facebookresearch/MetaICL

Issues

• MetaICL still needs to update its parameters during meta-training?

  #25 opened a year ago by SongDark
  3

• Inference time

  #24 opened a year ago by xipq
  3

• Additional information on `run_model` method from the `MetaICLModel` class

  #9 opened 3 years ago by MicPie
  5

• CHILD PROCESS FAILED WITH NO ERROR_FILE

  #8 opened 3 years ago by masoud-monajati
  6

• Inference with multi-GPU

  #22 opened 2 years ago by BlackieMia
  0

• Error when data tensorize

  #21 opened 2 years ago by fdszy
  0

• Can you release the checkpoints for the smaller models?

  #18 opened 2 years ago by thesofakillers
  1

• Confusion with classification when having multi tokens

  #20 opened 2 years ago by wookjeHan
  2

• Confusion about the initial evaluation results of the GPT2-LARGE model

  #19 opened 2 years ago by Lisennlp
  4

• Setting k for pre-processing train data

  #17 opened 2 years ago by thomaspzollo
  4

• How are the options shown to the model, if at all?

  #16 opened 2 years ago by thesofakillers
  1

• Using MetaICL for multi-label classification

  #15 opened 2 years ago by thesofakillers
  2

• Why using [SEP] token while using gpt2 tokenizer?

  #14 opened 2 years ago by wookjeHan
  2

• Checksum error in datasets==1.4.0

  #13 opened 2 years ago by AlexWan0
  1

• Using MetaICL for Unconstrained Generation

  #12 opened 2 years ago by aaronmueller
  3

• Max length of context

  #11 opened 3 years ago by bpucla
  1

• Questions around the data preprocessing

  #7 opened 3 years ago by JunShern
  5

• TypeError: can't pickle _thread.RLock objects

  #6 opened 3 years ago by masoud-monajati
  1

• Reproducibility of PMI methods?

  #5 opened 3 years ago by JunShern
  3

• Error in `_build_gym.py` when using `--do_train`

  #4 opened 3 years ago by JunShern
  1

• Data Downloading error

  #1 opened 3 years ago by BakingBrains
  1