Haimath commited on
Commit
bf191d8
1 Parent(s): 232bfca

Upload 10 files

Browse files
README.md ADDED
@@ -0,0 +1,152 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ pipeline_tag: image-to-text
3
+ tags:
4
+ - image-captioning
5
+ languages:
6
+ - en
7
+ license: bsd-3-clause
8
+ ---
9
+
10
+ # BLIP: Bootstrapping Language-Image Pre-training for Unified Vision-Language Understanding and Generation
11
+
12
+ Model card for image captioning pretrained on COCO dataset - base architecture (with ViT base backbone).
13
+
14
+ | ![BLIP.gif](https://cdn-uploads.huggingface.co/production/uploads/1670928184033-62441d1d9fdefb55a0b7d12c.gif) |
15
+ |:--:|
16
+ | <b> Pull figure from BLIP official repo | Image source: https://github.com/salesforce/BLIP </b>|
17
+
18
+ ## TL;DR
19
+
20
+ Authors from the [paper](https://arxiv.org/abs/2201.12086) write in the abstract:
21
+
22
+ *Vision-Language Pre-training (VLP) has advanced the performance for many vision-language tasks. However, most existing pre-trained models only excel in either understanding-based tasks or generation-based tasks. Furthermore, performance improvement has been largely achieved by scaling up the dataset with noisy image-text pairs collected from the web, which is a suboptimal source of supervision. In this paper, we propose BLIP, a new VLP framework which transfers flexibly to both vision-language understanding and generation tasks. BLIP effectively utilizes the noisy web data by bootstrapping the captions, where a captioner generates synthetic captions and a filter removes the noisy ones. We achieve state-of-the-art results on a wide range of vision-language tasks, such as image-text retrieval (+2.7% in average recall@1), image captioning (+2.8% in CIDEr), and VQA (+1.6% in VQA score). BLIP also demonstrates strong generalization ability when directly transferred to videolanguage tasks in a zero-shot manner. Code, models, and datasets are released.*
23
+
24
+ ## Usage
25
+
26
+ You can use this model for conditional and un-conditional image captioning
27
+
28
+ ### Using the Pytorch model
29
+
30
+ #### Running the model on CPU
31
+
32
+ <details>
33
+ <summary> Click to expand </summary>
34
+
35
+ ```python
36
+ import requests
37
+ from PIL import Image
38
+ from transformers import BlipProcessor, BlipForConditionalGeneration
39
+
40
+ processor = BlipProcessor.from_pretrained("Salesforce/blip-image-captioning-base")
41
+ model = BlipForConditionalGeneration.from_pretrained("Salesforce/blip-image-captioning-base")
42
+
43
+ img_url = 'https://storage.googleapis.com/sfr-vision-language-research/BLIP/demo.jpg'
44
+ raw_image = Image.open(requests.get(img_url, stream=True).raw).convert('RGB')
45
+
46
+ # conditional image captioning
47
+ text = "a photography of"
48
+ inputs = processor(raw_image, text, return_tensors="pt")
49
+
50
+ out = model.generate(**inputs)
51
+ print(processor.decode(out[0], skip_special_tokens=True))
52
+ # >>> a photography of a woman and her dog
53
+
54
+ # unconditional image captioning
55
+ inputs = processor(raw_image, return_tensors="pt")
56
+
57
+ out = model.generate(**inputs)
58
+ print(processor.decode(out[0], skip_special_tokens=True))
59
+ >>> a woman sitting on the beach with her dog
60
+ ```
61
+ </details>
62
+
63
+ #### Running the model on GPU
64
+
65
+ ##### In full precision
66
+
67
+ <details>
68
+ <summary> Click to expand </summary>
69
+
70
+ ```python
71
+ import requests
72
+ from PIL import Image
73
+ from transformers import BlipProcessor, BlipForConditionalGeneration
74
+
75
+ processor = BlipProcessor.from_pretrained("Salesforce/blip-image-captioning-base")
76
+ model = BlipForConditionalGeneration.from_pretrained("Salesforce/blip-image-captioning-base").to("cuda")
