Hugging Face's logo

Sahil-Kabir
/
colqwen2.5-v0.2-hf

Safetensors
colqwen2
Model card Files
xet
Community
1
Sahil-Kabir commited on
Commit
50eaf99
·
verified ·
1 Parent(s): 18d224e

Upload 8 files

Browse files
Files changed (4) hide show
  1. config.json +229 -108
  2. model-00001-of-00002.safetensors +2 -2
  3. model-00002-of-00002.safetensors +2 -2
  4. model.safetensors.index.json +0 -0
config.json CHANGED
@@ -1,87 +1,27 @@
1
  {
2
  "architectures": [
3
- "ColQwen2"
4
  ],
5
- "attention_dropout": 0.0,
6
- "bos_token_id": 151643,
7
- "eos_token_id": 151645,
8
- "hidden_act": "silu",
9
- "hidden_size": 2048,
10
- "image_token_id": 151655,
11
  "initializer_range": 0.02,
12
- "intermediate_size": 11008,
13
- "max_position_embeddings": 128000,
14
- "max_window_layers": 70,
15
  "model_type": "colqwen2",
16
- "num_attention_heads": 16,
17
- "num_hidden_layers": 36,
18
- "num_key_value_heads": 2,
19
- "rms_norm_eps": 1e-06,
20
- "rope_scaling": {
21
- "mrope_section": [
22
- 16,
23
- 24,
24
- 24
25
- ],
26
- "rope_type": "default",
27
- "type": "default"
28
- },
29
- "rope_theta": 1000000.0,
30
- "sliding_window": 32768,
31
- "text_config": {
32
  "_name_or_path": "Qwen/Qwen2.5-VL-3B-Instruct",
33
- "architectures": [
34
- "ColQwen2"
35
- ],
36
  "attention_dropout": 0.0,
37
  "bos_token_id": 151643,
 
38
  "eos_token_id": 151645,
39
  "hidden_act": "silu",
40
  "hidden_size": 2048,
41
- "image_token_id": null,
42
  "initializer_range": 0.02,
43
  "intermediate_size": 11008,
44
- "layer_types": [
45
- "full_attention",
46
- "full_attention",
47
- "full_attention",
48
- "full_attention",
49
- "full_attention",
50
- "full_attention",
51
- "full_attention",
52
- "full_attention",
53
- "full_attention",
54
- "full_attention",
55
- "full_attention",
56
- "full_attention",
57
- "full_attention",
58
- "full_attention",
59
- "full_attention",
60
- "full_attention",
61
- "full_attention",
62
- "full_attention",
63
- "full_attention",
64
- "full_attention",
65
- "full_attention",
66
- "full_attention",
67
- "full_attention",
68
- "full_attention",
69
- "full_attention",
70
- "full_attention",
71
- "full_attention",
72
- "full_attention",
73
- "full_attention",
74
- "full_attention",
75
- "full_attention",
76
- "full_attention",
77
- "full_attention",
78
- "full_attention",
79
- "full_attention",
80
- "full_attention"
81
- ],
82
  "max_position_embeddings": 128000,
83
  "max_window_layers": 70,
84
- "model_type": "qwen2_5_vl_text",
85
  "num_attention_heads": 16,
86
  "num_hidden_layers": 36,
87
  "num_key_value_heads": 2,
@@ -96,49 +36,230 @@
96
  "type": "default"
97
  },
98
  "rope_theta": 1000000.0,
99
- "sliding_window": null,
100
- "tie_word_embeddings": true,
101
- "torch_dtype": "bfloat16",
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
102
  "use_cache": true,
103
  "use_sliding_window": false,
104
- "video_token_id": null,
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
105
  "vision_end_token_id": 151653,
106
  "vision_start_token_id": 151652,
107
  "vision_token_id": 151654,
108
  "vocab_size": 151936
109
- },
110
- "torch_dtype": "bfloat16",
111
- "transformers_version": "4.53.3",
112
- "use_cache": true,
113
- "use_sliding_window": false,
114
- "video_token_id": 151656,
115
- "vision_config": {
116
- "depth": 32,
117
- "fullatt_block_indexes": [
118
- 7,
119
- 15,
120
- 23,
121
- 31
122
- ],
123
- "hidden_act": "silu",
124
- "hidden_size": 1280,
125
- "in_channels": 3,
126
- "in_chans": 3,
127
- "initializer_range": 0.02,
128
- "intermediate_size": 3420,
129
- "model_type": "qwen2_5_vl",
130
- "num_heads": 16,
131
- "out_hidden_size": 2048,
132
- "patch_size": 14,
133
- "spatial_merge_size": 2,
134
- "spatial_patch_size": 14,
135
- "temporal_patch_size": 2,
136
- "tokens_per_second": 2,
137
- "torch_dtype": "bfloat16",
138
- "window_size": 112
139
- },
140
- "vision_end_token_id": 151653,
141
- "vision_start_token_id": 151652,
142
- "vision_token_id": 151654,
143
- "vocab_size": 151936
144
  }
 
