Update app.py

app.py

@@ -1,18 +1,18 @@
 import streamlit as st
-from transformers import pipeline, AutoTokenizer, AutoModelForCausalLM
+import torch
+from transformers import AutoTokenizer, AutoModelForCausalLM
 
-# Load the tokenizer and model
+# Load the tokenizer and model using PyTorch
 tokenizer = AutoTokenizer.from_pretrained("MohamedMotaz/Examination-llama-8b-4bit")
-model = AutoModelForCausalLM.from_pretrained("MohamedMotaz/Examination-llama-8b-4bit")
-pipe = pipeline("text-generation", model=model, tokenizer=tokenizer)
+model = AutoModelForCausalLM.from_pretrained("MohamedMotaz/Examination-llama-8b-4bit", torch_dtype=torch.float16).to("cuda" if torch.cuda.is_available() else "cpu")
 
 # App Title
-st.title("Exam Corrector: Automated Grading with LLama 8b Model")
+st.title("Exam Corrector: Automated Grading with LLama 8b Model (PyTorch)")
 
 # Instructions
 st.markdown("""
 ### Instructions:
-- Upload or type both the **Model Answer** and the **Student Answer**.
+- Enter both the **Model Answer** and the **Student Answer**.
 - Click on the **Grade Answer** button to get the grade and explanation.
 """)
 
@@ -52,12 +52,15 @@ with st.expander("Click to View Documentation"):
 if st.button("Grade Answer"):
     # Combine inputs into the required prompt format
     inputs = f"Model Answer: {model_answer}\n\nStudent Answer: {student_answer}\n\nResponse:"
-
-    # Tokenize the inputs
-    input_ids = tokenizer(inputs, return_tensors="pt").input_ids
     
-    # Generate the response using the model
-    outputs = model.generate(input_ids)
+    # Tokenize the inputs using PyTorch tensors
+    input_ids = tokenizer(inputs, return_tensors="pt").input_ids.to(model.device)
+    
+    # Generate the response using the model (PyTorch)
+    with torch.no_grad():
+        outputs = model.generate(input_ids, max_length=200)
+    
+    # Decode the output
     response = tokenizer.decode(outputs[0], skip_special_tokens=True)
 
     # Display the grade and explanation