bigscience-workshop/xmtf

Issues

• Getting machine-translated prompts of xP3mt

  #24 opened 6 months ago by dptam
  4

• Parsing the xP3 dataset

  #23 opened 6 months ago by hmubarak
  1

• Dose mt0&bloomz trained on dev, devtest datasets of Flores-200?

  #22 opened 8 months ago by dsj96
  2

• Export mt0-xxl-mt to ONNX fails

  #21 opened 8 months ago by sh0tcall3r
  2

• bloomz-mt universal checkpoint

  #20 opened a year ago by LiuShixing
  2

• how to repreduce bloomz-*

  #18 opened a year ago by fpcsong
  6

• mT0-xxl finetuning

  #19 opened a year ago by sh0tcall3r
  6

• How to fineutne mT0 with specific down-stream data?

  #12 opened a year ago by benchen4395
  3

• how to convert model weights(e.g., bigscience/bloomz-560m-optimizer-states) to Hugging Face model.bin file?

  #16 opened a year ago by qazwsx042
  2

• Controlled generation

  #15 opened a year ago by sh0tcall3r
  1

• Use Petals without sharing GPU

  #14 opened a year ago by raihan0824
  11

• Questions on creating instruction data

  #13 opened a year ago by henryhungle
  1

• How to convert megatron-deepspeed checkpoints to huggingface checkpoints ?

  #11 opened a year ago by huybery
  4

• Were the checkpoints selected based on the held-out performance or seen task performance?

  #10 opened a year ago by MattYoon
  2

• Quesiton about MTFDataset

  #8 opened a year ago by noanti
  1

• Why does the number of templates differ between languages?

  #9 opened a year ago by MattYoon
  4

• I can't find the model weights that you used for experimentation.

  #7 opened a year ago by Mahyar-Ali
  1

• What is the training config?

  #6 opened a year ago by mkw18
  3

• Some datasets are not in xP3all

  #5 opened a year ago by mkw18
  4

• P3megds URL is not available

  #4 opened 2 years ago by hatvn
  1

• Questions about datas

  #3 opened 2 years ago by lbourdois
  1

• Is mT0 suitable for continued training on span corruption task?

  #2 opened 2 years ago by junwang-wish
  2