Hugging Face's logo

egemensert
/
inek-gemma-3-12b-pt

Image-Text-to-Text
Transformers
TensorBoard
Safetensors
gemma3
Generated from Trainer
trl
sft
conversational
text-generation-inference
Model card Files Metrics Community
egemensert commited on
Commit
7e0f29a
·
verified ·
1 Parent(s): 7524437

Upload Gemma3ForConditionalGeneration

Browse files
Files changed (15) hide show
  1. config.json +25 -3
  2. generation_config.json +13 -0
  3. model-00001-of-00012.safetensors +3 -0
  4. model-00002-of-00012.safetensors +3 -0
  5. model-00003-of-00012.safetensors +3 -0
  6. model-00004-of-00012.safetensors +3 -0
  7. model-00005-of-00012.safetensors +3 -0
  8. model-00006-of-00012.safetensors +3 -0
  9. model-00007-of-00012.safetensors +3 -0
  10. model-00008-of-00012.safetensors +3 -0
  11. model-00009-of-00012.safetensors +3 -0
  12. model-00010-of-00012.safetensors +3 -0
  13. model-00011-of-00012.safetensors +3 -0
  14. model-00012-of-00012.safetensors +3 -0
  15. model.safetensors.index.json +0 -0
config.json CHANGED
@@ -9,28 +9,50 @@
9
  "mm_tokens_per_image": 256,
10
  "model_type": "gemma3",
11
  "text_config": {
 
 
 
 
 
 
 
12
  "hidden_size": 3840,
 
13
  "intermediate_size": 15360,
 
14
  "model_type": "gemma3_text",
15
  "num_attention_heads": 16,
16
  "num_hidden_layers": 48,
17
  "num_key_value_heads": 8,
 
 
 
18
  "rope_scaling": {
19
  "factor": 8.0,
20
  "rope_type": "linear"
21
  },
22
- "sliding_window": 1024
 
 
 
 
 
23
  },
24
- "torch_dtype": "bfloat16",
25
- "transformers_version": "4.50.0.dev0",
26
  "vision_config": {
 
 
27
  "hidden_size": 1152,
28
  "image_size": 896,
29
  "intermediate_size": 4304,
 
30
  "model_type": "siglip_vision_model",
31
  "num_attention_heads": 16,
 
32
  "num_hidden_layers": 27,
33
  "patch_size": 14,
 
34
  "vision_use_head": false
35
  }
36
  }
 
9
  "mm_tokens_per_image": 256,
10
  "model_type": "gemma3",
11
  "text_config": {
12
+ "attention_bias": false,
13
+ "attention_dropout": 0.0,
14
+ "attn_logit_softcapping": null,
15
+ "cache_implementation": "hybrid",
16
+ "final_logit_softcapping": null,
17
+ "head_dim": 256,
18
+ "hidden_activation": "gelu_pytorch_tanh",
19
  "hidden_size": 3840,
20
+ "initializer_range": 0.02,
21
  "intermediate_size": 15360,
22
+ "max_position_embeddings": 131072,
23
  "model_type": "gemma3_text",
24
  "num_attention_heads": 16,
25
  "num_hidden_layers": 48,
26
  "num_key_value_heads": 8,
27
+ "query_pre_attn_scalar": 256,
28
+ "rms_norm_eps": 1e-06,
29
+ "rope_local_base_freq": 10000.0,
30
  "rope_scaling": {
31
  "factor": 8.0,
32
  "rope_type": "linear"
33
  },
34
+ "rope_theta": 1000000.0,
35
+ "sliding_window": 1024,
36
+ "sliding_window_pattern": 6,
37
+ "torch_dtype": "float32",
38
+ "use_cache": true,
39
+ "vocab_size": 262208
40
  },
41
+ "torch_dtype": "float32",
42
+ "transformers_version": "4.51.0.dev0",
43
  "vision_config": {
44
+ "attention_dropout": 0.0,
45
+ "hidden_act": "gelu_pytorch_tanh",
46
  "hidden_size": 1152,
47
  "image_size": 896,
48
  "intermediate_size": 4304,
49
+ "layer_norm_eps": 1e-06,
50
  "model_type": "siglip_vision_model",
51
  "num_attention_heads": 16,
52
+ "num_channels": 3,
53
  "num_hidden_layers": 27,
54
  "patch_size": 14,
55
+ "torch_dtype": "float32",
56
  "vision_use_head": false
57
  }
58
  }
generation_config.json ADDED
@@ -0,0 +1,13 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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.dev0"
13
+ }
model-00001-of-00012.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:08ac4ddeeddcb48b549aef357aff564017473f0cd0738dfcef3f468cbac6a501
3
+ size 1685223128
model-00002-of-00012.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8b1b855fd49d925477971067f9797d31ee1e206f073c3a1db086f800d54d4e60
3
+ size 4987027384
model-00003-of-00012.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b4791e85f0e6c60169546cd823d7dcbb43f6797923307d0fe82b1b04662c5f95
3
+ size 4844749824
model-00004-of-00012.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:218ec7cc55691c2be1cc28ae07c2ac821e23a2db34fb816635319f5a858ac4bb
3
+ size 4954909736
model-00005-of-00012.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9221b6efa97794ecd2f2df1391da104b304ed9067e98cf6eeea349d01c52a923
3
+ size 4907664584
model-00006-of-00012.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:152951cc6b0c10a25a392c78cf8807d75a6041d2d0d3f26ab16c1527f8df7ac2
3
+ size 4954909792
model-00007-of-00012.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:037d0adfcc0b5ec168be8f29c9baa9fb402c13c4f524f666391adb05616f43d6
3
+ size 4907664584
model-00008-of-00012.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d394ebebca91eac0df4d127dd6773c85042cf211453022c6000ca4f9a200f5c3
3
+ size 4954909792
model-00009-of-00012.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7399c6fa8f98030c90d875c5000891cd0faa2325c3d4ce293cc7ce3df65f15d6
3
+ size 4907664584
model-00010-of-00012.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ff7da59a922b2d69fb264cbaab632712d0d1c4eb56b6cd8093dd747c7b13f656
3
+ size 4954909792
model-00011-of-00012.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a65f03eb7c0a953355b44887332e1b1ff1baddb9a786d6411152e189c74bc31c
3
+ size 2689808472
model-00012-of-00012.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4a1e0bf97783f3ecc4003a927ef11bd0567faf85f78c1ee9923116d7600755a5
3
+ size 4027515024
model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff