fast-reid/fastreid/data
liaoxingyu fd90555e19 feat: add multi-dataset joint training
new feature that can support joint training, and find some bugs in funtion combine_all of datasets/bases.py that assume person id in dataset has been relabeld from 0 to num_class.
Another bug appears in msmt17 which trainset and testset person id both begin from 0, and we should change testset id from num_class of trainset.
2020-05-18 20:06:04 +08:00
..
datasets feat: add multi-dataset joint training 2020-05-18 20:06:04 +08:00
samplers fix(data/samplers): fix bug in triplet sampler 2020-04-27 15:25:29 +08:00
transforms feat(engine/defaults): add DefaultPredictor to get image reid features 2020-05-08 19:24:27 +08:00
__init__.py 1. upload circle loss and arcface 2020-04-05 23:54:26 +08:00
build.py feat: add multi-dataset joint training 2020-05-18 20:06:04 +08:00
common.py fix(data): fix resume training bug 2020-05-05 23:20:42 +08:00
data_utils.py Update sampler code 2020-02-10 07:38:56 +08:00