Hugging Face's logo

unsloth
/
gemma-3-27b-pt-unsloth-bnb-4bit

Any-to-Any
Transformers
Safetensors
English
gemma3
unsloth
gemma
google
text-generation-inference
4-bit precision
bitsandbytes
Model card Files
xet
Community
danielhanchen commited on
Commit
a130367
·
verified ·
1 Parent(s): 90b28dd

Add files using upload-large-folder tool

Browse files
Files changed (9) hide show
  1. config.json +73 -46
  2. generation_config.json +8 -3
  3. model-00001-of-00006.safetensors +3 -0
  4. model-00002-of-00006.safetensors +3 -0
  5. model-00003-of-00006.safetensors +3 -0
  6. model-00004-of-00006.safetensors +3 -0
  7. model-00005-of-00006.safetensors +3 -0
  8. model-00006-of-00006.safetensors +3 -0
  9. model.safetensors.index.json +0 -0
config.json CHANGED
@@ -25,68 +25,95 @@
25
  "multi_modal_projector",
26
  "merger",
27
  "modality_projection",
28
- "lm_head",
29
- "multi_modal_projector",
30
- "merger",
31
- "modality_projection",
32
- "vision_tower.vision_model.encoder.layers.24.self_attn",
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
33
  "vision_tower.vision_model.encoder.layers.26.self_attn",
34
- "vision_tower.vision_model.encoder.layers.22.self_attn",
35
- "language_model.model.layers.27.mlp",
36
  "vision_tower.vision_model.encoder.layers.23.self_attn",
37
  "vision_tower.vision_model.encoder.layers.25.self_attn",
38
- "vision_tower.vision_model.encoder.layers.23.mlp",
39
- "vision_tower.vision_model.encoder.layers.25.mlp",
40
- "vision_tower.vision_model.encoder.layers.18.self_attn",
41
- "vision_tower.vision_model.encoder.layers.17.self_attn",
42
- "vision_tower.vision_model.encoder.layers.15.mlp",
43
  "vision_tower.vision_model.encoder.layers.20.self_attn",
 
 
 
 
 
44
  "vision_tower.vision_model.encoder.layers.21.self_attn",
45
- "vision_tower.vision_model.encoder.layers.19.mlp",
46
- "vision_tower.vision_model.encoder.layers.16.mlp",
47
  "vision_tower.vision_model.encoder.layers.24.mlp",
48
- "vision_tower.vision_model.encoder.layers.13.mlp",
49
- "vision_tower.vision_model.encoder.layers.18.mlp",
50
- "vision_tower.vision_model.encoder.layers.8.mlp",
51
- "vision_tower.vision_model.encoder.layers.22.mlp",
52
- "vision_tower.vision_model.encoder.layers.4.mlp",
53
- "vision_tower.vision_model.encoder.layers.16.self_attn",
54
  "vision_tower.vision_model.encoder.layers.21.mlp",
 
 
 
 
 
 
 
 
 
 
 
 
55
  "vision_tower.vision_model.encoder.layers.9.mlp",
 
56
  "vision_tower.vision_model.encoder.layers.13.self_attn",
 
57
  "vision_tower.vision_model.encoder.layers.10.mlp",
58
- "vision_tower.vision_model.encoder.layers.19.self_attn",
59
- "vision_tower.vision_model.encoder.layers.11.self_attn",
60
- "vision_tower.vision_model.encoder.layers.1.mlp",
61
- "vision_tower.vision_model.encoder.layers.17.mlp",
62
- "vision_tower.vision_model.encoder.layers.15.self_attn",
63
- "vision_tower.vision_model.encoder.layers.12.self_attn",
64
- "vision_tower.vision_model.encoder.layers.20.mlp",
65
  "vision_tower.vision_model.encoder.layers.10.self_attn",
66
- "vision_tower.vision_model.encoder.layers.0.self_attn",
67
- "vision_tower.vision_model.encoder.layers.3.mlp",
68
  "vision_tower.vision_model.encoder.layers.9.self_attn",
69
- "vision_tower.vision_model.encoder.layers.1.self_attn",
70
- "vision_tower.vision_model.encoder.layers.14.mlp",
71
- "vision_tower.vision_model.encoder.layers.0.mlp",
72
- "language_model.model.layers.27.self_attn",
73
- "vision_tower.vision_model.encoder.layers.5.mlp",
74
- "language_model.model.layers.5.mlp",
75
- "vision_tower.vision_model.encoder.layers.11.mlp",
76
- "vision_tower.vision_model.encoder.layers.7.mlp",
77
  "vision_tower.vision_model.encoder.layers.7.self_attn",
 
