Model

The model subcommand is used to generate and train models that can be evaluated and used for different tasks.

Available commands

stray model generate

Options

namedefaultchoicesdescription
--model-typedetectron2detectron2The model type to use. Currently only detectron2 is supported
--repositoryWhere to save the newly created models
--help, -hShow help

stray model bake

  • "Bakes" a dataset into a given model. Saves the model as model.pth into the model directory. The model training happens inside a Docker container.

Options

namedefaultchoicesdescription
<scenes>Paths to the scenes which are used in model training
--modelPath to the model to be used in baking
--num-gpus0Number of GPUs to use in baking
--resumeFalse(flag)Resume training from previous run
--segmentationFalse(flag)Train a segmentation model, requires segmentation masks to exist (masks can be created with stray dataset bake
--bbox-from-maskFalse(flag)Determine the 2D bounding boxes from segmentation masks, requires segmentation masks to exist (masks can be created with stray dataset bake
--help, -hShow help

stray model evaluate

  • Evaluates model performance against the given evaluation dataset

Options

namedefaultchoicesdescription
<scenes>Paths to the scenes to use for evaluation
--modelPath to the model to evaluate
--weightsmodel/output/model_final.pthPath to the weights file to use for evaluation
--threshold0.7Prediction confidence threshold
--help, -hShow help

Help

Visit our issue tracker for help and direct support.