Hugging Face's logo

ZinengTang
/
image-qformer

Transformers
PyTorch
blip-2
Inference Endpoints
Model card Files Community
ZinengTang commited on
Commit
08dd609
•
1 Parent(s): 2908f6a

Update config.json

Browse files
Files changed (1) hide show
  1. config.json +45 -49
config.json CHANGED
@@ -7,57 +7,53 @@
7
  "model_type": "blip-2",
8
  "num_query_tokens": 32,
9
  "qformer_config": {
10
- {
11
- "architectures": [
12
- "Blip2QFormerModel"
13
- ],
14
- "num_query_tokens": 32,
15
- "attention_probs_dropout_prob": 0.1,
16
- "classifier_dropout": null,
17
- "cross_attention_frequency": 2,
18
- "encoder_hidden_size": 1408,
19
- "hidden_act": "gelu",
20
- "hidden_dropout_prob": 0.1,
21
- "hidden_size": 768,
22
- "initializer_range": 0.02,
23
- "intermediate_size": 3072,
24
- "layer_norm_eps": 1e-12,
25
- "max_position_embeddings": 512,
26
- "model_type": "blip_2_qformer",
27
- "num_attention_heads": 12,
28
- "num_hidden_layers": 12,
29
- "pad_token_id": 0,
30
- "position_embedding_type": "absolute",
31
- "torch_dtype": "float32",
32
- "transformers_version": "4.34.1",
33
- "vocab_size": 30522
34
- }
35
  },
36
  "torch_dtype": "float32",
37
  "transformers_version": "4.34.1",
38
  "vision_config": {
39
- {
40
- "architectures": [
41
- "Blip2VisionModel"
42
- ],
43
- "attention_dropout": 0.0,
44
- "dropout": 0.0,
45
- "hidden_act": "gelu",
46
- "hidden_size": 1408,
47
- "image_size": 224,
48
- "initializer_factor": 1.0,
49
- "initializer_range": 1e-10,
50
- "intermediate_size": 6144,
51
- "layer_norm_eps": 1e-06,
52
- "model_type": "blip_2_vision_model",
53
- "num_attention_heads": 16,
54
- "num_channels": 3,
55
- "num_hidden_layers": 39,
56
- "patch_size": 14,
57
- "projection_dim": 512,
58
- "qkv_bias": true,
59
- "torch_dtype": "float32",
60
- "transformers_version": "4.34.1"
61
- }
62
  }
63
- }
 
7
  "model_type": "blip-2",
8
  "num_query_tokens": 32,
9
  "qformer_config": {
10
+ "architectures": [
11
+ "Blip2QFormerModel"
12
+ ],
13
+ "num_query_tokens": 32,
14
+ "attention_probs_dropout_prob": 0.1,
15
+ "classifier_dropout": null,
16
+ "cross_attention_frequency": 2,
17
+ "encoder_hidden_size": 1408,
18
+ "hidden_act": "gelu",
19
+ "hidden_dropout_prob": 0.1,
20
+ "hidden_size": 768,
21
+ "initializer_range": 0.02,
22
+ "intermediate_size": 3072,
23
+ "layer_norm_eps": 1e-12,
24
+ "max_position_embeddings": 512,
25
+ "model_type": "blip_2_qformer",
26
+ "num_attention_heads": 12,
27
+ "num_hidden_layers": 12,
28
+ "pad_token_id": 0,
29
+ "position_embedding_type": "absolute",
30
+ "torch_dtype": "float32",
31
+ "transformers_version": "4.34.1",
32
+ "vocab_size": 30522
 
 
33
  },
34
  "torch_dtype": "float32",
35
  "transformers_version": "4.34.1",
36
  "vision_config": {
37
+ "architectures": [
38
+ "Blip2VisionModel"
39
+ ],
40
+ "attention_dropout": 0.0,
41
+ "dropout": 0.0,
42
+ "hidden_act": "gelu",
43
+ "hidden_size": 1408,
44
+ "image_size": 224,
45
+ "initializer_factor": 1.0,
46
+ "initializer_range": 1e-10,
47
+ "intermediate_size": 6144,
48
+ "layer_norm_eps": 1e-06,
49
+ "model_type": "blip_2_vision_model",
50
+ "num_attention_heads": 16,
51
+ "num_channels": 3,
52
+ "num_hidden_layers": 39,
53
+ "patch_size": 14,
54
+ "projection_dim": 512,
55
+ "qkv_bias": true,
56
+ "torch_dtype": "float32",
57
+ "transformers_version": "4.34.1"
 
 
58
  }
59
+ }