How to train for Video ReID dataset?

[FatimaZulfiqar]

Hi it might be a simple question but I need assistance regarding training model for video reid dataset. I have specified an engine as follows:

datamanager = torchreid.data.ImageDataManager( root='/content/DukeMTMC-VideoReID', sources='dukemtmcvidreid', targets='dukemtmcvidreid', height=256, width=128, batch_size_train=32, batch_size_test=100, transforms=['random_flip', 'random_crop'] )

but it is giving me following error

ValueError Traceback (most recent call last) in () 6 width=128, 7 batch_size_train=32, ----> 8 batch_size_test=100, 9 #transforms=['random_flip', 'random_crop'] 10 )

1 frames /content/Final/deep-person-reid/torchreid/data/datasets/init.py in init_image_dataset(name, **kwargs) 37 raise ValueError( 38 'Invalid dataset name. Received "{}", ' ---> 39 'but expected to be one of {}'.format(name, avai_datasets) 40 ) 41 return __image_datasetsname

ValueError: Invalid dataset name. Received "dukemtmcvidreid", but expected to be one of ['market1501', 'cuhk03', 'dukemtmcreid', 'msmt17', 'viper', 'grid', 'cuhk01', 'ilids', 'sensereid', 'prid', 'cuhk02', 'university1652']

need assistance in this regard

[thuangb]

Just change ImageDataManager to VideoDataManager. And you have to prepare the dataset yourseld also