Hugging Face's logo

unsloth
/
gemma-3-12b-pt-unsloth-bnb-4bit

Image-Text-to-Text
Transformers
Safetensors
English
gemma3
unsloth
gemma
google
text-generation-inference
4-bit precision
bitsandbytes
Model card Files
xet
Community
danielhanchen commited on
Commit
fee248b
·
verified ·
1 Parent(s): 18c5c2c

Add files using upload-large-folder tool

Browse files
Files changed (14) hide show
  1. .gitattributes +1 -0
  2. added_tokens.json +3 -0
  3. config.json +161 -0
  4. generation_config.json +8 -0
  5. model-00001-of-00003.safetensors +3 -0
  6. model-00002-of-00003.safetensors +3 -0
  7. model-00003-of-00003.safetensors +3 -0
  8. model.safetensors.index.json +0 -0
  9. preprocessor_config.json +29 -0
  10. processor_config.json +4 -0
  11. special_tokens_map.json +33 -0
  12. tokenizer.json +3 -0
  13. tokenizer.model +3 -0
  14. tokenizer_config.json +0 -0
.gitattributes CHANGED
@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
+ tokenizer.json filter=lfs diff=lfs merge=lfs -text
added_tokens.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ {
2
+ "<image_soft_token>": 262144
3
+ }
config.json ADDED
@@ -0,0 +1,161 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "Gemma3ForConditionalGeneration"
4
+ ],
5
+ "boi_token_index": 255999,
6
+ "bos_token_id": 2,
7
+ "eoi_token_index": 256000,
8
+ "eos_token_id": 1,
9
+ "image_token_index": 262144,
10
+ "initializer_range": 0.02,
11
+ "mm_tokens_per_image": 256,
12
+ "model_type": "gemma3",
13
+ "pad_token_id": 0,
14
+ "quantization_config": {
15
+ "_load_in_4bit": true,
16
+ "_load_in_8bit": false,
17
+ "bnb_4bit_compute_dtype": "bfloat16",
18
+ "bnb_4bit_quant_storage": "uint8",
19
+ "bnb_4bit_quant_type": "nf4",
20
+ "bnb_4bit_use_double_quant": true,
21
+ "llm_int8_enable_fp32_cpu_offload": false,
22
+ "llm_int8_has_fp16_weight": false,
23
+ "llm_int8_skip_modules": [
24
+ "lm_head",
25
+ "multi_modal_projector",
26
+ "merger",
27
+ "modality_projection",
28
+ "lm_head",
29
+ "multi_modal_projector",
30
+ "merger",
31
+ "modality_projection",
32
+ "language_model.model.layers.14.mlp",
33
+ "language_model.model.layers.15.self_attn",
34
+ "language_model.model.layers.2.mlp",
35
+ "language_model.model.layers.18.self_attn",
36
+ "language_model.model.layers.15.mlp",
37
+ "language_model.model.layers.20.self_attn",
38
+ "language_model.model.layers.11.mlp",
39
+ "language_model.model.layers.4.mlp",
40
+ "language_model.model.layers.0.mlp",
41
+ "language_model.model.layers.17.self_attn",
42
+ "language_model.model.layers.9.mlp",
43
+ "language_model.model.layers.10.mlp",
44
+ "language_model.model.layers.2.self_attn",
45
+ "language_model.model.layers.20.mlp",
46
+ "language_model.model.layers.16.mlp",
47
+ "vision_tower.vision_model.encoder.layers.22.self_attn",
48
+ "vision_tower.vision_model.encoder.layers.21.self_attn",
49
+ "language_model.model.layers.8.mlp",
50
+ "language_model.model.layers.1.mlp",
51
+ "vision_tower.vision_model.encoder.layers.24.self_attn",
52
+ "vision_tower.vision_model.encoder.layers.19.self_attn",
53
+ "language_model.model.layers.7.mlp",
54
+ "vision_tower.vision_model.encoder.layers.17.mlp",
55
+ "vision_tower.vision_model.encoder.layers.15.mlp",
56
+ "vision_tower.vision_model.encoder.layers.17.self_attn",
57
+ "language_model.model.layers.6.mlp",
58
+ "language_model.model.layers.3.mlp",
59
+ "vision_tower.vision_model.encoder.layers.12.mlp",
60
+ "vision_tower.vision_model.encoder.layers.26.self_attn",
61
+ "language_model.model.layers.5.mlp",
62
