Update quick_start output behavior and trim README args

README.md

@@ -60,9 +60,9 @@ python quick_start/run_example.py \
 - `--task-instruction`: task description, unless `--prompt` is used instead.
 - `--task-plan`: optional step-by-step plan text for the default `chunk_all` prompt.
 - `--model-path`: optional. Defaults to the parent directory of `quick_start/`.
-- `--output-jsonl`: optional. Defaults to `quick_start/outputs/<video_stem>.jsonl`.
+- `--output-jsonl`: optional. If omitted, the script prints the prediction only and does not save a JSONL file.
 - `--prompt`: optional full prompt override.
-- `--sample-hz`, `--max-new-tokens`: optional inference controls.
+- `--sample-hz`: optional inference control. Default: `2.0`.
 
 Render a prediction-only preview video from the JSONL output:
 

quick_start/README.md

@@ -36,9 +36,8 @@ Important arguments:
 - `--task-instruction`: optional if `--prompt` is provided. Natural-language task description.
 - `--task-plan`: optional. Multi-line plan text appended after `--task-instruction` when building the default `chunk_all` prompt.
 - `--model-path`: optional. Defaults to the parent directory of `quick_start/`.
-- `--output-jsonl`: optional. Defaults to `quick_start/outputs/<video_stem>.jsonl`.
+- `--output-jsonl`: optional. If omitted, the script prints the prediction only and does not save a JSONL file.
 - `--sample-hz`: optional. Video sampling rate for inference. Default: `2.0`.
-- `--max-new-tokens`: optional. Generation cap. Default: `1024`.
 - `--prompt`: optional full prompt override. If set, `--task-instruction` and `--task-plan` are ignored.
 
 Render a preview video:

quick_start/run_example.py

@@ -68,7 +68,7 @@ def parse_args() -> argparse.Namespace:
         "--output-jsonl",
         type=Path,
         default=None,
-        help="Optional output path. Defaults to quick_start/outputs/<video_stem>.jsonl",
+        help="Optional output path. If omitted, print the response only and do not save a JSONL file.",
     )
     parser.add_argument(
         "--max-new-tokens",
@@ -282,8 +282,9 @@ def main() -> int:
     prompt, prompt_source, task_instruction, task_plan = resolve_prompt(args)
     video_path = resolve_video_path(args.video_path)
     video_path_for_output = maybe_relativize_to_model_root(video_path)
-    output_path = args.output_jsonl or (THIS_DIR / "outputs" / f"{video_path.stem}.jsonl")
-    output_path.parent.mkdir(parents=True, exist_ok=True)
+    output_path = args.output_jsonl.resolve() if args.output_jsonl is not None else None
+    if output_path is not None:
+        output_path.parent.mkdir(parents=True, exist_ok=True)
 
     InferRequest, PtEngine, RequestConfig = load_swift_runtime()
     engine = PtEngine(
@@ -341,10 +342,11 @@ def main() -> int:
             "pred_curve_point_progress": pred_point_progress,
         }
     )
-    output_path.write_text(json.dumps(output_row, ensure_ascii=False) + "\n", encoding="utf-8")
+    if output_path is not None:
+        output_path.write_text(json.dumps(output_row, ensure_ascii=False) + "\n", encoding="utf-8")
 
     print(f"video_path={video_path}")
-    print(f"output_jsonl={output_path.resolve()}")
+    print(f"output_jsonl={output_path if output_path is not None else '(not saved)'}")
     print(f"input_sample_hz={float(args.sample_hz):.4f}")
     print(f"input_frame_count={len(sampled_indices)}")
     print(f"pred_keypoint_count={len(pred_point_progress)}")