Hugging Face's logo

Novasy
/
nova-vision-7b

Image-Text-to-Text
Transformers
Safetensors
English
qwen2_5_vl
text-generation-inference
unsloth
conversational
Model card Files
xet
 Community
Novasy commited on
Commit
434ccc0
·
verified ·
1 Parent(s): a7f3659

Upload folder using huggingface_hub

Browse files
Files changed (4) hide show
  1. config.json +103 -101
  2. generation_config.json +1 -0
  3. model-00001-of-00004.safetensors +1 -1
  4. tokenizer_config.json +6 -0
config.json CHANGED
@@ -1,109 +1,111 @@
1
  {
 
 
 
 
 
 
 
2
  "architectures": [
3
- "Qwen2_5_VLForConditionalGeneration"
4
  ],
 
 
5
  "torch_dtype": "bfloat16",
6
- "image_token_id": 151655,
7
- "model_type": "qwen2_5_vl",
8
- "text_config": {
9
- "architectures": [
10
- "Qwen2_5_VLForConditionalGeneration"
11
- ],
12
- "attention_dropout": 0.0,
13
- "bos_token_id": 151643,
14
- "torch_dtype": "bfloat16",
15
- "eos_token_id": 151645,
16
- "hidden_act": "silu",
17
- "hidden_size": 3584,
18
- "image_token_id": null,
19
- "initializer_range": 0.02,
20
- "intermediate_size": 18944,
21
- "layer_types": [
22
- "full_attention",
23
- "full_attention",
24
- "full_attention",
25
- "full_attention",
26
- "full_attention",
27
- "full_attention",
28
- "full_attention",
29
- "full_attention",
30
- "full_attention",
31
- "full_attention",
32
- "full_attention",
33
- "full_attention",
34
- "full_attention",
35
- "full_attention",
36
- "full_attention",
37
- "full_attention",
38
- "full_attention",
39
- "full_attention",
40
- "full_attention",
41
- "full_attention",
42
- "full_attention",
43
- "full_attention",
44
- "full_attention",
45
- "full_attention",
46
- "full_attention",
47
- "full_attention",
48
- "full_attention",
49
- "full_attention"
50
- ],
51
- "max_position_embeddings": 128000,
52
- "max_window_layers": 28,
53
- "model_type": "qwen2_5_vl_text",
54
- "num_attention_heads": 28,
55
- "num_hidden_layers": 28,
56
- "num_key_value_heads": 4,
57
- "pad_token_id": null,
58
- "rms_norm_eps": 1e-06,
59
- "rope_parameters": {
60
- "mrope_section": [
61
- 16,
62
- 24,
63
- 24
64
- ],
65
- "rope_theta": 1000000.0,
66
- "rope_type": "default",
67
- "type": "default"
68
- },
69
- "sliding_window": null,
70
- "use_cache": true,
71
- "use_sliding_window": false,
72
- "video_token_id": null,
73
- "vision_end_token_id": 151653,
74
- "vision_start_token_id": 151652,
75
- "vision_token_id": 151654,
76
- "vocab_size": 152064
77
- },
78
- "tie_word_embeddings": false,
79
- "unsloth_fixed": true,
80
- "video_token_id": 151656,
81
- "vision_config": {
82
- "depth": 32,
83
- "torch_dtype": "bfloat16",
84
- "fullatt_block_indexes": [
85
- 7,
86
- 15,
87
- 23,
88
- 31
89
- ],
90
- "hidden_act": "silu",
91
- "hidden_size": 1280,
92
- "in_channels": 3,
93
- "in_chans": 3,
94
- "initializer_range": 0.02,
95
- "intermediate_size": 3420,
96
- "model_type": "qwen2_5_vl",
97
- "num_heads": 16,
98
- "out_hidden_size": 3584,
99
- "patch_size": 14,
100
- "spatial_merge_size": 2,
101
- "spatial_patch_size": 14,
102
- "temporal_patch_size": 2,
103
- "tokens_per_second": 2,
104
- "window_size": 112
105
  },
 
 
 
 
106
  "vision_end_token_id": 151653,
107
  "vision_start_token_id": 151652,
108
- "vision_token_id": 151654
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
109
  }
 
