lxuechen/private-transformers

Issues

• Feature Request: Context Manager for training mixed private and public data

  #40 opened 3 months ago by bnestor
  0

• Can not train with a single GPU

  #39 opened 8 months ago by CHAOS-Yang
  1

• Table2text support for facebook/opt models

  #38 opened a year ago by zredlined
  1

• RuntimeError: CUDA error: device-side assert triggered CUDA kernel errors might be asynchronously reported at some other API call,so the stacktrace below might be incorrect.

  #37 opened a year ago by JeffffffFu
  3

• no metrics: BLEU or Rouge-L

  #36 opened a year ago by adam-dziedzic
  2

• text infilling task

  #34 opened a year ago by adam-dziedzic
  2

• table2text

  #35 opened a year ago by adam-dziedzic
  2

• Support for multi-gpu private fine-tuning

  #32 opened a year ago by Pier297
  2

• compatibility issue with transformers

  #33 opened a year ago by kshll6
  0

• What is the best way to handle large models?

  #31 opened a year ago by Pier297
  2

• v0.4.0 fixes

  #24 opened 2 years ago by lxuechen
  0

• Customize loss function / adding regularizer under privacy setting?

  #30 opened 2 years ago by shi-kejian
  4

• classification example fails with the latest release

  #28 opened 2 years ago by lxuechen
  1

• No such file or directory

  #27 opened 2 years ago by trestad
  2

• v0.3.0 fixes

  #22 opened 2 years ago by lxuechen
  0

• v0.2.2 fixes

  #25 opened 2 years ago by lxuechen
  0

• Support Poisson sampling + profile speed

  #11 opened 2 years ago by lxuechen
  0

• Support BART model

  #12 opened 2 years ago by SeolhwaLee
  7

• v0.2.0 refactor, fixes, docs, additional features

  #21 opened 2 years ago by lxuechen
  0

• Refine spectrum eval

  #17 opened 2 years ago by lxuechen
  0

• How to set max_compositions

  #19 opened 2 years ago by hlzhang109
  3

• Setting small target epsilon like 0.1 fails training

  #15 opened 2 years ago by LinkToPast1900
  3

• [DistilBERT] RuntimeError: stack expects each tensor to be equal size

  #14 opened 2 years ago by LinkToPast1900
  1

• Issue with privacy_engine

  #13 opened 2 years ago by acphile
  1

• Set another seed won't change the result

  #10 opened 2 years ago by JunyiZhu-AI
  6

• Private gradient seemingly has been overwritten by non-private gradient.

  #9 opened 2 years ago by JunyiZhu-AI
  3

• Questions about sigma search and epsilon from composed tradeoff functions

  #8 opened 2 years ago by xplip
  3

• Using dataloader with fixed batch size

  #7 opened 3 years ago by xplip
  4

• Function for "get_privacy_stats()"

  #1 opened 3 years ago by XuandongZhao
  1