1
  {
2
  "architectures": [
3
+ "ColQwen2ForRetrieval"
4
  ],
5
+ "dtype": "bfloat16",
6
+ "embedding_dim": 128,
 
 
 
 
7
  "initializer_range": 0.02,
8
+ "is_composition": false,
 
 
9
  "model_type": "colqwen2",
10
+ "transformers_version": "4.57.0.dev0",
11
+ "vlm_config": {
 
 
 
 
 
 
 
 
 
 
 
 
 
 
12
  "_name_or_path": "Qwen/Qwen2.5-VL-3B-Instruct",
 
 
 
13
  "attention_dropout": 0.0,
14
  "bos_token_id": 151643,
15
+ "dtype": "bfloat16",
16
  "eos_token_id": 151645,
17
  "hidden_act": "silu",
18
  "hidden_size": 2048,
19
+ "image_token_id": 151655,
20
  "initializer_range": 0.02,
21
  "intermediate_size": 11008,
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
22
  "max_position_embeddings": 128000,
23
  "max_window_layers": 70,
24
+ "model_type": "qwen2_5_vl",
25
  "num_attention_heads": 16,
26
  "num_hidden_layers": 36,
27
  "num_key_value_heads": 2,
 
36
  "type": "default"
37
  },
38
  "rope_theta": 1000000.0,
39
+ "sliding_window": 32768,
40
+ "text_config": {
41
+ "_name_or_path": "",
42
+ "add_cross_attention": false,
43
+ "architectures": [
44
+ "ColQwen2_5"
45
+ ],
46
+ "attention_dropout": 0.0,
47
+ "bad_words_ids": null,
48
+ "begin_suppress_tokens": null,
49
+ "bos_token_id": 151643,
50
+ "chunk_size_feed_forward": 0,
51
+ "cross_attention_hidden_size": null,
52
+ "decoder_start_token_id": null,
53
+ "diversity_penalty": 0.0,
54
+ "do_sample": false,
55
+ "dtype": "bfloat16",
56
+ "early_stopping": false,
57
+ "encoder_no_repeat_ngram_size": 0,
58
+ "eos_token_id": 151645,
59
+ "exponential_decay_length_penalty": null,
60
+ "finetuning_task": null,
61
+ "forced_bos_token_id": null,
62
+ "forced_eos_token_id": null,
63
+ "hidden_act": "silu",
64
+ "hidden_size": 2048,
65
+ "id2label": {
66
+ "0": "LABEL_0",
67
+ "1": "LABEL_1"
68
+ },
69
+ "image_token_id": null,
70
+ "initializer_range": 0.02,
71
+ "intermediate_size": 11008,
72
+ "is_decoder": false,
73
+ "is_encoder_decoder": false,
74
+ "label2id": {
75
+ "LABEL_0": 0,
76
+ "LABEL_1": 1
77
+ },
78
+ "layer_types": [
79
+ "full_attention",
80
+ "full_attention",
81
+ "full_attention",
82
+ "full_attention",
83
+ "full_attention",
84
+ "full_attention",
85
+ "full_attention",
86
+ "full_attention",
87
+ "full_attention",
88
+ "full_attention",
89
+ "full_attention",
90
+ "full_attention",
91
+ "full_attention",
92
+ "full_attention",
93
+ "full_attention",
94
+ "full_attention",
95
+ "full_attention",
96
+ "full_attention",
97
+ "full_attention",
98
+ "full_attention",
99
+ "full_attention",
100
+ "full_attention",
101
+ "full_attention",
102
+ "full_attention",
103
+ "full_attention",
104
+ "full_attention",
105
+ "full_attention",
106
+ "full_attention",
107
+ "full_attention",
108
+ "full_attention",
109
+ "full_attention",
110
+ "full_attention",
111
+ "full_attention",
112
+ "full_attention",
113
+ "full_attention",
114
+ "full_attention"
115
+ ],
116
+ "length_penalty": 1.0,
117
+ "max_length": 20,
118
+ "max_position_embeddings": 128000,
119
+ "max_window_layers": 70,
120
+ "min_length": 0,
121
+ "model_type": "qwen2_5_vl_text",
122
+ "no_repeat_ngram_size": 0,
123
+ "num_attention_heads": 16,
124
+ "num_beam_groups": 1,
125
+ "num_beams": 1,
126
+ "num_hidden_layers": 36,
127
+ "num_key_value_heads": 2,
128
+ "num_return_sequences": 1,
129
+ "output_attentions": false,
130
+ "output_hidden_states": false,
131
+ "output_scores": false,
132
+ "pad_token_id": null,
133
+ "prefix": null,
134
+ "problem_type": null,
135
+ "pruned_heads": {},
136
+ "remove_invalid_values": false,
137
+ "repetition_penalty": 1.0,
138
+ "return_dict": true,
139
+ "return_dict_in_generate": false,
140
+ "rms_norm_eps": 1e-06,
141
+ "rope_scaling": {
142
+ "mrope_section": [
143
+ 16,
144
+ 24,
145
+ 24
146
+ ],
147
+ "rope_type": "default",
148
+ "type": "default"
149
+ },
150
+ "rope_theta": 1000000.0,
151
+ "sep_token_id": null,
152
+ "sliding_window": null,
