ConnorJL/GPT2

Issues

• Question about the metric reported in the paper?

  #38 opened 2 years ago by dsj96
  0

• create_tfrecords.py。Dealing with problems with your own data set

  #36 opened 3 years ago by xsyzka
  0

• where is the length of the forecast article set? Thank you!

  #35 opened 3 years ago by xsyzka
  0

• Samples?

  #34 opened 4 years ago by sleepinyourhat
  0

• Training 1.5B?

  #33 opened 4 years ago by JulesGM
  0

• Retraining a new model, only gpu 0 can be used

  #32 opened 4 years ago by yds1024
  1

• Error on output

  #31 opened 4 years ago by silicahd
  1

• GPT vs BERT, under same computation and data resource, which one is better for downstream tasks like GLUE?

  #30 opened 4 years ago by guotong1988
  0

• 117M/model.ckpt.index is corrupted?

  #29 opened 4 years ago by ksjae
  0

• format dataset

  #13 opened 6 years ago by khaerulumam42
  6

• character-level

  #28 opened 4 years ago by amacfie
  1

• about encoder.json

  #27 opened 5 years ago by fnyhy
  4

• I figured out how to cram GPT-2 1.5B onto a single TPU core with Adam optimizer

  #23 opened 5 years ago by shawwn
  3

• DOCKER: Web interface doesn't work

  #26 opened 5 years ago by fartwhif
  0

• Docker documentation for CUDA

  #25 opened 5 years ago by fartwhif
  0

• Training on artificial language data (server logs, medical records, etc.)

  #24 opened 5 years ago by klimentij
  1

• How can i create smaller sized file for inference of 1.5B model

  #22 opened 5 years ago by pragnakalpdev6
  1

• Are there some research papers about text-to-set generation?

  #21 opened 5 years ago by guotong1988
  1

• error when using create_tfrecords.py

  #20 opened 5 years ago by CrackerHax
  3

• Your 1.5B model

  #19 opened 5 years ago by 4R7I5T
  2

• when reading metadata of gs://openwebtext/stuff/encoder/encoder.json

  #18 opened 5 years ago by makamkkumar
  1

• Downloading Encoder Model fails

  #16 opened 5 years ago by PickHub
  2

• Training problem

  #15 opened 5 years ago by DrYangLiu
  1

• Input Chinese, the predicted is Japanese.

  #5 opened 6 years ago by dpyneo
  9

• quirks that hold the model back

  #11 opened 6 years ago by murpen
  4

• Why gpt-2 could apply to other tasks without fine-tune?

  #14 opened 6 years ago by guotong1988
  2

• Predicting with PrettyBigModel `InvalidArgumentError: indices[0,0] = 1024 is not in [0, 1024)`

  #4 opened 6 years ago by pkmital
  5

• what's the difference between sample and sample_free?

  #12 opened 6 years ago by Tianyu00
  1

• To train my model means fit-tuning or retrain a model?

  #10 opened 6 years ago by wjy979769265
  4

• A meaningful performance comparison with OpenAI's models

  #3 opened 6 years ago by lostmsu
  5

• Has anyone managed to work it on Windows? Which OS did you use to make it work?

  #6 opened 6 years ago by FurkanGozukara
  2

• How to process raw text files to create similar "PrettyBig" model?

  #2 opened 6 years ago by GenTxt
  5

• Unable to predict with bfloat16 model

  #1 opened 6 years ago by kizinfo
  2