77
+
78
+ img_url = 'https://storage.googleapis.com/sfr-vision-language-research/BLIP/demo.jpg'
79
+ raw_image = Image.open(requests.get(img_url, stream=True).raw).convert('RGB')
80
+
81
+ # conditional image captioning
82
+ text = "a photography of"
83
+ inputs = processor(raw_image, text, return_tensors="pt").to("cuda")
84
+
85
+ out = model.generate(**inputs)
86
+ print(processor.decode(out[0], skip_special_tokens=True))
87
+ # >>> a photography of a woman and her dog
88
+
89
+ # unconditional image captioning
90
+ inputs = processor(raw_image, return_tensors="pt").to("cuda")
91
+
92
+ out = model.generate(**inputs)
93
+ print(processor.decode(out[0], skip_special_tokens=True))
94
+ >>> a woman sitting on the beach with her dog
95
+ ```
96
+ </details>
97
+
98
+ ##### In half precision (`float16`)
99
+
100
+ <details>
101
+ <summary> Click to expand </summary>
102
+
103
+ ```python
104
+ import torch
105
+ import requests
106
+ from PIL import Image
107
+ from transformers import BlipProcessor, BlipForConditionalGeneration
108
+
109
+ processor = BlipProcessor.from_pretrained("Salesforce/blip-image-captioning-base")
110
+ model = BlipForConditionalGeneration.from_pretrained("Salesforce/blip-image-captioning-base", torch_dtype=torch.float16).to("cuda")
111
+
112
+ img_url = 'https://storage.googleapis.com/sfr-vision-language-research/BLIP/demo.jpg'
113
+ raw_image = Image.open(requests.get(img_url, stream=True).raw).convert('RGB')
114
+
115
+ # conditional image captioning
116
+ text = "a photography of"
117
+ inputs = processor(raw_image, text, return_tensors="pt").to("cuda", torch.float16)
118
+
119
+ out = model.generate(**inputs)
120
+ print(processor.decode(out[0], skip_special_tokens=True))
121
+ # >>> a photography of a woman and her dog
122
+
123
+ # unconditional image captioning
124
+ inputs = processor(raw_image, return_tensors="pt").to("cuda", torch.float16)
125
+
126
+ out = model.generate(**inputs)
127
+ print(processor.decode(out[0], skip_special_tokens=True))
128
+ >>> a woman sitting on the beach with her dog
129
+ ```
130
+ </details>
131
+
132
+ ## BibTex and citation info
133
+
134
+ ```
135
+ @misc{https://doi.org/10.48550/arxiv.2201.12086,
136
+ doi = {10.48550/ARXIV.2201.12086},
137
+
138
+ url = {https://arxiv.org/abs/2201.12086},
139
+
140
+ author = {Li, Junnan and Li, Dongxu and Xiong, Caiming and Hoi, Steven},
141
+
142
+ keywords = {Computer Vision and Pattern Recognition (cs.CV), FOS: Computer and information sciences, FOS: Computer and information sciences},
143
+
144
+ title = {BLIP: Bootstrapping Language-Image Pre-training for Unified Vision-Language Understanding and Generation},
145
+
146
+ publisher = {arXiv},
147
+
148
+ year = {2022},
149
+
150
+ copyright = {Creative Commons Attribution 4.0 International}
151
+ }
152
+ ```
config.json ADDED
@@ -0,0 +1,169 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_commit_hash": null,
3
+ "architectures": [
4
+ "BlipForConditionalGeneration"
5
+ ],
6
+ "image_text_hidden_size": 256,
7
+ "initializer_factor": 1.0,
8
+ "logit_scale_init_value": 2.6592,
9
+ "model_type": "blip",
10
+ "projection_dim": 512,
11
+ "text_config": {
12
+ "_name_or_path": "",
13
+ "add_cross_attention": false,
14
+ "architectures": null,
15
+ "attention_probs_dropout_prob": 0.0,
16
+ "bad_words_ids": null,
17
+ "begin_suppress_tokens": null,
18
+ "bos_token_id": 30522,
19
+ "chunk_size_feed_forward": 0,
20
+ "cross_attention_hidden_size": null,
21
+ "decoder_start_token_id": null,
22
+ "diversity_penalty": 0.0,
23
+ "do_sample": false,
24
+ "early_stopping": false,
25
+ "encoder_no_repeat_ngram_size": 0,
26
+ "eos_token_id": 2,
27
