Update docs/guides/multimodal/video.md

Co-authored-by: Samuel Monson <[email protected]>
Signed-off-by: Mark Kurtz <[email protected]>

Author: markurtz

docs/guides/multimodal/video.md

@@ -153,7 +153,8 @@ guidellm benchmark \
 - `--target`: The base URL of the inference server.
 - `--model`: The model name to use for requests.
 - `--request-type`: chat_completions, supporting multimodal inputs.
-- `--data`: The dataset identifier (lmms-lab/Video-MME).
+- `--data`: The dataset identifier (lmms-lab/Video-MME). 
+- `--data-args`: Configuration for the dataset loading, selecting the "test" split. See [`datasets.load_dataset`](https://huggingface.co/docs/datasets/v4.5.0/en/package_reference/loading_methods#datasets.load_dataset) for full list of valid options.
 - `--data-column-mapper`: Maps the dataset’s `url` column (containing the video link) to `video_column` and `question` to `text_column`.
 
 ### 2. Video Captioning