Update app.py

app.py

@@ -124,8 +124,8 @@ def stage_3_4_gguf_quantize(model_path_or_id: str, original_model_id: str, quant
     f16_gguf_path = os.path.join(gguf_path, "model-f16.gguf")
     quantized_gguf_path = os.path.join(gguf_path, "model.gguf")
     try:
-        convert_command = ["python3", str(LLAMA_CPP_CONVERT_SCRIPT), model_path_or_id, "--outfile", f16_gguf_path, "--outtype", "f16"]
-        process = subprocess.run(convert_command, check=True, capture_output=True, text=True)
+        convert_command = ["python3", "convert.py", model_path_or_id, "--outfile", f16_gguf_path, "--outtype", "f16"]
+        process = subprocess.run(convert_command, check=True, capture_output=True, text=True, cwd=str(LLAMA_CPP_DIR))
         log_stream += f"Executing llama.cpp conversion script on '{model_path_or_id}'...\n{process.stdout}\n"
         if process.stderr: log_stream += f"[STDERR]\n{process.stderr}\n"
         quantize_map = {"q4_k_m": "Q4_K_M", "q5_k_m": "Q5_K_M", "q8_0": "Q8_0", "f16": "F16"}
@@ -135,8 +135,8 @@ def stage_3_4_gguf_quantize(model_path_or_id: str, original_model_id: str, quant
             os.rename(f16_gguf_path, quantized_gguf_path)
         else:
             log_stream += f"Quantizing FP16 GGUF to {target_quant_name}...\n"
-            quantize_command = [str(LLAMA_CPP_QUANTIZE_SCRIPT), f16_gguf_path, quantized_gguf_path, target_quant_name]
-            process = subprocess.run(quantize_command, check=True, capture_output=True, text=True)
+            quantize_command = ["./quantize", f16_gguf_path, quantized_gguf_path, target_quant_name]
+            process = subprocess.run(quantize_command, check=True, capture_output=True, text=True, cwd=str(LLAMA_CPP_DIR))
             log_stream += f"{process.stdout}\n"
             if process.stderr: log_stream += f"[STDERR]\n{process.stderr}\n"
             os.remove(f16_gguf_path)
@@ -204,7 +204,7 @@ def run_amop_pipeline(model_id: str, pipeline_type: str, do_prune: bool, prune_p
             raise ValueError("Invalid pipeline type selected.")
         full_log += log
         yield {final_output: "Packaging & Uploading (4/5)", log_output: full_log}
-        final_message, log = stage_5_package_and_upload(model_id, optimized_path, full_log, options)
+        final_message, log = stage_5_package_and_upload(model_id, optimized_model_path, full_log, options)
         full_log += log
         yield {final_output: gr.update(value="SUCCESS", label="Status"), log_output: full_log, success_box: gr.Markdown(f"✅ **Success!** Model available: [{repo_id_for_link}](https://huggingface.co/{repo_id_for_link})", visible=True), run_button: gr.Button(interactive=True, value="Run Optimization Pipeline", variant="primary"), analyze_button: gr.Button(interactive=True, value="Analyze Model")}
     except Exception as e: