vic-yes/comparing-captioning-models-clone
0
1import gradio as gr2from transformers import AutoProcessor, AutoTokenizer, AutoImageProcessor, AutoModelForCausalLM, BlipForConditionalGeneration, Blip2ForConditionalGeneration, VisionEncoderDecoderModel3import torch4import open_clip5 6from huggingface_hub import hf_hub_download7 8torch.hub.download_url_to_file('http://images.cocodataset.org/val2017/000000039769.jpg', 'cats.jpg')9torch.hub.download_url_to_file('https://huggingface.co/datasets/nielsr/textcaps-sample/resolve/main/stop_sign.png', 'stop_sign.png')10torch.hub.download_url_to_file('https://cdn.openai.com/dall-e-2/demos/text2im/astronaut/horse/photo/0.jpg', 'astronaut.jpg')11 12# git_processor_base = AutoProcessor.from_pretrained("microsoft/git-base-coco")13# git_model_base = AutoModelForCausalLM.from_pretrained("microsoft/git-base-coco")14 15git_processor_large_coco = AutoProcessor.from_pretrained("microsoft/git-large-coco")16git_model_large_coco = AutoModelForCausalLM.from_pretrained("microsoft/git-large-coco")17 18git_processor_large_textcaps = AutoProcessor.from_pretrained("microsoft/git-large-r-textcaps")19git_model_large_textcaps = AutoModelForCausalLM.from_pretrained("microsoft/git-large-r-textcaps")20 21# blip_processor_base = AutoProcessor.from_pretrained("Salesforce/blip-image-captioning-base")22# blip_model_base = BlipForConditionalGeneration.from_pretrained("Salesforce/blip-image-captioning-base")23 24blip_processor_large = AutoProcessor.from_pretrained("Salesforce/blip-image-captioning-large")25blip_model_large = BlipForConditionalGeneration.from_pretrained("Salesforce/blip-image-captioning-large")26 27# blip2_processor = AutoProcessor.from_pretrained("Salesforce/blip2-opt-2.7b")28# blip2_model = Blip2ForConditionalGeneration.from_pretrained("Salesforce/blip2-opt-2.7b", torch_dtype=torch.float16)29 30blip2_processor_8_bit = AutoProcessor.from_pretrained("Salesforce/blip2-opt-6.7b")31blip2_model_8_bit = Blip2ForConditionalGeneration.from_pretrained("Salesforce/blip2-opt-6.7b", device_map="auto", load_in_8bit=True)32 33# vitgpt_processor = AutoImageProcessor.from_pretrained("nlpconnect/vit-gpt2-image-captioning")34# vitgpt_model = VisionEncoderDecoderModel.from_pretrained("nlpconnect/vit-gpt2-image-captioning")35# vitgpt_tokenizer = AutoTokenizer.from_pretrained("nlpconnect/vit-gpt2-image-captioning")36 37coca_model, _, coca_transform = open_clip.create_model_and_transforms(38 model_name="coca_ViT-L-14",39 pretrained="mscoco_finetuned_laion2B-s13B-b90k"40)41 42device = "cuda" if torch.cuda.is_available() else "cpu"43 44# git_model_base.to(device)45# blip_model_base.to(device)46git_model_large_coco.to(device)47git_model_large_textcaps.to(device)48blip_model_large.to(device)49# vitgpt_model.to(device)50coca_model.to(device)51# blip2_model.to(device)52 53def generate_caption(processor, model, image, tokenizer=None, use_float_16=False):54 inputs = processor(images=image, return_tensors="pt").to(device)55 56 if use_float_16:57 inputs = inputs.to(torch.float16)58 59 generated_ids = model.generate(pixel_values=inputs.pixel_values, max_length=50)60 61 if tokenizer is not None:62 generated_caption = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]63 else:64 generated_caption = processor.batch_decode(generated_ids, skip_special_tokens=True)[0]65 66 return generated_caption67 68 69def generate_caption_coca(model, transform, image):70 im = transform(image).unsqueeze(0).to(device)71 with torch.no_grad(), torch.cuda.amp.autocast():72 generated = model.generate(im, seq_len=20)73 return open_clip.decode(generated[0].detach()).split("<end_of_text>")[0].replace("<start_of_text>", "")74 75 76def generate_captions(image):77 # caption_git_base = generate_caption(git_processor_base, git_model_base, image)78 79 caption_git_large_coco = generate_caption(git_processor_large_coco, git_model_large_coco, image)80 81 caption_git_large_textcaps = generate_caption(git_processor_large_textcaps, git_model_large_textcaps, image)82 83 # caption_blip_base = generate_caption(blip_processor_base, blip_model_base, image)84 85 caption_blip_large = generate_caption(blip_processor_large, blip_model_large, image)86 87 # caption_vitgpt = generate_caption(vitgpt_processor, vitgpt_model, image, vitgpt_tokenizer)88 89 caption_coca = generate_caption_coca(coca_model, coca_transform, image)90 91 # caption_blip2 = generate_caption(blip2_processor, blip2_model, image, use_float_16=True).strip()92 93 caption_blip2_8_bit = generate_caption(blip2_processor_8_bit, blip2_model_8_bit, image, use_float_16=True).strip()94 95 return caption_git_large_coco, caption_git_large_textcaps, caption_blip_large, caption_coca, caption_blip2_8_bit96 97 98examples = [["cats.jpg"], ["stop_sign.png"], ["astronaut.jpg"]]99outputs = [gr.outputs.Textbox(label="Caption generated by GIT-large fine-tuned on COCO"), gr.outputs.Textbox(label="Caption generated by GIT-large fine-tuned on TextCaps"), gr.outputs.Textbox(label="Caption generated by BLIP-large"), gr.outputs.Textbox(label="Caption generated by CoCa"), gr.outputs.Textbox(label="Caption generated by BLIP-2 OPT 6.7b")] 100 101title = "Interactive demo: comparing image captioning models"102description = "Gradio Demo to compare GIT, BLIP, CoCa, and BLIP-2, 4 state-of-the-art vision+language models. To use it, simply upload your image and click 'submit', or click one of the examples to load them. Read more at the links below."103article = "<p style='text-align: center'><a href='https://huggingface.co/docs/transformers/main/model_doc/blip' target='_blank'>BLIP docs</a> | <a href='https://huggingface.co/docs/transformers/main/model_doc/git' target='_blank'>GIT docs</a></p>"104 105interface = gr.Interface(fn=generate_captions, 106 inputs=gr.inputs.Image(type="pil"),107 outputs=outputs,108 examples=examples, 109 title=title,110 description=description,111 article=article, 112 enable_queue=True)113interface.launch(debug=True)