Update app.py

app.py

@@ -1,21 +1,14 @@
+from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
+import requests
 import PyPDF2
-import torch
 import gradio as gr
 
-# Make sure the model path is correct for your system!
-llm = LlamaCpp(
-    model_path="./llama-2-7b-chat.Q4_K_S.gguf",
-    temperature=0.2,
-    n_ctx=512,
-    max_tokens=2000,
-    top_p=1,
-)
+# Replace with your Hugging Face API token
+api_token = "YOUR_HUGGING_FACE_TOKEN"
+mistral_model_id = "NousResearch/Hermes-2-Pro-Mistral-7B"  # Choose appropriate model version
 
-template = """Generate only one MCQ question based on text \
-that is delimited by triple backticks \
-with {pattern} pattern. \
-text: `{text}` \
-"""
+tokenizer = AutoTokenizer.from_pretrained(mistral_model_id)
+model = AutoModelForSeq2SeqLM.from_pretrained(mistral_model_id)
 
 
 def extract_paragraphs(pdf_file):
@@ -49,13 +42,27 @@ def extract_paragraphs(pdf_file):
 def Generate_mcq_from_pdf(pdf_file):
     paragraphs = extract_paragraphs(pdf_file)
     for para in paragraphs:
-        input_msg = PromptTemplate.from_template(template=template)
-        input_s = input_msg.format(pattern=pattern, text=para)
-        output_msg = llm(input_s)
+        template = """Generate only one MCQ question based on text \
+        that is delimited by triple backticks \
+        with {pattern} pattern. \
+        text: `{text}` \
+        """
+        prompt = template.format(pattern="IIT GATE", text=para)
+        inputs = tokenizer(prompt, return_tensors="pt")
+
+        headers = {"Authorization": f"Bearer {api_token}"}
+        url = f"https://api-inference.huggingface.co/models/{mistral_model_id}"
+        response = requests.post(url, headers=headers, json=inputs)
+        response.raise_for_status()  # Raise an error if request fails
+
+        output_ids = response.json()["generated_ids"]
+        output_text = tokenizer.batch_decode(output_ids, skip_special_tokens=True)[0]
+
         output_file = "questions.txt"
         with open(output_file, "w") as f:
-            f.write(output_msg)
-    return output_msg, output_file
+            f.write(output_text)
+
+    return output_text, output_file
 
 
 app = gr.Interface(