Update app.py

app.py

@@ -6,9 +6,11 @@ app = FastAPI()
 
 # Download and initialize the model when the server starts
 llm = Llama.from_pretrained(
-    repo_id="Qwen/Qwen2.5-Coder-1.5B-Instruct-GGUF",
-    filename="*q4_k_m.gguf", # 4-bit quantization for speed and low memory
-    n_ctx=2048 # Context window size
+    repo_id="Qwen/Qwen2.5-Coder-0.5B-Instruct-GGUF", # Make sure this says 0.5B!
+    filename="*q4_k_m.gguf", 
+    n_ctx=2048, 
+    n_threads=2, 
+    n_batch=512  
 )
 
 class EvalRequest(BaseModel):