baihuaxie/ConvLab

Issues

• fix MNIST data loading channel (1 instead of 3) bug

  #24 opened 4 years ago
  0

• (flow) add automatic mixed-precision (FP16/32) training

  #23 opened 4 years ago
  0

• (flow) add utility for gradient checkpointing

  #22 opened 4 years ago
  0

• (flow) add utility to track GPU memory usage

  #21 opened 4 years ago
  0

• (flow) resolve .gitignore not ommiting .vscode/settings.json issue

  #19 opened 4 years ago
  0

• test launch training with ImageNet

  #17 opened 4 years ago
  0

• resolve CIFAR-100 training error in flow

  #16 opened 4 years ago
  0

• train ResNet-50 on ImageNet to >= 70% top-1

  #14 opened 4 years ago
  0

• Inspect ImageNet samples

  #13 opened 4 years ago
  1

• implement & train >= 2 contrastive learning frameworks

  #12 opened 4 years ago
  0

• train >= 3 semantic segmentation networks

  #11 opened 4 years ago
  0

• train >= 3 object detection networks on COCO or PASCAL

  #10 opened 4 years ago
  0

• train >= 10 CNN's on ImageNet to >= 70% top-1

  #9 opened 4 years ago
  0

• implement Lambda-ResNet

  #8 opened 4 years ago
  0

• validate training flow

  #7 opened 4 years ago
  0

• add plotting utilities

  #6 opened 4 years ago
  0

• update training flow with experiment launch utilities

  #5 opened 4 years ago
  0

• update readme.md

  #4 opened 4 years ago
  0

• add several small-scale datasets

  #3 opened 4 years ago
  0

• fix tqdm progress bar training loss stagnation issue

  #2 opened 4 years ago
  0

• update training flow for utilites of data inspection

  #1 opened 4 years ago
  0