Hugging Face's logo

OrionCAF
/
qwen2_5_turkish_vlm

Safetensors
qwen2_5_vl
8-bit precision
bitsandbytes
Model card Files Community
1
Cosmobillian commited on
Commit
899ed2d
·
verified ·
1 Parent(s): 43c750a

Upload model trained with Unsloth

Browse files

Upload model trained with Unsloth 2x faster

Files changed (5) hide show
  1. config.json +140 -0
  2. generation_config.json +13 -0
  3. model-00001-of-00002.safetensors +3 -0
  4. model-00002-of-00002.safetensors +3 -0
  5. model.safetensors.index.json +0 -0
config.json ADDED
@@ -0,0 +1,140 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "Qwen2_5_VLForConditionalGeneration"
4
+ ],
5
+ "attention_dropout": 0.0,
6
+ "eos_token_id": 151645,
7
+ "hidden_act": "silu",
8
+ "hidden_size": 3584,
9
+ "image_token_id": 151655,
10
+ "initializer_range": 0.02,
11
+ "intermediate_size": 18944,
12
+ "max_position_embeddings": 128000,
13
+ "max_window_layers": 28,
14
+ "model_type": "qwen2_5_vl",
15
+ "num_attention_heads": 28,
16
+ "num_hidden_layers": 28,
17
+ "num_key_value_heads": 4,
18
+ "pad_token_id": 151654,
19
+ "quantization_config": {
20
+ "bnb_4bit_compute_dtype": "bfloat16",
21
+ "bnb_4bit_quant_type": "nf4",
22
+ "bnb_4bit_use_double_quant": true,
23
+ "llm_int8_enable_fp32_cpu_offload": false,
24
+ "llm_int8_has_fp16_weight": false,
25
+ "llm_int8_skip_modules": null,
26
+ "llm_int8_threshold": 6.0,
27
+ "load_in_4bit": true,
28
+ "load_in_8bit": false,
29
+ "quant_method": "bitsandbytes"
30
+ },
31
+ "rms_norm_eps": 1e-06,
32
+ "rope_scaling": {
33
+ "mrope_section": [
34
+ 16,
35
+ 24,
36
+ 24
37
+ ],
38
+ "rope_type": "default",
39
+ "type": "default"
40
+ },
41
+ "rope_theta": 1000000.0,
42
+ "sliding_window": 32768,
43
+ "text_config": {
44
+ "architectures": [
45
+ "Qwen2_5_VLForConditionalGeneration"
46
+ ],
47
+ "attention_dropout": 0.0,
48
+ "eos_token_id": 151645,
49
+ "hidden_act": "silu",
50
+ "hidden_size": 3584,
51
+ "image_token_id": null,
52
+ "initializer_range": 0.02,
53
+ "intermediate_size": 18944,
54
+ "max_position_embeddings": 128000,
55
+ "max_window_layers": 28,
56
+ "model_type": "qwen2_5_vl_text",
57
+ "num_attention_heads": 28,
58
+ "num_hidden_layers": 28,
59
+ "num_key_value_heads": 4,
60
+ "pad_token_id": 151654,
61
+ "quantization_config": {
62
+ "_load_in_4bit": true,
63
+ "_load_in_8bit": false,
64
+ "bnb_4bit_compute_dtype": "bfloat16",
65
+ "bnb_4bit_quant_storage": "uint8",
66
+ "bnb_4bit_quant_type": "nf4",
67
+ "bnb_4bit_use_double_quant": true,
68
+ "llm_int8_enable_fp32_cpu_offload": false,
69
+ "llm_int8_has_fp16_weight": false,
70
+ "llm_int8_skip_modules": [
71
+ "lm_head",
72
+ "multi_modal_projector",
73
+ "merger",
74
+ "modality_projection"
75
+ ],
76
+ "llm_int8_threshold": 6.0,
77
+ "load_in_4bit": true,
78
+ "load_in_8bit": false,
79
+ "quant_method": "bitsandbytes"
80
+ },
81
+ "rms_norm_eps": 1e-06,
82
+ "rope_scaling": {
83
+ "mrope_section": [
84
+ 16,
85
+ 24,
86
+ 24
87
+ ],
88
+ "rope_type": "default",
89
+ "type": "default"
90
+ },
91
+ "rope_theta": 1000000.0,
92
+ "sliding_window": 32768,
93
+ "torch_dtype": "bfloat16",
94
+ "unsloth_fixed": true,
95
+ "use_cache": true,
96
+ "use_sliding_window": false,
97
+ "video_token_id": null,
98
+ "vision_end_token_id": 151653,
99
+ "vision_start_token_id": 151652,
100
+ "vision_token_id": 151654,
101
+ "vocab_size": 152064
102
+ },
103
+ "tie_word_embeddings": false,
104
+ "torch_dtype": "bfloat16",
105
+ "transformers_version": "4.52.2",
106
+ "unsloth_fixed": true,
107
+ "unsloth_version": "2025.5.9",
108
+ "use_cache": true,
109
+ "use_sliding_window": false,
110
+ "video_token_id": 151656,
111
+ "vision_config": {
112
+ "depth": 32,
113
+ "fullatt_block_indexes": [
114
+ 7,
115
+ 15,
116
+ 23,
117
+ 31
118
+ ],
119
+ "hidden_act": "silu",
120
+ "hidden_size": 1280,
121
+ "in_channels": 3,
122
+ "in_chans": 3,
123
+ "initializer_range": 0.02,
124
+ "intermediate_size": 3420,
125
+ "model_type": "qwen2_5_vl",
126
+ "num_heads": 16,
127
+ "out_hidden_size": 3584,
128
+ "patch_size": 14,
129
+ "spatial_merge_size": 2,
130
+ "spatial_patch_size": 14,
131
+ "temporal_patch_size": 2,
132
+ "tokens_per_second": 2,
133
+ "torch_dtype": "bfloat16",
134
+ "window_size": 112
135
+ },
136
+ "vision_end_token_id": 151653,
137
+ "vision_start_token_id": 151652,
138
+ "vision_token_id": 151654,
139
+ "vocab_size": 152064
140
+ }
generation_config.json ADDED
@@ -0,0 +1,13 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token_id": 151643,
3
+ "do_sample": true,
4
+ "eos_token_id": [
5
+ 151645,
6
+ 151643
7
+ ],
8
+ "max_length": 128000,
9
+ "pad_token_id": 151654,
10
+ "repetition_penalty": 1.05,
11
+ "temperature": 1e-06,
12
+ "transformers_version": "4.52.2"
13
+ }
model-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e822b5e0732193833cd3167a75314fc1686492fb29ef80f2cc29d28a52f81e81
3
+ size 3998906443
model-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f259da87b4d09d3795dc9440da90284acf1b8c54744d45550c979effcb3ac5d7
3
+ size 1966863129
model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff