.experts.50.down_proj",
229
+ "mlp.experts.51.down_proj",
230
+ "mlp.experts.52.down_proj",
231
+ "mlp.experts.53.down_proj",
232
+ "mlp.experts.54.down_proj",
233
+ "mlp.experts.55.down_proj",
234
+ "mlp.experts.56.down_proj",
235
+ "mlp.experts.57.down_proj",
236
+ "mlp.experts.58.down_proj",
237
+ "mlp.experts.59.down_proj"
238
+ ]
239
+ ],
240
+ "pad_token_id": null,
241
+ "quant_method": "gptq",
242
+ "sym": true,
243
+ "tokenizer": null,
244
+ "true_sequential": true,
245
+ "use_cuda_fp16": false,
246
+ "use_exllama": true
247
+ },
248
+ "rms_norm_eps": 1e-06,
249
+ "rope_theta": 1000000.0,
250
+ "router_aux_loss_coef": 0.001,
251
+ "shared_expert_intermediate_size": 5632,
252
+ "sliding_window": 4096,
253
+ "tie_word_embeddings": false,
254
+ "torch_dtype": "float16",
255
+ "transformers_version": "4.39.0.dev0",
256
+ "use_cache": true,
257
+ "use_sliding_window": false,
258
+ "vocab_size": 151936
259
+ }
generation_config.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token_id": 151643,
3
+ "do_sample": true,
4
+ "eos_token_id": [
5
+ 151645,
6
+ 151643
7
+ ],
8
+ "pad_token_id": 151643,
9
+ "repetition_penalty": 1.05,
10
+ "temperature": 0.7,
11
+ "top_k": 20,
12
+ "top_p": 0.8,
13
+ "transformers_version": "4.39.0.dev0"
14
+ }
merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
model-00001-of-00003.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9fc14789a514d8b6fa6163c24b072fbcde6f061c52e48182ebebd215bbd80726
3
+ size 4000527016
model-00002-of-00003.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9d16d52c40f5c2fd32ab050fdbcee923db6b7355b5511f2514cd464cf9f8afb6
3
+ size 3791651312
model-00003-of-00003.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1fa7dea175bfbc5fafb5ee7b098f72acacb4a1744078ea067346f6d63b729481
3
+ size 622329984
model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer_config.json ADDED
@@ -0,0 +1,41 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "added_tokens_decoder": {
4
+ "151643": {
5
+ "content": "<|endoftext|>",
6
+ "lstrip": false,
7
+ "normalized": false,
8
+ "rstrip": false,
9
+ "single_word": false,
10
+ "special": true
11
+ },
12
+ "151644": {
13
+ "content": "<|im_start|>",
14
+ "lstrip": false,
15
+ "normalized": false,
16
+ "rstrip": false,
17
+ "single_word": false,
18
+ "special": true
19
+ },
20
+ "151645": {
21
+ "content": "<|im_end|>",
22
+ "lstrip": false,
23
+ "normalized": false,
24
+ "rstrip": false,
25
+ "single_word": false,
26
+ "special": true
27
+ }
28
+ },
29
+ "additional_special_tokens": ["<|im_start|>", "<|im_end|>"],
30
+ "bos_token": null,
31
+ "chat_template": "{% for message in messages %}{% if loop.first and messages[0]['role'] != 'system' %}{{ '<|im_start|>system\nYou are a helpful assistant<|im_end|>\n' }}{% endif %}{{'<|im_start|>' + message['role'] + '\n' + message['content']}}{% if (loop.last and add_generation_prompt) or not loop.last %}{{ '<|im_end|>' + '\n'}}{% endif %}{% endfor %}{% if add_generation_prompt and messages[-1]['role'] != 'assistant' %}{{ '<|im_start|>assistant\n' }}{% endif %}",
32
+ "clean_up_tokenization_spaces": false,
33
+ "eos_token": "<|im_end|>",
34
+ "errors": "replace",
35
+ "model_max_length": 32768,
36
+ "pad_token": "<|endoftext|>",
37
+ "split_special_tokens": false,
38
+ "tokenizer_class": "Qwen2Tokenizer",
39
+ "unk_token": null
40
+ }
41
+
vocab.json ADDED
The diff for this file is too large to render. See raw diff