Metrics not being shown in execution-run.json after training.
A metric calculation step is being started directly after the training. Training might take longer but in this example a small dataset is being used so the training is as fast as a inference. Although model has logger.info('MericsAvailable') and logging metrics as shown in the screenshot below:
Here is what is being logged:
Logged metric
2023-12-12 08:57:26,409 __main__ INFO {'metrics': {'type': 'coco-metrics/v1', 'date_time': '2023-12-12 08:57:22', 'status_text': 'success', 'more_is_better': {'AP': 0.0, 'AP_50': 0.0}, 'less_is_better': {}}}And here is how the execution-run.json after the training: