Closed chrise96 closed 1 year ago
Batch inference implementation using build_model and DataLoader.
ContainerDataset(Dataset) class is partly based on DefaultPredictor https://github.com/facebookresearch/detectron2/blob/main/detectron2/engine/defaults.py#L252
Results are saved directly to database instead of blob storage -> database
To test the detection results locally, please clone this commit: d3428a07dc5fd354cdb30c3da672d257f695545a
TODO: implement select query to database in postprocessing.py
Batch inference implementation using build_model and DataLoader.
ContainerDataset(Dataset) class is partly based on DefaultPredictor https://github.com/facebookresearch/detectron2/blob/main/detectron2/engine/defaults.py#L252
Results are saved directly to database instead of blob storage -> database
To test the detection results locally, please clone this commit: d3428a07dc5fd354cdb30c3da672d257f695545a
TODO: implement select query to database in postprocessing.py