153
+ "suppress_tokens": null,
154
+ "task_specific_params": null,
155
+ "temperature": 1.0,
156
+ "tf_legacy_loss": false,
157
+ "tie_encoder_decoder": false,
158
+ "tie_word_embeddings": true,
159
+ "tokenizer_class": null,
160
+ "top_k": 50,
161
+ "top_p": 1.0,
162
+ "torchscript": false,
163
+ "typical_p": 1.0,
164
+ "use_bfloat16": false,
165
+ "use_cache": true,
166
+ "use_sliding_window": false,
167
+ "video_token_id": null,
168
+ "vision_end_token_id": 151653,
169
+ "vision_start_token_id": 151652,
170
+ "vision_token_id": 151654,
171
+ "vocab_size": 151936
172
+ },
173
  "use_cache": true,
174
  "use_sliding_window": false,
175
+ "video_token_id": 151656,
176
+ "vision_config": {
177
+ "_name_or_path": "",
178
+ "add_cross_attention": false,
179
+ "architectures": null,
180
+ "bad_words_ids": null,
181
+ "begin_suppress_tokens": null,
182
+ "bos_token_id": null,
183
+ "chunk_size_feed_forward": 0,
184
+ "cross_attention_hidden_size": null,
185
+ "decoder_start_token_id": null,
186
+ "depth": 32,
187
+ "diversity_penalty": 0.0,
188
+ "do_sample": false,
189
+ "dtype": "bfloat16",
190
+ "early_stopping": false,
191
+ "encoder_no_repeat_ngram_size": 0,
192
+ "eos_token_id": null,
193
+ "exponential_decay_length_penalty": null,
194
+ "finetuning_task": null,
195
+ "forced_bos_token_id": null,
196
+ "forced_eos_token_id": null,
197
+ "fullatt_block_indexes": [
198
+ 7,
199
+ 15,
200
+ 23,
201
+ 31
202
+ ],
203
+ "hidden_act": "silu",
204
+ "hidden_size": 1280,
205
+ "id2label": {
206
+ "0": "LABEL_0",
207
+ "1": "LABEL_1"
208
+ },
209
+ "in_channels": 3,
210
+ "in_chans": 3,
211
+ "initializer_range": 0.02,
212
+ "intermediate_size": 3420,
213
+ "is_decoder": false,
214
+ "is_encoder_decoder": false,
215
+ "label2id": {
216
+ "LABEL_0": 0,
217
+ "LABEL_1": 1
218
+ },
219
+ "length_penalty": 1.0,
220
+ "max_length": 20,
221
+ "min_length": 0,
222
+ "model_type": "qwen2_5_vl",
223
+ "no_repeat_ngram_size": 0,
224
+ "num_beam_groups": 1,
225
+ "num_beams": 1,
226
+ "num_heads": 16,
227
+ "num_return_sequences": 1,
228
+ "out_hidden_size": 2048,
229
+ "output_attentions": false,
230
+ "output_hidden_states": false,
231
+ "output_scores": false,
232
+ "pad_token_id": null,
233
+ "patch_size": 14,
234
+ "prefix": null,
235
+ "problem_type": null,
236
+ "pruned_heads": {},
237
+ "remove_invalid_values": false,
238
+ "repetition_penalty": 1.0,
239
+ "return_dict": true,
240
+ "return_dict_in_generate": false,
241
+ "sep_token_id": null,
242
+ "spatial_merge_size": 2,
243
+ "spatial_patch_size": 14,
244
+ "suppress_tokens": null,
245
+ "task_specific_params": null,
246
+ "temperature": 1.0,
247
+ "temporal_patch_size": 2,
248
+ "tf_legacy_loss": false,
249
+ "tie_encoder_decoder": false,
250
+ "tie_word_embeddings": true,
251
+ "tokenizer_class": null,
252
+ "tokens_per_second": 2,
253
+ "top_k": 50,
254
+ "top_p": 1.0,
255
+ "torchscript": false,
256
+ "typical_p": 1.0,
257
+ "use_bfloat16": false,
258
+ "window_size": 112
259
+ },
260
  "vision_end_token_id": 151653,
261
  "vision_start_token_id": 151652,
262
  "vision_token_id": 151654,
263
  "vocab_size": 151936
264
+ }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
265
  }
model-00001-of-00002.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:e36caeaa49db50be5795471f0aaee792a64f15c9bf196589aee34c912016d6f7
3
- size 4997750760
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:46b4c359abc980a09b547f339ec238fae2685d2bf28e8e8a86945bdb8ae68b59
3
+ size 4997759176
model-00002-of-00002.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:034a413aec84ee5504a7d85e801b34ba75033014bd2090678d77ec2d6db43d7a
3
- size 2512111904
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f938b787dbea325fa78dc54c494c028cb06fbcf19a5d4cc1c2f44a57990adf70
3
+ size 2512115640
model.safetensors.index.json CHANGED
The diff for this file is too large to render. See raw diff