+ "exponential_decay_length_penalty": null,
28
+ "finetuning_task": null,
29
+ "forced_bos_token_id": null,
30
+ "forced_eos_token_id": null,
31
+ "hidden_act": "gelu",
32
+ "hidden_dropout_prob": 0.0,
33
+ "hidden_size": 768,
34
+ "id2label": {
35
+ "0": "LABEL_0",
36
+ "1": "LABEL_1"
37
+ },
38
+ "initializer_factor": 1.0,
39
+ "initializer_range": 0.02,
40
+ "intermediate_size": 3072,
41
+ "is_decoder": true,
42
+ "is_encoder_decoder": false,
43
+ "label2id": {
44
+ "LABEL_0": 0,
45
+ "LABEL_1": 1
46
+ },
47
+ "layer_norm_eps": 1e-12,
48
+ "length_penalty": 1.0,
49
+ "max_length": 20,
50
+ "max_position_embeddings": 512,
51
+ "min_length": 0,
52
+ "model_type": "blip_text_model",
53
+ "no_repeat_ngram_size": 0,
54
+ "num_attention_heads": 12,
55
+ "num_beam_groups": 1,
56
+ "num_beams": 1,
57
+ "num_hidden_layers": 12,
58
+ "num_return_sequences": 1,
59
+ "output_attentions": false,
60
+ "output_hidden_states": false,
61
+ "output_scores": false,
62
+ "pad_token_id": 0,
63
+ "prefix": null,
64
+ "problem_type": null,
65
+ "projection_dim": 768,
66
+ "pruned_heads": {},
67
+ "remove_invalid_values": false,
68
+ "repetition_penalty": 1.0,
69
+ "return_dict": true,
70
+ "return_dict_in_generate": false,
71
+ "sep_token_id": 102,
72
+ "suppress_tokens": null,
73
+ "task_specific_params": null,
74
+ "temperature": 1.0,
75
+ "tf_legacy_loss": false,
76
+ "tie_encoder_decoder": false,
77
+ "tie_word_embeddings": true,
78
+ "tokenizer_class": null,
79
+ "top_k": 50,
80
+ "top_p": 1.0,
81
+ "torch_dtype": null,
82
+ "torchscript": false,
83
+ "transformers_version": "4.26.0.dev0",
84
+ "typical_p": 1.0,
85
+ "use_bfloat16": false,
86
+ "use_cache": true,
87
+ "vocab_size": 30524
88
+ },
89
+ "torch_dtype": "float32",
90
+ "transformers_version": null,
91
+ "vision_config": {
92
+ "_name_or_path": "",
93
+ "add_cross_attention": false,
94
+ "architectures": null,
95
+ "attention_dropout": 0.0,
96
+ "bad_words_ids": null,
97
+ "begin_suppress_tokens": null,
98
+ "bos_token_id": null,
99
+ "chunk_size_feed_forward": 0,
100
+ "cross_attention_hidden_size": null,
101
+ "decoder_start_token_id": null,
102
+ "diversity_penalty": 0.0,
103
+ "do_sample": false,
104
+ "dropout": 0.0,
105
+ "early_stopping": false,
106
+ "encoder_no_repeat_ngram_size": 0,
107
+ "eos_token_id": null,
108
+ "exponential_decay_length_penalty": null,
109
+ "finetuning_task": null,
110
+ "forced_bos_token_id": null,
111
+ "forced_eos_token_id": null,
112
+ "hidden_act": "gelu",
113
+ "hidden_size": 768,
114
+ "id2label": {
115
+ "0": "LABEL_0",
116
+ "1": "LABEL_1"
117
+ },
118
+ "image_size": 384,
119
+ "initializer_factor": 1.0,
120
+ "initializer_range": 0.02,
121
+ "intermediate_size": 3072,
122
+ "is_decoder": false,
123
+ "is_encoder_decoder": false,
124
+ "label2id": {
125
+ "LABEL_0": 0,
126
+ "LABEL_1": 1
127
+ },
128
+ "layer_norm_eps": 1e-05,
129
+ "length_penalty": 1.0,
130
+ "max_length": 20,
131
+ "min_length": 0,
132
+ "model_type": "blip_vision_model",
133
+ "no_repeat_ngram_size": 0,
134
+ "num_attention_heads": 12,
135
+ "num_beam_groups": 1,
136
+ "num_beams": 1,
137
+ "num_channels": 3,
138
+ "num_hidden_layers": 12,
139
+ "num_return_sequences": 1,
140
+ "output_attentions": false,
141
+ "output_hidden_states": false,
142
+ "output_scores": false,
143
+ "pad_token_id": null,
144
+ "patch_size": 16,
145
+ "prefix": null,
146
+ "problem_type": null,
147
+ "projection_dim": 512,
148
+ "pruned_heads": {},
149
+ "remove_invalid_values": false,
150
+ "repetition_penalty": 1.0,
151
+ "return_dict": true,
152
+ "return_dict_in_generate": false,
153
+ "sep_token_id": null,
154
+ "suppress_tokens": null,
155
+ "task_specific_params": null,
156
+ "temperature": 1.0,
157
+ "tf_legacy_loss": false,
