Update app.py

app.py

@@ -34,10 +34,10 @@ optimal_threads = max(4, cpu_count - 1)  # Leave one core free
 print(f"Using {optimal_threads} of {cpu_count} CPU cores")
 
 # Download model files
-def get_model_path(repo_id, filename):
+def get_model_path(repo_id):
     print(f"Obtaining {filename}...")
     # Download to our custom cache location
-    return hf_hub_download(repo_id=repo_id, filename=filename, cache_dir=MODEL_CACHE)
+    return hf_hub_download(repo_id=repo_id, cache_dir=MODEL_CACHE)
 
 # Function to quantize model to int4 or int8
 def quantize_model(input_model_path, output_model_path, quantization_type="q4_0"):
@@ -68,12 +68,10 @@ def quantize_model(input_model_path, output_model_path, quantization_type="q4_0"
 
 # Download models
 base_model_path = get_model_path(
-    "johnpaulbin/articulate-11-expspanish-base-merged-Q8_0-GGUF", 
-    "articulate-11-expspanish-base-merged-q8_0.gguf"
+    "johnpaulbin/articulate-11-expspanish-base-merged"
 )
 adapter_path = get_model_path(
-    "johnpaulbin/articulate-V1-Q8_0-GGUF", 
-    "articulate-V1-q8_0.gguf"
+    "johnpaulbin/articulate-V1"
 )
 
 # Quantize models (creates int4 versions for faster CPU inference)