chenhao2345 / ICE

Implementation for ICCV 2021 paper "ICE: Inter-instance Contrastive Encoding for Unsupervised Person Re-identification"
MIT License
44 stars 15 forks source link

Following Issues during execution #21

Open vibhugupta1 opened 1 week ago

vibhugupta1 commented 1 week ago

Dear Author, Hope you are fine. Kindly look into the following issues

  1. The code is working only when iters is set -1. Otherwise the following error is showing: raise ValueError('Expected more than 1 value per channel when training, got input size {}'.format(size)) ValueError: Expected more than 1 value per channel when training, got input size torch.Size([1, 2048])

  2. When iters is set to -1, then mAP, top1, top5, top10 scores not increasing and fixed from epoch 1 to last epoch (40). For dukemtmc-reid dataset: Mean AP: 1.8%, top-1: 5.6%, top-5: 11.5%, top-10: 14.9% For market-1501 dataset: Mean AP: 2.6%, top-1: 8.1%, top-5: 17.5%, top-10: 23.6%

  3. raise StopIteration, StopIteration error, whenever batch size is large.