diff --git a/quantize_streaming.py b/quantize_streaming.py index 29305d6..b6d9a19 100644 --- a/quantize_streaming.py +++ b/quantize_streaming.py @@ -90,7 +90,7 @@ def streaming_quantize(model_path: str, output_path: str): if __name__ == "__main__": parser = argparse.ArgumentParser() parser.add_argument("--model-path", type=str, default="/data/models/Ornith-1.0-35B") - parser.add_argument("--output-path", type=str, default="/data/models/Ornith-1.0-35B-4bit-streaming") + parser.add_argument("--output-path", type=str, default="/data/models/Ornith-1.0-35B-nf4") args = parser.parse_args() streaming_quantize(args.model_path, args.output_path)