Update app.py

app.py

@@ -2,10 +2,20 @@ import streamlit as st
 from transformers import AutoModelForCausalLM, AutoTokenizer
 import torch
 
+# Disable safetensors fast GPU loading (if needed)
+import os
+os.environ["SAFETENSORS_FAST_GPU"] = "0"
+
+# Cache the model and tokenizer
+@st.cache_resource
+def load_model_and_tokenizer():
+    model_name = "rajrakeshdr/IntelliSoc"
+    tokenizer = AutoTokenizer.from_pretrained(model_name)
+    model = AutoModelForCausalLM.from_pretrained(model_name, use_safetensors=False)
+    return model, tokenizer
+
 # Load the model and tokenizer
-model_name = "rajrakeshdr/IntelliSoc"
-tokenizer = AutoTokenizer.from_pretrained(model_name)
-model = AutoModelForCausalLM.from_pretrained(model_name)
+model, tokenizer = load_model_and_tokenizer()
 
 # Streamlit app title
 st.title("IntelliSoc Text Generation")
@@ -13,9 +23,6 @@ st.title("IntelliSoc Text Generation")
 # Input prompt
 prompt = st.text_area("Enter your prompt:", "Once upon a time")
 
-# Slider for max length
-max_length = st.slider("Max length of generated text", 50, 200, 100)
-
 # Generate text on button click
 if st.button("Generate Text"):
     # Tokenize input
@@ -25,7 +32,7 @@ if st.button("Generate Text"):
     with torch.no_grad():
         outputs = model.generate(
             inputs.input_ids,
-            max_length=max_length,
+            max_length=100,
             num_return_sequences=1,
             no_repeat_ngram_size=2,
             top_k=50,