78
  "vision_tower.vision_model.encoder.layers.8.self_attn",
79
- "language_model.model.layers.23.self_attn",
80
- "vision_tower.vision_model.encoder.layers.26.mlp",
81
  "vision_tower.vision_model.encoder.layers.2.mlp",
82
- "vision_tower.vision_model.encoder.layers.6.self_attn",
83
- "vision_tower.vision_model.encoder.layers.14.self_attn",
84
- "vision_tower.vision_model.encoder.layers.12.mlp",
85
  "vision_tower.vision_model.encoder.layers.3.self_attn",
86
- "vision_tower.vision_model.encoder.layers.4.self_attn",
87
- "vision_tower.vision_model.encoder.layers.6.mlp",
88
  "vision_tower.vision_model.encoder.layers.5.self_attn",
89
- "vision_tower.vision_model.encoder.layers.2.self_attn"
 
 
90
  ],
91
  "llm_int8_threshold": 6.0,
92
  "load_in_4bit": true,
@@ -124,7 +151,7 @@
124
  "vocab_size": 262208
125
  },
126
  "torch_dtype": "bfloat16",
127
- "transformers_version": "4.50.0.dev0",
128
  "unsloth_fixed": true,
129
  "vision_config": {
130
  "attention_dropout": 0.0,
 
25
  "multi_modal_projector",
26
  "merger",
27
  "modality_projection",
28
+ "language_model.model.layers.17.self_attn",
29
+ "language_model.model.layers.23.self_attn",
30
+ "language_model.model.layers.5.self_attn",
31
+ "language_model.model.layers.2.self_attn",
32
+ "language_model.model.layers.6.self_attn",
33
+ "language_model.model.layers.19.mlp",
34
+ "language_model.model.layers.17.mlp",
35
+ "language_model.model.layers.16.mlp",
36
+ "language_model.model.layers.18.mlp",
37
+ "language_model.model.layers.4.self_attn",
38
+ "language_model.model.layers.2.mlp",
39
+ "language_model.model.layers.1.mlp",
40
+ "language_model.model.layers.0.self_attn",
41
+ "language_model.model.layers.15.mlp",
42
+ "language_model.model.layers.14.mlp",
43
+ "language_model.model.layers.1.self_attn",
44
+ "language_model.model.layers.7.self_attn",
45
+ "language_model.model.layers.6.mlp",
46
+ "language_model.model.layers.7.mlp",
47
+ "language_model.model.layers.4.mlp",
48
+ "language_model.model.layers.8.self_attn",
49
+ "language_model.model.layers.12.self_attn",
50
+ "language_model.model.layers.0.mlp",
51
+ "language_model.model.layers.13.mlp",
52
+ "language_model.model.layers.8.mlp",
53
+ "language_model.model.layers.5.mlp",
54
+ "language_model.model.layers.9.self_attn",
55
+ "language_model.model.layers.3.mlp",
56
+ "language_model.model.layers.52.self_attn",
57
+ "language_model.model.layers.11.self_attn",
58
+ "language_model.model.layers.11.mlp",
59
+ "language_model.model.layers.10.self_attn",
60
+ "language_model.model.layers.10.mlp",
61
+ "vision_tower.vision_model.encoder.layers.19.self_attn",
62
+ "language_model.model.layers.12.mlp",
63
+ "language_model.model.layers.9.mlp",
64
  "vision_tower.vision_model.encoder.layers.26.self_attn",
65
+ "vision_tower.vision_model.encoder.layers.24.self_attn",
 
66
  "vision_tower.vision_model.encoder.layers.23.self_attn",
67
  "vision_tower.vision_model.encoder.layers.25.self_attn",
68
+ "vision_tower.vision_model.encoder.layers.22.self_attn",
 
 
 
 
69
  "vision_tower.vision_model.encoder.layers.20.self_attn",
70
+ "vision_tower.vision_model.encoder.layers.20.mlp",
71
+ "vision_tower.vision_model.encoder.layers.25.mlp",
72
+ "vision_tower.vision_model.encoder.layers.16.self_attn",
73
+ "vision_tower.vision_model.encoder.layers.17.mlp",
74
+ "vision_tower.vision_model.encoder.layers.18.mlp",
75
  "vision_tower.vision_model.encoder.layers.21.self_attn",
76
+ "vision_tower.vision_model.encoder.layers.15.mlp",
 
77
  "vision_tower.vision_model.encoder.layers.24.mlp",
 
 
 
 
 
