MMDetWandbHookto the MMDetection
log_configmethod of the config system.
MMDetWandbHook. This argument takes in a dictionary of key-value pairs which in turn is passed to
wandb.initwhich controls which project your run is logged to as well as other features of your run.
True, every checkpoint version will have metadata associated with it. This feature depends on the
CheckpointHookas well as
DistEvalHook. The metadata is logged only when the checkpoint interval is divisible by evaluation interval.
MMDetWandbHookyou can now log the validation data as W&B Tables and create versioned W&B Tables for model prediction.
num_eval_imagesargument controls the number of validation samples that are logged as W&B Tables. Here are a few things to note:
num_eval_images=0the validation data, as well as model predictions, will not be logged.
num_eval_imagesis greater than the total number of validation samples, the complete validation dataset is logged.