Update app.py

app.py

@@ -3,15 +3,15 @@ import requests
 from bs4 import BeautifulSoup
 from transformers import pipeline
 
-# Load Mixtral model (via Transformers pipeline)
-refiner = pipeline("text2text-generation", model="mistralai/Mixtral-8x7B-Instruct-v0.1")
+# Use open-access model
+refiner = pipeline("text2text-generation", model="google/flan-t5-large")
 
 def refine_from_url(url, instruction):
     try:
         response = requests.get(url, timeout=5)
         soup = BeautifulSoup(response.text, "html.parser")
         raw_text = soup.get_text(separator="\n")
-        prompt = f"{instruction}\n\n{raw_text[:4000]}"  # truncate for token limit
+        prompt = f"{instruction}\n\n{raw_text[:3000]}"  # truncate for token limit
         output = refiner(prompt, max_new_tokens=512)[0]["generated_text"]
         return output
     except Exception as e:
@@ -24,8 +24,8 @@ demo = gr.Interface(
         gr.Textbox(label="Refinement Instruction", placeholder="e.g. Clean and structure this for AI training")
     ],
     outputs=gr.Textbox(label="Refined Output"),
-    title="🔍 Data Refiner with Mixtral",
-    description="Crawl any webpage and refine its content using Mixtral 8x7B for AI training or research."
+    title="🧠 Data Refiner with Flan-T5",
+    description="Crawl any webpage and refine its content using Flan-T5 for AI training or research."
 )
 
 if __name__ == "__main__":