added initial version

src/app.py

@@ -10,11 +10,12 @@ from fastapi.templating import Jinja2Templates
 from fastapi.middleware.cors import CORSMiddleware
 from fastapi.responses import HTMLResponse
 from fastapi.staticfiles import StaticFiles
+from pathlib import Path
 
 # Get the absolute path to the templates directory
 TEMPLATES_DIR = os.path.join(os.path.dirname(__file__), "templates")
 
-MODEL_ID = "sagargurujula/text-generator"
+MODEL_ID = "sagargurujala/text-generator"
 
 # Initialize FastAPI
 app = FastAPI(title="GPT Text Generator")
@@ -34,13 +35,22 @@ app.add_middleware(
 # Set device
 device = 'cuda' if torch.cuda.is_available() else 'cpu'
 
+# Create cache directory in the current working directory
+cache_dir = Path("model_cache")
+cache_dir.mkdir(exist_ok=True)
+
+# Set environment variable for Hugging Face cache
+os.environ['TRANSFORMERS_CACHE'] = str(cache_dir)
+os.environ['HF_HOME'] = str(cache_dir)
+
 # Load model from Hugging Face Hub
 def load_model():
     try:
-        # Download the model file from HF Hub
+        # Download the model file from HF Hub with custom cache directory
         model_path = hf_hub_download(
             repo_id=MODEL_ID,
-            filename="best_model.pth"
+            filename="best_model.pth",
+            cache_dir=cache_dir
         )
         
         # Initialize our custom GPT model