Update app.py

app.py

@@ -1,12 +1,41 @@
-from transformers import BlipForQuestionAnswering
-from transformers import AutoProcessor
+from transformers import BlipForQuestionAnswering, AutoProcessor
 from PIL import Image
 import gradio as gr
-model=BlipForQuestionAnswering.from_pretrained("Salesforce/blip-vqa-base")
-processor=AutoProcessor.from_pretrained("Salesforce/blip-vqa-base")
-def qna(image,question):
-  inputs=processor(image,question,return_tensors="pt")
-  out=model.generate(**inputs)
-  return processor.decode(out[0],skip_special_tokens=True)
-interf=gr.Interface(qna,inputs=["image","text"],outputs="text")
-interf.launch()
+import requests  # To call Groq API
+
+# Load the BLIP model and processor
+model = BlipForQuestionAnswering.from_pretrained("Salesforce/blip-vqa-base")
+processor = AutoProcessor.from_pretrained("Salesforce/blip-vqa-base")
+
+# Define Groq API details (Replace with actual API endpoint and key)
+GROQ_API_ENDPOINT = "https://api.groq.com/v1/completions"
+GROQ_API_KEY = "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"  # Replace with your actual Groq API key
+
+# Function to generate the initial answer with BLIP and expand it with Groq API
+def qna(image, question):
+    # Step 1: Get initial short answer from BLIP
+    inputs = processor(image, question, return_tensors="pt")
+    out = model.generate(**inputs)
+    short_answer = processor.decode(out[0], skip_special_tokens=True)
+
+    # Step 2: Construct prompt for Groq API
+    prompt = f"Question: {question}\nShort Answer: {short_answer}\nProvide a detailed explanation based on this answer."
+
+    # Step 3: Send prompt to Groq API for a paragraph-length answer
+    response = requests.post(
+        GROQ_API_ENDPOINT,
+        headers={"Authorization": f"Bearer {GROQ_API_KEY}"},
+        json={"prompt": prompt, "max_tokens": 200}  # Adjust max_tokens as needed
+    )
+
+    if response.status_code == 200:
+        # Extract generated paragraph from Groq API response
+        detailed_answer = response.json().get("choices")[0].get("text")
+    else:
+        detailed_answer = "Failed to get response from Groq API."
+
+    return detailed_answer
+
+# Create Gradio interface
+interf = gr.Interface(fn=qna, inputs=["image", "text"], outputs="text")
+interf.launch()