+ "vision_tower.vision_model.encoder.layers.24.mlp",
63
+ "vision_tower.vision_model.encoder.layers.23.self_attn",
64
+ "vision_tower.vision_model.encoder.layers.20.mlp",
65
+ "vision_tower.vision_model.encoder.layers.25.self_attn",
66
+ "vision_tower.vision_model.encoder.layers.20.self_attn",
67
+ "vision_tower.vision_model.encoder.layers.19.mlp",
68
+ "vision_tower.vision_model.encoder.layers.14.mlp",
69
+ "vision_tower.vision_model.encoder.layers.18.self_attn",
70
+ "vision_tower.vision_model.encoder.layers.16.self_attn",
71
+ "vision_tower.vision_model.encoder.layers.10.mlp",
72
+ "vision_tower.vision_model.encoder.layers.14.self_attn",
73
+ "vision_tower.vision_model.encoder.layers.15.self_attn",
74
+ "vision_tower.vision_model.encoder.layers.22.mlp",
75
+ "vision_tower.vision_model.encoder.layers.13.self_attn",
76
+ "vision_tower.vision_model.encoder.layers.13.mlp",
77
+ "vision_tower.vision_model.encoder.layers.25.mlp",
78
+ "vision_tower.vision_model.encoder.layers.23.mlp",
79
+ "vision_tower.vision_model.encoder.layers.11.mlp",
80
+ "vision_tower.vision_model.encoder.layers.21.mlp",
81
+ "vision_tower.vision_model.encoder.layers.16.mlp",
82
+ "vision_tower.vision_model.encoder.layers.9.self_attn",
83
+ "vision_tower.vision_model.encoder.layers.12.self_attn",
84
+ "vision_tower.vision_model.encoder.layers.18.mlp",
85
+ "vision_tower.vision_model.encoder.layers.4.mlp",
86
+ "vision_tower.vision_model.encoder.layers.7.mlp",
87
+ "vision_tower.vision_model.encoder.layers.9.mlp",
88
+ "vision_tower.vision_model.encoder.layers.11.self_attn",
89
+ "vision_tower.vision_model.encoder.layers.2.mlp",
90
+ "vision_tower.vision_model.encoder.layers.5.self_attn",
91
+ "vision_tower.vision_model.encoder.layers.1.mlp",
92
+ "vision_tower.vision_model.encoder.layers.10.self_attn",
93
+ "vision_tower.vision_model.encoder.layers.8.mlp",
94
+ "vision_tower.vision_model.encoder.layers.4.self_attn",
95
+ "vision_tower.vision_model.encoder.layers.5.mlp",
96
+ "vision_tower.vision_model.encoder.layers.7.self_attn",
97
+ "vision_tower.vision_model.encoder.layers.6.mlp",
98
+ "vision_tower.vision_model.encoder.layers.3.self_attn",
99
+ "vision_tower.vision_model.encoder.layers.3.mlp",
100
+ "vision_tower.vision_model.encoder.layers.1.self_attn",
101
+ "vision_tower.vision_model.encoder.layers.8.self_attn",
102
+ "vision_tower.vision_model.encoder.layers.6.self_attn",
103
+ "vision_tower.vision_model.encoder.layers.26.mlp",
104
+ "vision_tower.vision_model.encoder.layers.0.self_attn",
105
+ "vision_tower.vision_model.encoder.layers.0.mlp",
106
+ "vision_tower.vision_model.encoder.layers.2.self_attn"
107
+ ],
108
+ "llm_int8_threshold": 6.0,
109
+ "load_in_4bit": true,
110
+ "load_in_8bit": false,
111
+ "quant_method": "bitsandbytes"
112
+ },
113
+ "text_config": {
114
+ "attention_bias": false,
115
+ "attention_dropout": 0.0,
116
+ "attn_logit_softcapping": null,
117
+ "cache_implementation": "hybrid",
118
+ "final_logit_softcapping": null,
119
+ "head_dim": 256,
120
+ "hidden_activation": "gelu_pytorch_tanh",
121
+ "hidden_size": 3840,
122
+ "initializer_range": 0.02,
123
+ "intermediate_size": 15360,
124
+ "max_position_embeddings": 131072,
125
+ "model_type": "gemma3_text",
126
+ "num_attention_heads": 16,
127
+ "num_hidden_layers": 48,
128
+ "num_key_value_heads": 8,
129
+ "query_pre_attn_scalar": 256,
130
+ "rms_norm_eps": 1e-06,
131
+ "rope_local_base_freq": 10000.0,
132
+ "rope_scaling": {
133
+ "factor": 8.0,
134
+ "rope_type": "linear"
135
+ },
136
+ "rope_theta": 1000000.0,
