Update main.py

main.py

@@ -1,4 +1,9 @@
 import os
+
+# Set cache directories to a writable location
+os.environ["TRANSFORMERS_CACHE"] = "/tmp/hf"
+os.environ["HF_HOME"] = "/tmp/hf"
+
 from langchain.agents import initialize_agent
 from langchain.llms import HuggingFacePipeline
 from langchain.tools import BaseTool
@@ -7,13 +12,13 @@ from transformers import pipeline
 # Set up the Hugging Face pipeline to run on CPU (device=-1 ensures CPU usage)
 pipe = pipeline(
     "text-generation",
-    model="bigscience/bloom",       # You can change to another CPU-friendly model if desired
-    tokenizer="bigscience/bloom",
-    device=-1,                      # -1 forces CPU mode
+    model="bigscience/bloom-560m",       # Use a smaller, CPU-friendly model
+    tokenizer="bigscience/bloom-560m",
+    device=-1,                      # Force CPU mode
     max_new_tokens=512
 )
 
-# Initialize the LangChain LLM with our Hugging Face pipeline
+# Initialize the LangChain LLM with our HuggingFace pipeline
 llm = HuggingFacePipeline(pipeline=pipe)
 
 # Define a custom tool that compiles the research report