Update app.py

app.py

@@ -17,22 +17,21 @@ if not HUGGINGFACE_API_TOKEN:
 # Title of the app
 st.title("Optimized LLaMA 2 Chatbot")
 
-# Load the smaller LLaMA model and tokenizer from Hugging Face (optimized for smaller size)
+# Load the LLaMA model and tokenizer from Hugging Face
 @st.cache_resource
 def load_model_and_tokenizer():
-    """Load the tokenizer and model with quantization and efficient caching."""
+    """Load the tokenizer and model."""
     tokenizer = AutoTokenizer.from_pretrained(
-        "meta-llama/Llama-2-3b-hf",  # Switch to LLaMA 2 3B model for faster response
+        "meta-llama/Llama-2-7b-hf",  # Correct model identifier
         use_auth_token=HUGGINGFACE_API_TOKEN
     )
 
-    # Load the model with quantization and device_map auto
     model = LlamaForCausalLM.from_pretrained(
-        "meta-llama/Llama-2-3b-hf",  # Use the smaller 3B model
+        "meta-llama/Llama-2-7b-hf",  # Correct model identifier
         use_auth_token=HUGGINGFACE_API_TOKEN,
-        torch_dtype=torch.float16,  # Efficient half precision
-        device_map="auto",  # Auto distribute the model efficiently
-        low_cpu_mem_usage=True  # Optimize memory usage for CPU-based inference
+        torch_dtype=torch.float16,
+        device_map="auto",
+        low_cpu_mem_usage=True
     )
 
     return tokenizer, model
@@ -46,12 +45,12 @@ def generate_text(prompt, tokenizer, model, max_length=256, temperature=0.6):
     with torch.no_grad():
         generate_ids = model.generate(
             inputs.input_ids,
-            max_length=max_length,  # Reduced max length for faster generation
-            temperature=temperature,  # Lower temperature for less computation
+            max_length=max_length,
+            temperature=temperature,
             do_sample=True,
-            top_k=30,  # Lower top_k for faster generation
-            top_p=0.85,  # Adjust top_p to balance diversity and speed
-            repetition_penalty=1.1,  # Adjusted repetition penalty
+            top_k=30,
+            top_p=0.85,
+            repetition_penalty=1.1,
         )
     
     return tokenizer.decode(generate_ids[0], skip_special_tokens=True)
@@ -70,8 +69,8 @@ if st.button("Generate Response"):
 
 # Optional settings in sidebar
 st.sidebar.header("Settings")
-max_length = st.sidebar.slider("Max Length", min_value=50, max_value=512, value=256)  # Reduced default max_length
-temperature = st.sidebar.slider("Temperature", min_value=0.1, max_value=1.0, value=0.6)  # Lower default temperature for speed
+max_length = st.sidebar.slider("Max Length", min_value=50, max_value=512, value=256)
+temperature = st.sidebar.slider("Temperature", min_value=0.1, max_value=1.0, value=0.6)
 
 # Regenerate response with updated settings
 if st.sidebar.button("Regenerate"):