158
+ "tie_encoder_decoder": false,
159
+ "tie_word_embeddings": true,
160
+ "tokenizer_class": null,
161
+ "top_k": 50,
162
+ "top_p": 1.0,
163
+ "torch_dtype": null,
164
+ "torchscript": false,
165
+ "transformers_version": "4.26.0.dev0",
166
+ "typical_p": 1.0,
167
+ "use_bfloat16": false
168
+ }
169
+ }
gitattributes.txt ADDED
@@ -0,0 +1,34 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ *.7z filter=lfs diff=lfs merge=lfs -text
2
+ *.arrow filter=lfs diff=lfs merge=lfs -text
3
+ *.bin filter=lfs diff=lfs merge=lfs -text
4
+ *.bz2 filter=lfs diff=lfs merge=lfs -text
5
+ *.ckpt filter=lfs diff=lfs merge=lfs -text
6
+ *.ftz filter=lfs diff=lfs merge=lfs -text
7
+ *.gz filter=lfs diff=lfs merge=lfs -text
8
+ *.h5 filter=lfs diff=lfs merge=lfs -text
9
+ *.joblib filter=lfs diff=lfs merge=lfs -text
10
+ *.lfs.* filter=lfs diff=lfs merge=lfs -text
11
+ *.mlmodel filter=lfs diff=lfs merge=lfs -text
12
+ *.model filter=lfs diff=lfs merge=lfs -text
13
+ *.msgpack filter=lfs diff=lfs merge=lfs -text
14
+ *.npy filter=lfs diff=lfs merge=lfs -text
15
+ *.npz filter=lfs diff=lfs merge=lfs -text
16
+ *.onnx filter=lfs diff=lfs merge=lfs -text
17
+ *.ot filter=lfs diff=lfs merge=lfs -text
18
+ *.parquet filter=lfs diff=lfs merge=lfs -text
19
+ *.pb filter=lfs diff=lfs merge=lfs -text
20
+ *.pickle filter=lfs diff=lfs merge=lfs -text
21
+ *.pkl filter=lfs diff=lfs merge=lfs -text
22
+ *.pt filter=lfs diff=lfs merge=lfs -text
23
+ *.pth filter=lfs diff=lfs merge=lfs -text
24
+ *.rar filter=lfs diff=lfs merge=lfs -text
25
+ *.safetensors filter=lfs diff=lfs merge=lfs -text
26
+ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
27
+ *.tar.* filter=lfs diff=lfs merge=lfs -text
28
+ *.tflite filter=lfs diff=lfs merge=lfs -text
29
+ *.tgz filter=lfs diff=lfs merge=lfs -text
30
+ *.wasm filter=lfs diff=lfs merge=lfs -text
31
+ *.xz filter=lfs diff=lfs merge=lfs -text
32
+ *.zip filter=lfs diff=lfs merge=lfs -text
33
+ *.zst filter=lfs diff=lfs merge=lfs -text
34
+ *tfevents* filter=lfs diff=lfs merge=lfs -text
preprocessor_config.json ADDED
@@ -0,0 +1,17 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_normalize": true,
3
+ "do_resize": true,
4
+ "image_mean": [
5
+ 0.48145466,
6
+ 0.4578275,
7
+ 0.40821073
8
+ ],
9
+ "image_processor_type": "BlipImageProcessor",
10
+ "image_std": [
11
+ 0.26862954,
12
+ 0.26130258,
13
+ 0.27577711
14
+ ],
15
+ "processor_class": "BlipProcessor",
16
+ "size": 384
17
+ }
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a014e2a72e3435fd3f5836965d85fecbc79e1df1c325fc84dd942f89b303b796
3
+ size 989821731
special_tokens_map.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "cls_token": "[CLS]",
3
+ "mask_token": "[MASK]",
4
+ "pad_token": "[PAD]",
5
+ "sep_token": "[SEP]",
6
+ "unk_token": "[UNK]"
7
+ }
tf_model.h5 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d0aaa4c0e003f599d8baa53a9dee85af14eef20554cf2f8113a2673e25a59f8c
3
+ size 990275136
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer_config.json ADDED
@@ -0,0 +1,21 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "cls_token": "[CLS]",
3
+ "do_basic_tokenize": true,
4
+ "do_lower_case": true,
5
+ "mask_token": "[MASK]",
6
+ "model_max_length": 512,
7
+ "name_or_path": "bert-base-uncased",
8
+ "never_split": null,
9
+ "pad_token": "[PAD]",
10
+ "processor_class": "BlipProcessor",
11
+ "sep_token": "[SEP]",
12
+ "special_tokens_map_file": null,
13
+ "strip_accents": null,
14
+ "tokenize_chinese_chars": true,
15
+ "tokenizer_class": "BertTokenizer",
16
+ "unk_token": "[UNK]",
17
+ "model_input_names": [
18
+ "input_ids",
19
+ "attention_mask"
20
+ ]
21
+ }
vocab.txt ADDED
The diff for this file is too large to render. See raw diff