Switch to Llama 3.2 3B Instruct Q4_K_M

app.py

@@ -1,13 +1,13 @@
 import gradio as gr
 from ctransformers import AutoModelForCausalLM
 
-MODEL_REPO = "Qwen/Qwen2.5-3B-Instruct-GGUF"
-MODEL_FILE = "qwen2.5-3b-instruct-q4_k_m.gguf"
+MODEL_REPO = "hugging-quants/Llama-3.2-3B-Instruct-Q4_K_M-GGUF"
+MODEL_FILE = "llama-3.2-3b-instruct-q4_k_m.gguf"
 
 llm = AutoModelForCausalLM.from_pretrained(
     MODEL_REPO,
     model_file=MODEL_FILE,
-    model_type="qwen",
+    model_type="llama",
     gpu_layers=0,
     context_length=8192,
 )