 
78
  "vision_tower.vision_model.encoder.layers.21.mlp",
79
+ "vision_tower.vision_model.encoder.layers.18.self_attn",
80
+ "vision_tower.vision_model.encoder.layers.19.mlp",
81
+ "vision_tower.vision_model.encoder.layers.23.mlp",
82
+ "vision_tower.vision_model.encoder.layers.22.mlp",
83
+ "vision_tower.vision_model.encoder.layers.17.self_attn",
84
+ "vision_tower.vision_model.encoder.layers.12.mlp",
85
+ "vision_tower.vision_model.encoder.layers.15.self_attn",
86
+ "vision_tower.vision_model.encoder.layers.14.self_attn",
87
+ "vision_tower.vision_model.encoder.layers.13.mlp",
88
+ "vision_tower.vision_model.encoder.layers.16.mlp",
89
+ "vision_tower.vision_model.encoder.layers.14.mlp",
90
+ "vision_tower.vision_model.encoder.layers.11.mlp",
91
  "vision_tower.vision_model.encoder.layers.9.mlp",
92
+ "vision_tower.vision_model.encoder.layers.8.mlp",
93
  "vision_tower.vision_model.encoder.layers.13.self_attn",
94
+ "vision_tower.vision_model.encoder.layers.5.mlp",
95
  "vision_tower.vision_model.encoder.layers.10.mlp",
96
+ "vision_tower.vision_model.encoder.layers.7.mlp",
 
 
 
 
 
 
97
  "vision_tower.vision_model.encoder.layers.10.self_attn",
98
+ "vision_tower.vision_model.encoder.layers.6.self_attn",
 
99
  "vision_tower.vision_model.encoder.layers.9.self_attn",
100
+ "vision_tower.vision_model.encoder.layers.6.mlp",
101
+ "vision_tower.vision_model.encoder.layers.1.mlp",
102
+ "vision_tower.vision_model.encoder.layers.11.self_attn",
103
+ "vision_tower.vision_model.encoder.layers.4.mlp",
 
 
 
 
104
  "vision_tower.vision_model.encoder.layers.7.self_attn",
105
+ "vision_tower.vision_model.encoder.layers.12.self_attn",
106
  "vision_tower.vision_model.encoder.layers.8.self_attn",
107
+ "vision_tower.vision_model.encoder.layers.4.self_attn",
108
+ "vision_tower.vision_model.encoder.layers.3.mlp",
109
  "vision_tower.vision_model.encoder.layers.2.mlp",
110
+ "vision_tower.vision_model.encoder.layers.26.mlp",
111
+ "vision_tower.vision_model.encoder.layers.1.self_attn",
 
112
  "vision_tower.vision_model.encoder.layers.3.self_attn",
 
 
113
  "vision_tower.vision_model.encoder.layers.5.self_attn",
114
+ "vision_tower.vision_model.encoder.layers.0.self_attn",
115
+ "vision_tower.vision_model.encoder.layers.2.self_attn",
116
+ "vision_tower.vision_model.encoder.layers.0.mlp"
117
  ],
118
  "llm_int8_threshold": 6.0,
119
  "load_in_4bit": true,
 
151
  "vocab_size": 262208
152
  },
153
  "torch_dtype": "bfloat16",
154
+ "transformers_version": "4.51.0",
155
  "unsloth_fixed": true,
156
  "vision_config": {
157
  "attention_dropout": 0.0,
generation_config.json CHANGED
@@ -1,8 +1,13 @@
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
  }
 
1
  {
 
2
  "bos_token_id": 2,
3
  "cache_implementation": "hybrid",
4
+ "do_sample": true,
5
+ "eos_token_id": [
6
+ 1,
7
+ 106
8
+ ],
9
  "pad_token_id": 0,
10
+ "top_k": 64,
11
+ "top_p": 0.95,
12
+ "transformers_version": "4.51.0"
13
  }
model-00001-of-00006.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fbd8138d46f6f44099c4c2488cd02d738ce82af8d69e076f905d1ad8e2e228ca
3
+ size 4854573696
model-00002-of-00006.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:79ff76d7684e8fd593a1cd8df8135122a86c6de8bfe608db71d8055be2acf4f2
3
+ size 4856758553
model-00003-of-00006.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ab2946ad224a1a5390bec0330f1b814b6497b3495c008d02921da98913aa6091
3
+ size 4856758609
model-00004-of-00006.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:673f01fe49a23bc036a807e2c99f7b8c82ed93fd3961dbcba2e9d1847f1023b6
3
+ size 4961174560
model-00005-of-00006.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dc51d4f8627efb8736143206634c41dc1b6ef7216209e0a31ef9933c6f53e54f
3
+ size 4998165819
model-00006-of-00006.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3a22684c1b86f129ca8bf3962af7889625192c8ca19c59761a3e5d306abe1d8d
3
+ size 4111892059
model.safetensors.index.json CHANGED
The diff for this file is too large to render. See raw diff