1
  {
2
+ "architectures": [
3
+ "Qwen2_5_VLForConditionalGeneration"
4
+ ],
5
+ "torch_dtype": "bfloat16",
6
+ "image_token_id": 151655,
7
+ "model_type": "qwen2_5_vl",
8
+ "text_config": {
9
  "architectures": [
10
+ "Qwen2_5_VLForConditionalGeneration"
11
  ],
12
+ "attention_dropout": 0.0,
13
+ "bos_token_id": 151643,
14
  "torch_dtype": "bfloat16",
15
+ "eos_token_id": 151645,
16
+ "hidden_act": "silu",
17
+ "hidden_size": 3584,
18
+ "image_token_id": null,
19
+ "initializer_range": 0.02,
20
+ "intermediate_size": 18944,
21
+ "layer_types": [
22
+ "full_attention",
23
+ "full_attention",
24
+ "full_attention",
25
+ "full_attention",
26
+ "full_attention",
27
+ "full_attention",
28
+ "full_attention",
29
+ "full_attention",
30
+ "full_attention",
31
+ "full_attention",
32
+ "full_attention",
33
+ "full_attention",
34
+ "full_attention",
35
+ "full_attention",
36
+ "full_attention",
37
+ "full_attention",
38
+ "full_attention",
39
+ "full_attention",
40
+ "full_attention",
41
+ "full_attention",
42
+ "full_attention",
43
+ "full_attention",
44
+ "full_attention",
45
+ "full_attention",
46
+ "full_attention",
47
+ "full_attention",
48
+ "full_attention",
49
+ "full_attention"
50
+ ],
51
+ "max_position_embeddings": 128000,
52
+ "max_window_layers": 28,
53
+ "model_type": "qwen2_5_vl_text",
54
+ "num_attention_heads": 28,
55
+ "num_hidden_layers": 28,
56
+ "num_key_value_heads": 4,
57
+ "pad_token_id": null,
58
+ "rms_norm_eps": 1e-06,
59
+ "rope_parameters": {
60
+ "mrope_section": [
61
+ 16,
62
+ 24,
63
+ 24
64
+ ],
65
+ "rope_theta": 1000000.0,
66
+ "rope_type": "default",
67
+ "type": "default"
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
68
  },
69
+ "sliding_window": null,
70
+ "use_cache": true,
71
+ "use_sliding_window": false,
72
+ "video_token_id": null,
73
  "vision_end_token_id": 151653,
74
  "vision_start_token_id": 151652,
75
+ "vision_token_id": 151654,
76
+ "vocab_size": 152064
77
+ },
78
+ "tie_word_embeddings": false,
79
+ "unsloth_fixed": true,
80
+ "video_token_id": 151656,
81
+ "vision_config": {
82
+ "depth": 32,
83
+ "torch_dtype": "bfloat16",
84
+ "fullatt_block_indexes": [
85
+ 7,
86
+ 15,
87
+ 23,
88
+ 31
89
+ ],
90
+ "hidden_act": "silu",
91
+ "hidden_size": 1280,
92
+ "in_channels": 3,
93
+ "in_chans": 3,
94
+ "initializer_range": 0.02,
95
+ "intermediate_size": 3420,
96
+ "model_type": "qwen2_5_vl",
97
+ "num_heads": 16,
98
+ "out_hidden_size": 3584,
99
+ "patch_size": 14,
100
+ "spatial_merge_size": 2,
101
+ "spatial_patch_size": 14,
102
+ "temporal_patch_size": 2,
103
+ "tokens_per_second": 2,
104
+ "window_size": 112
105
+ },
106
+ "vision_end_token_id": 151653,
107
+ "vision_start_token_id": 151652,
108
+ "vision_token_id": 151654,
109
+ "_name_or_path": "Novasy/nova-vision-7b",
110
+ "nova_brand_name": "Nova AI"
111
  }
generation_config.json CHANGED
@@ -1,6 +1,7 @@
1
  {
2
  "do_sample": true,
3
  "eos_token_id": [
 
4
  151645,
5
  151645,
6
  151643
 
1
  {
2
  "do_sample": true,
3
  "eos_token_id": [
4
+ 151645,
5
  151645,
6
  151645,
7
  151643
model-00001-of-00004.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:8c9b042ef61fe5258fd6bc1b46af63a99c282ac739b287d873782a0924861d36
3
  size 4968243304
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cae5c08174bd05b1e1d96b7b56ebf4e2526d2ddd4af76e2f5a147d7928e92c78
3
  size 4968243304
tokenizer_config.json CHANGED
@@ -6,12 +6,18 @@
6
  "eos_token": "<|im_end|>",
7
  "errors": "replace",
8
  "is_local": false,
 
9
  "model_max_length": 32768,
 
10
  "pad_token": "<|vision_pad|>",
 
11
  "padding_side": "right",
12
  "processor_class": "Qwen2_5_VLProcessor",
13
  "split_special_tokens": false,
 
14
  "tokenizer_class": "Qwen2Tokenizer",
 
 
15
  "unk_token": null,
16
  "added_tokens_decoder": {
17
  "151643": {
 
6
  "eos_token": "<|im_end|>",
7
  "errors": "replace",
8
  "is_local": false,
9
+ "max_length": 2048,
10
  "model_max_length": 32768,
11
+ "pad_to_multiple_of": null,
12
  "pad_token": "<|vision_pad|>",
13
+ "pad_token_type_id": 0,
14
  "padding_side": "right",
15
  "processor_class": "Qwen2_5_VLProcessor",
16
  "split_special_tokens": false,
17
+ "stride": 0,
18
  "tokenizer_class": "Qwen2Tokenizer",
19
+ "truncation_side": "right",
20
+ "truncation_strategy": "longest_first",
21
  "unk_token": null,
22
  "added_tokens_decoder": {
23
  "151643": {