Update app.py

app.py

@@ -5,10 +5,6 @@ from huggingface_hub import snapshot_download
 
 import subprocess
 
-# Recompile llama.cpp before running
-subprocess.run(["make", "clean"], cwd="/home/user/app/llama.cpp", check=True)
-subprocess.run(["make"], cwd="/home/user/app/llama.cpp", check=True)
-
 def check_directory_path(directory_name: str) -> str:
     if os.path.exists(directory_name):
         path = os.path.abspath(directory_name)
@@ -39,7 +35,7 @@ def convert_to_gguf(model_dir, output_file):
     os.makedirs(os.path.dirname(output_file), exist_ok=True)
     st.write(model_dir_path)
     cmd = [
-    "python3", f"{model_dir_path}/convert_hf_to_gguf.py", model_dir,
+    "python3", "/app/llama.cpp/convert_hf_to_gguf.py", model_dir,
     "--outtype", "f16", "--outfile", output_file
     ]
     process = subprocess.run(cmd, text=True, stdout=subprocess.PIPE, stderr=subprocess.PIPE)
@@ -54,11 +50,11 @@ def quantize_llama(model_path, quantized_output_path, quant_type):
     """
     st.write(f"⚡ Quantizing `{model_path}` with `{quant_type}` precision...")
     os.makedirs(os.path.dirname(quantized_output_path), exist_ok=True)
-    quantize_path = f"{model_dir_path}/build/bin/llama-quantize"
+    quantize_path = "/app/llama.cpp/build/bin/llama-quantize"
     subprocess.run(["chmod", "+x", quantize_path], check=True)
     
     cmd = [
-    f"{model_dir_path}/build/bin/llama-quantize", 
+    "/app/llama.cpp/build/bin/llama-quantize", 
     model_path, 
     quantized_output_path,
     quant_type