137
+ "sliding_window": 1024,
138
+ "sliding_window_pattern": 6,
139
+ "torch_dtype": "bfloat16",
140
+ "use_cache": true,
141
+ "vocab_size": 262208
142
+ },
143
+ "torch_dtype": "bfloat16",
144
+ "transformers_version": "4.50.0.dev0",
145
+ "unsloth_fixed": true,
146
+ "vision_config": {
147
+ "attention_dropout": 0.0,
148
+ "hidden_act": "gelu_pytorch_tanh",
149
+ "hidden_size": 1152,
150
+ "image_size": 896,
151
+ "intermediate_size": 4304,
152
+ "layer_norm_eps": 1e-06,
153
+ "model_type": "siglip_vision_model",
154
+ "num_attention_heads": 16,
155
+ "num_channels": 3,
156
+ "num_hidden_layers": 27,
157
+ "patch_size": 14,
158
+ "torch_dtype": "bfloat16",
159
+ "vision_use_head": false
160
+ }
161
+ }
generation_config.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 2,
4
+ "cache_implementation": "hybrid",
5
+ "eos_token_id": 1,
6
+ "pad_token_id": 0,
7
+ "transformers_version": "4.50.0.dev0"
8
+ }
model-00001-of-00003.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:636caa7592c252469d386c7a567c5ef6b34bac754edb95a033721e0ff3021ab9
3
+ size 4960125164
model-00002-of-00003.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ed20ac4b90eada6a5c89633f495aad5bd41e0a186421a0ebfed9ac1ccbe64701
3
+ size 4992846171
model-00003-of-00003.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a7bedb8fc3a7b4ea28c590264748a908f868256fc48930c13692606e4e6dab21
3
+ size 3007436388
model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
preprocessor_config.json ADDED
@@ -0,0 +1,29 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_convert_rgb": null,
3
+ "do_normalize": true,
4
+ "do_pan_and_scan": null,
5
+ "do_rescale": true,
6
+ "do_resize": true,
7
+ "image_mean": [
8
+ 0.5,
9
+ 0.5,
10
+ 0.5
11
+ ],
12
+ "image_processor_type": "Gemma3ImageProcessor",
13
+ "image_seq_length": 256,
14
+ "image_std": [
15
+ 0.5,
16
+ 0.5,
17
+ 0.5
18
+ ],
19
+ "pan_and_scan_max_num_crops": null,
20
+ "pan_and_scan_min_crop_size": null,
21
+ "pan_and_scan_min_ratio_to_activate": null,
22
+ "processor_class": "Gemma3Processor",
23
+ "resample": 2,
24
+ "rescale_factor": 0.00392156862745098,
25
+ "size": {
26
+ "height": 896,
27
+ "width": 896
28
+ }
29
+ }
processor_config.json ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ {
2
+ "image_seq_length": 256,
3
+ "processor_class": "Gemma3Processor"
4
+ }
special_tokens_map.json ADDED
@@ -0,0 +1,33 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "boi_token": "<start_of_image>",
3
+ "bos_token": {
4
+ "content": "<bos>",
5
+ "lstrip": false,
6
+ "normalized": false,
7
+ "rstrip": false,
8
+ "single_word": false
9
+ },
10
+ "eoi_token": "<end_of_image>",
11
+ "eos_token": {
12
+ "content": "<eos>",
13
+ "lstrip": false,
14
+ "normalized": false,
15
+ "rstrip": false,
16
+ "single_word": false
17
+ },
18
+ "image_token": "<image_soft_token>",
19
+ "pad_token": {
20
+ "content": "<pad>",
21
+ "lstrip": false,
22
+ "normalized": false,
23
+ "rstrip": false,
24
+ "single_word": false
25
+ },
26
+ "unk_token": {
27
+ "content": "<unk>",
28
+ "lstrip": false,
29
+ "normalized": false,
30
+ "rstrip": false,
31
+ "single_word": false
32
+ }
33
+ }
tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4667f2089529e8e7657cfb6d1c19910ae71ff5f28aa7ab2ff2763330affad795
3
+ size 33384568
tokenizer.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1299c11d7cf632ef3b4e11937501358ada021bbdf7c47638d13c0ee982f2e79c
3
+ size 4689074
tokenizer_config.json ADDED
The diff for this file is too large to render. See raw diff