mismatch in the configuration file with the provided weights

[hirunima]

I was trying to use Partitioned_COI_RS101_2x and 'Partitioned_COI_RS101_2x.yaml' need to be changed as follows:

MULTI_DATASET: DATASETS: ['objects365', 'coco', 'oid'] NUM_CLASSES: [ 365, 80, 500]

Otherwise the weight won't load correctly and would showup as following warning
skip loading parameter 'roi_heads.box_predictor.1.cls_score.1.bias' to the model due to incompatible shapes: (81,) in the checkpoint but (366,) which inturn would results in zero instance detected.