Command-line Interface¶
DREEM provides several types of functionality accessible through a command prompt.
Training¶
dreem-train¶
dreem-train is the command-line interface for training. Use this for training on a remote machine/cluster/colab notebook instead of using the API directly.
Usage¶
usage: dreem-train [-h] [--hydra-help] [--config-dir] [--config-name] [+params_config] [+batch_config]
positional arguments:
--config-dir Path to configs dir
--config-name Name of the .yaml config file stored in config-dir without the .yaml extension
optional arguments:
-h, --help Shows the application's help and exit.
--hydra-help Shows Hydra specific flags (recommended over -h)
+params_config Path to .yaml file containing subset of params to override
+batch_config Path to .csv file where each row indicates params to override for a single task in a batch job
See the usage guide for a more in-depth explanation on how to use dreem-train and see the training config walkthrough for all available parameters.
dreem-track¶
dreem-track is the command-line interface for inference. Use this for tracking using a pretrained model on a remote machine/cluster/notebook instead of using the API directly.
Usage¶
usage: dreem-track [-h] [--hydra-help] [--config-dir] [--config-name] [+params_config] [+batch_config]
positional arguments:
--config-dir Path to configs dir
--config-name Name of the .yaml config file stored in config-dir without the .yaml extension
optional arguments:
-h, --help Shows the application's help and exit.
--hydra-help Shows Hydra specific flags (recommended over -h)
+params_config Path to .yaml file containing subset of params to override
+batch_config Path to .csv file where each row indicates params to override for a single task in a batch job
dreem-track and see the inference config walkthrough for all available parameters.
dreem-visualize¶
dreem-visualize is the command-line utility for generating annotated videos based on DREEM output
Usage¶
usage: dreem-visualize [-h] [--hydra-help] [+labels_path] [+vid_path] [+save_path] [+annotate.key] [--config-path] [+annotate.color_palette="tab20"] [+annotate.trails=2] [+annotate.boxes=64] [+annotate.names=True] [+annotate.track_scores=0.5] [+annotate.centroids=4] [+annotate.poses=False] [+annotate.fps=30] [+annotate.alpha=0.2]
positional arguments:
+labels_path The path to the dreem tracks as a `.csv` file containing the following keys:
- "X": the X position of the instance in pixels
- "Y": the Y position of the instance in pixels
- "Frame_id": The frame of the video in which the instance occurs
- "Pred_track_id": The track id output from DREEM
- "Track_score": The trajectory score output from DREEM
- (Optional) "Gt_track_id": the gt track id for the instance (mostly used for debugging)
where each row is an instance. See `dreem.inference.track.export_trajectories`
+vid_path The path to the video file in an `imageio`-accepted format (See: https://imageio.readthedocs.io/en/v2.4.1/formats.html)
+save_path The path to where you want to store the annotated video (e.g "./tracked_videos/dreem_output.mp4")
+annotate.key The key where labels are stored in the dataframe - mostly used for choosing whether to annotate based on pred or gt labels
optional arguments
-h, --help Shows the application's help and exit.
--hydra-help Shows Hydra specific flags (recommended over -h)
--config-path The path to the config .yaml file containing params for annotating. We recommend using this file in lieu of any "annotate.*" arguments.
+annotate.color_palette The matplotlib colorpalette to use for annotating the video. Defaults to `tab10`
+annotate.trails The size of the track trails. If the trail size <=0 or none then it is not used.
+annotate.boxes The size of the bbox. If bbox size <= 0 or None then it is not added
+annotate.names Whether or not to annotate with name
+annotate.centroids The size of the centroid. If centroid size <=0 or None then it is not added.
+annotate.poses Whether or not to annotate with poses
+annotate.fps The frames-per-second of the annotated video
+annotate.alpha The opacity of the centroid annotations.
As noted above, instead of specifying each individual +annotate.* param we recommend setting up a visualize config (e.g ./configs/visualize.yaml) which looks like:
annotate:
color_palette: "tab10"
trails: 5
boxes: (64, 64)
names: true
centroids: 10
poses: false
fps: 30
alpha: 0.2
and then running: