Duplicate from nielsr/comparing-VQA-models

Co-authored-by: Niels Rogge <nielsr@users.noreply.huggingface.co>

.gitattributes

@@ -0,0 +1,34 @@
+*.7z filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.bz2 filter=lfs diff=lfs merge=lfs -text
+*.ckpt filter=lfs diff=lfs merge=lfs -text
+*.ftz filter=lfs diff=lfs merge=lfs -text
+*.gz filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.joblib filter=lfs diff=lfs merge=lfs -text
+*.lfs.* filter=lfs diff=lfs merge=lfs -text
+*.mlmodel filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.npy filter=lfs diff=lfs merge=lfs -text
+*.npz filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.ot filter=lfs diff=lfs merge=lfs -text
+*.parquet filter=lfs diff=lfs merge=lfs -text
+*.pb filter=lfs diff=lfs merge=lfs -text
+*.pickle filter=lfs diff=lfs merge=lfs -text
+*.pkl filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text
+*.rar filter=lfs diff=lfs merge=lfs -text
+*.safetensors filter=lfs diff=lfs merge=lfs -text
+saved_model/**/* filter=lfs diff=lfs merge=lfs -text
+*.tar.* filter=lfs diff=lfs merge=lfs -text
+*.tflite filter=lfs diff=lfs merge=lfs -text
+*.tgz filter=lfs diff=lfs merge=lfs -text
+*.wasm filter=lfs diff=lfs merge=lfs -text
+*.xz filter=lfs diff=lfs merge=lfs -text
+*.zip filter=lfs diff=lfs merge=lfs -text
+*.zst filter=lfs diff=lfs merge=lfs -text
+*tfevents* filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,13 @@
+---
+title: Comparing VQA Models
+emoji: 📊
+colorFrom: gray
+colorTo: green
+sdk: gradio
+sdk_version: 3.15.0
+app_file: app.py
+pinned: false
+duplicated_from: nielsr/comparing-VQA-models
+---
+
+Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference

app.py

@@ -0,0 +1,98 @@
+import gradio as gr
+from transformers import AutoProcessor, AutoModelForCausalLM, BlipForQuestionAnswering, ViltForQuestionAnswering
+import torch
+
+torch.hub.download_url_to_file('http://images.cocodataset.org/val2017/000000039769.jpg', 'cats.jpg')
+torch.hub.download_url_to_file('https://huggingface.co/datasets/nielsr/textcaps-sample/resolve/main/stop_sign.png', 'stop_sign.png')
+torch.hub.download_url_to_file('https://cdn.openai.com/dall-e-2/demos/text2im/astronaut/horse/photo/0.jpg', 'astronaut.jpg')
+
+git_processor_base = AutoProcessor.from_pretrained("microsoft/git-base-vqav2")
+git_model_base = AutoModelForCausalLM.from_pretrained("microsoft/git-base-vqav2")
+
+git_processor_large = AutoProcessor.from_pretrained("microsoft/git-large-vqav2")
+git_model_large = AutoModelForCausalLM.from_pretrained("microsoft/git-large-vqav2")
+
+blip_processor_base = AutoProcessor.from_pretrained("Salesforce/blip-vqa-base")
+blip_model_base = BlipForQuestionAnswering.from_pretrained("Salesforce/blip-vqa-base")
+
+blip_processor_large = AutoProcessor.from_pretrained("Salesforce/blip-vqa-capfilt-large")
+blip_model_large = BlipForQuestionAnswering.from_pretrained("Salesforce/blip-vqa-capfilt-large")
+
+vilt_processor = AutoProcessor.from_pretrained("dandelin/vilt-b32-finetuned-vqa")
+vilt_model = ViltForQuestionAnswering.from_pretrained("dandelin/vilt-b32-finetuned-vqa")
+
+device = "cuda" if torch.cuda.is_available() else "cpu"
+
+git_model_base.to(device)
+blip_model_base.to(device)
+git_model_large.to(device)
+blip_model_large.to(device)
+vilt_model.to(device)
+
+def generate_answer_git(processor, model, image, question):
+    # prepare image
+    pixel_values = processor(images=image, return_tensors="pt").pixel_values
+
+    # prepare question
+    input_ids = processor(text=question, add_special_tokens=False).input_ids
+    input_ids = [processor.tokenizer.cls_token_id] + input_ids
+    input_ids = torch.tensor(input_ids).unsqueeze(0)
+    
+    generated_ids = model.generate(pixel_values=pixel_values, input_ids=input_ids, max_length=50)
+    generated_answer = processor.batch_decode(generated_ids, skip_special_tokens=True)
+   
+    return generated_answer
+
+
+def generate_answer_blip(processor, model, image, question):
+    # prepare image + question
+    inputs = processor(images=image, text=question, return_tensors="pt")
+    
+    generated_ids = model.generate(**inputs, max_length=50)
+    generated_answer = processor.batch_decode(generated_ids, skip_special_tokens=True)
+   
+    return generated_answer
+
+
+def generate_answer_vilt(processor, model, image, question):
+    # prepare image + question
+    encoding = processor(images=image, text=question, return_tensors="pt")
+
+    with torch.no_grad():
+        outputs = model(**encoding)
+
+    predicted_class_idx = outputs.logits.argmax(-1).item()
+    
+    return model.config.id2label[predicted_class_idx]
+
+
+def generate_answers(image, question):
+    answer_git_base = generate_answer_git(git_processor_base, git_model_base, image, question)
+
+    answer_git_large = generate_answer_git(git_processor_large, git_model_large, image, question)
+
+    answer_blip_base = generate_answer_blip(blip_processor_base, blip_model_base, image, question)
+
+    answer_blip_large = generate_answer_blip(blip_processor_large, blip_model_large, image, question)
+
+    answer_vilt = generate_answer_vilt(vilt_processor, vilt_model, image, question)
+
+    return answer_git_base, answer_git_large, answer_blip_base, answer_blip_large, answer_vilt
+
+   
+examples = [["cats.jpg", "How many cats are there?"], ["stop_sign.png", "What's behind the stop sign?"], ["astronaut.jpg", "What's the astronaut riding on?"]]
+outputs = [gr.outputs.Textbox(label="Answer generated by GIT-base"), gr.outputs.Textbox(label="Answer generated by GIT-large"), gr.outputs.Textbox(label="Answer generated by BLIP-base"), gr.outputs.Textbox(label="Answer generated by BLIP-large"), gr.outputs.Textbox(label="Answer generated by ViLT")] 
+
+title = "Interactive demo: comparing visual question answering (VQA) models"
+description = "Gradio Demo to compare GIT, BLIP and ViLT, 3 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."
+article = "<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>"
+
+interface = gr.Interface(fn=generate_answers, 
+                         inputs=[gr.inputs.Image(type="pil"), gr.inputs.Textbox(label="Question")],
+                         outputs=outputs,
+                         examples=examples, 
+                         title=title,
+                         description=description,
+                         article=article, 
+                         enable_queue=True)
+interface.launch(debug=True)

requirements.txt

@@ -0,0 +1,2 @@
+git+https://github.com/huggingface/transformers.git
+torch