salesforce/ctrl

Issues

• tips and scripts related to data collection

  #74 opened 4 years ago by alvinchangw
  1

• control code

  #85 opened 2 years ago by dedededefo
  0

• A transformer decoer-based model or seq2seq model?

  #84 opened 2 years ago by zhaowei-wang-nlp
  0

• How to finetuning with lower memory fp16 version for p100 GPUs?

  #39 opened 5 years ago by xurongqiang
  5

• repeats the last word on AWS

  #71 opened 4 years ago by dataSci-rigo
  1

• Cuda out of memory issue.

  #83 opened 3 years ago by jamalabdul1
  0

• 12 layer (huggingface gpt-2 equivalent) ctrl model?

  #82 opened 3 years ago by zichaow
  0

• CTRL model can not work in huggingface transformers

  #80 opened 3 years ago by yananchen1989
  2

• Issues with pytorch_generation.py when running the Colab exercise

  #79 opened 3 years ago by yaw2014
  0

• Sampling settings used in the paper

  #52 opened 5 years ago by bilal2vec
  2

• why set "seq_length = min(args.generate_num, 256)"

  #68 opened 4 years ago by luweishuang
  1

• control code not recognised

  #78 opened 4 years ago by cainesap
  2

• Fine-tuning on Colab

  #63 opened 5 years ago by manueltonneau
  4

• Out of memory when fine-tuning

  #43 opened 5 years ago by orenmelamud
  9

• Will BERT+transformer-decoder better than tensor2tensor for text-generation?

  #77 opened 4 years ago by guotong1988
  0

• Using ctrl for summarization

  #64 opened 5 years ago by Hellisotherpeople
  2

• Altering the tone of the output

  #76 opened 4 years ago by qdx45
  0

• Are control codes required for finetuning?

  #75 opened 4 years ago by umhau
  0

• Sampling method used for translation

  #73 opened 4 years ago by MichaelZhouwang
  0

• License for pre-trained model

  #72 opened 4 years ago by MobiusLooper
  0

• training curriculum used

  #70 opened 4 years ago by v1nc3nt27
  0

• Source attribution - Cannot replicate results

  #67 opened 4 years ago by sipity19
  0

• How to finetune on TPU v3-8 nodes? It runs without error but does not seem to progress.

  #38 opened 5 years ago by eurka
  5

• Is that a way to do "general" generation?

  #66 opened 5 years ago by cloudygoose
  1

• TPU configuration - fine tuning

  #65 opened 5 years ago by pgrandinetti
  0

• 'tensorflow._api.v1.keras.layers' has no attribute 'LayerNormalization'

  #62 opened 5 years ago by manueltonneau
  2

• Penalize/punish some generated words more than others??!

  #61 opened 5 years ago by hypnoai
  1

• Using new top-level domain with Links control code

  #60 opened 5 years ago by pgrandinetti
  1

• Is it possible to run CTRL (full) on Gradient?

  #59 opened 5 years ago by ckoshka
  0

• ❓ Question about testing procedure

  #58 opened 5 years ago by astariul
  0

• Custom model conversion to Hugging Face

  #57 opened 5 years ago by orenmelamud
  1

• Line breaks in prompts?

  #56 opened 5 years ago by vince-lynch
  1

• Running full model on V100 outputs last word

  #49 opened 5 years ago by dimitri320
  12

• training_utils on new vocab and codes

  #53 opened 5 years ago by ctoffo
  1

• Running the model on TPUs?

  #55 opened 5 years ago by vessenes
  2

• What's the main practical difference between 256 and 512 models?

  #54 opened 5 years ago by dimitri320
  1

• OutOfMemory in fine-tuning.

  #46 opened 5 years ago by LiuYixian
  2

• more than 600 labels

  #45 opened 5 years ago by leejason
  1

• How to add new control code into vocabulary?

  #37 opened 5 years ago by leejason
  4

• benchmarking with GPT-2

  #47 opened 5 years ago by leejason
  2

• Finetuning Errors

  #32 opened 5 years ago by nickwalton
  5

• Enhancement request: How to read prompts from a text file?

  #48 opened 5 years ago by GenTxt
  1

• I can do inference with pretrained model but face error when finetuning.

  #40 opened 5 years ago by LiuYixian
  2

• Just for fun: How long would training this model take on a Nvidia 1080Ti GPU (12gb)

  #44 opened 5 years ago by timpal0l
  1

• using pytorch_generation.py: setting --temperature argument to any value causes a failure

  #41 opened 5 years ago by tanselmi
  2

• Update Google Colab Notebook to Correct Model Path Flag

  #34 opened 5 years ago by trisongz
  0

• multiple tags as control code

  #35 opened 5 years ago by leejason
  1

• smaller model

  #36 opened 5 years ago by leejason
  1

• Question about vocabulary file

  #42 opened 5 years ago by htw2012
  1

• Can a conditional control code be multiple tags or multiple keywords？

  #33 opened 5 years ago by xurongqiang
  1