a

app.py

@@ -0,0 +1,38 @@
+import gradio as gr
+from transformers import ViltProcessor, ViltForQuestionAnswering
+import torch
+
+torch.hub.download_url_to_file('http://images.cocodataset.org/val2017/000000039769.jpg', 'cats.jpg')
+
+processor = ViltProcessor.from_pretrained("dandelin/vilt-b32-finetuned-vqa")
+model = ViltForQuestionAnswering.from_pretrained("dandelin/vilt-b32-finetuned-vqa")
+
+def answer_question(image, text):
+    encoding = processor(image, text, return_tensors="pt")
+    
+    # forward pass
+    with torch.no_grad():
+     outputs = model(**encoding)
+     
+    logits = outputs.logits
+    idx = logits.argmax(-1).item()
+    predicted_answer = model.config.id2label[idx]
+   
+    return predicted_answer
+   
+image = gr.inputs.Image(type="pil")
+question = gr.inputs.Textbox(label="Question")
+answer = gr.outputs.Textbox(label="Predicted answer")
+examples = [["cats.jpg", "How many cats are there?"]]
+
+title = "Cross-lingual VQA"
+description = "ViLT (Vision and Language Transformer), fine-tuned on VQAv2 "
+
+interface = gr.Interface(fn=answer_question, 
+                         inputs=[image, question], 
+                         outputs=answer, 
+                         examples=examples, 
+                         title=title,
+                         description=description,
+                         enable_queue=True)
+interface.launch(debug=True)

requirements.txt

@@ -0,0 +1,3 @@
+gradio
+torch
+git+https://github.com/huggingface/transformers.git