microsoft/DeBERTa

Issues

• How to pre training mDeBERTa model?

  #155 opened 13 days ago by 9mean2
  0

• Install fails due to use of deprecated `sklearn` package

  #134 opened 2 years ago by benfogelson
  1

• Pretraining the deberta-v3 by larger context length.

  #153 opened 4 months ago by sherlcok314159
  2

• RTD is not registed

  #154 opened 2 months ago by Eric-Chen-007
  1

• AssertionError: RTD is not registed.

  #129 opened 2 years ago by StephennFernandes
  3

• Fine-tune DeBERTa v3 language model, worthwhile endeavour?

  #151 opened 8 months ago by shensmobile
  5

• Generator weights

  #152 opened 7 months ago by ir2718
  0

• Deberta-v3-base Generator model

  #131 opened 2 years ago by sharanyarc96
  2

• How can I evaluate COPA dataset?

  #150 opened 10 months ago by KwanghyeonLee
  0

• Reason for missing values in table for the Roberta-base, mrpc entry

  #149 opened a year ago by Aradhye2002
  0

• Evaluation hangs for distributed MLM task

  #104 opened 3 years ago by dannyel2511
  7

• No assert: Training does not start when using different tokenizer/ tokenized-data

  #148 opened a year ago by adriwitek
  0

• Inference gives different results when using multiple gpus (distributed mode) vs just one gpu (not distributed mode)

  #147 opened a year ago by ThuongTNguyen
  0

• Model is not initialized correctly when path to a pretrained model is provided via `pre_trained`

  #146 opened a year ago by ThuongTNguyen
  0

• Question regarding symmetric KL Loss

  #145 opened a year ago by skbaur
  0

• EOF error while running the rtd.sh script

  #139 opened a year ago by BartWesthoff
  1

• Load deberta-v3-large but got deberta-v2 model

  #132 opened 2 years ago by ChengsongLu
  2

• out of memory

  #109 opened 2 years ago by Amazing-J
  18

• Trying to initialize model "large"

  #140 opened a year ago by Saivaks
  0

• Trying to run rtd_task.py on Windows

  #137 opened a year ago by Yuri-Albuquerque
  1

• Eligibility for Commercial Use

  #135 opened a year ago by Hegelim
  1

• When calculating Qr, why is the W of content used instead of the W of position used?

  #136 opened a year ago by nebula303
  0

• Error when running the example code for pretraining the rtd model.

  #127 opened 2 years ago by soonilbae
  15

• n/a

  #130 opened 2 years ago by StephennFernandes
  0

• No module named 'torch._six'

  #128 opened 2 years ago by StephennFernandes
  2

• mDeBERTa Generator model

  #123 opened 2 years ago by dadelani
  3

• effectiveness of RTD

  #126 opened 2 years ago by martin-reczko
  0

• Info on Deberta-v2-xlarge training infra

  #125 opened 2 years ago by karthickgopalswamy
  0

• Microsoft

  #124 opened 2 years ago by omniteams
  0

• This model for MLM is waste of time, why did you even made it if it cannot be used?

  #99 opened 2 years ago by Oxi84
  6

• How to pretrain DeBERTa v3 ??

  #108 opened 2 years ago by BinhMinhs10
  2

• Where is the Gradient-Disentangled Embedding Sharing(GDES) part in the code?

  #111 opened 2 years ago by Cakeyan
  3

• Code about deberta_v3

  #116 opened 2 years ago by BAOOOOOM
  1

• which version is torch ?

  #119 opened 2 years ago by XuJianzhi
  0

• Generator Model

  #121 opened 2 years ago by prajwal967
  1

• Convert DeBERTa model to ONNX with mixed precision

  #120 opened 2 years ago by SergeyShk
  0

• why vocab.txt and tokenizer.json not in pretrained model in huggingface ??

  #117 opened 2 years ago by XuJianzhi
  1

• AssertionError: [] in google coab

  #115 opened 2 years ago by yupesh
  0

• Can you upload the code finetuned in SQuad 2.0? Thank you very much.

  #114 opened 2 years ago by junzai0215
  0

• mDeBERTa large

  #113 opened 2 years ago by djstrong
  0

• Can you tell me which token represents the overall representation of the sentence in the task of feature-extraction? The first token or the last token?

  #112 opened 2 years ago by junzai0215
  0

• Can't run bash commands in /DeBERTa/experiments/glue/

  #110 opened 2 years ago by heya5
  0

• Why does the size of DeBERTaV3 double on disk after finetuning?

  #106 opened 2 years ago by nadahlberg
  2

• Embedding layer vocab size not match to tokenizer length

  #103 opened 3 years ago by kingbone9
  1

• where is ENHANCED MASK DECODER ACCOUNTS part in code?

  #105 opened 3 years ago by tjshu
  1

• DeXLNeta

  #102 opened 3 years ago by LifeIsStrange
  0

• Pre-training times: v2 vs. v3

  #100 opened 3 years ago by stefan-it
  1

• AttributeError: 'DebertaV2Tokenizer' object has no attribute 'get_vocab_size'

  #101 opened 3 years ago by pn12
  0

• How to use this model for MLM task?

  #98 opened 3 years ago by Oxi84
  0

• Release source distribution through PyPI or GitHub releases

  #97 opened 3 years ago by BastianZim
  1