google-research/electra

Issues

• Electric model release?

  #142 opened 6 months ago by donglixp
  0

• What is the maximum acceptance for the sentence length for the ELECTRA model?

  #139 opened a year ago by saiefulEZO
  0

• Sampling step?

  #87 opened 4 years ago by anshulsamar
  2

• Why RoBERTa-500K has 4.5x more computation than ELECTRA-400K?

  #96 opened 4 years ago by rabbitwayne
  3

• How many parameters do discriminator and generator have?

  #138 opened a year ago by zjk000
  0

• Cannot import trace from tensorflow.python.profiler

  #132 opened 3 years ago by n-garc
  4

• ValueError: Couldn't find 'checkpoint' file or checkpoints in given directory

  #106 opened 4 years ago by etetteh
  1

• ValueError: Must specify max_steps > 0, given: 0

  #104 opened 4 years ago by etetteh
  5

• finetune preprocessing adding padding to the dataset error

  #137 opened 2 years ago by joelsprunger
  1

• failed to run cuBLAS routine: CUBLAS_STATUS_EXECUTION_FAILED

  #136 opened 2 years ago by EJDU21
  1

• some confusions about paper

  #124 opened 4 years ago by leileilin
  1

• Using own data to continue pre-training from the released ELECTRA checkpoints

  #97 opened 4 years ago
  4

• Can I used run_mlm.py in transformer for fine-tuning generator(mlm) of electra

  #134 opened 2 years ago by ToanKGO
  0

• How can I draw this?

  #130 opened 3 years ago by zshy1205
  1

• sequence tagging tasks fails at metric reporting

  #133 opened 3 years ago by Joseph-Vineland
  1

• Finetuning on Tagging task for custom dataset outputs f1 = 0 and

  #110 opened 4 years ago by etetteh
  1

• Tagging Task Segment ids

  #131 opened 3 years ago by kamalkraj
  0

• pretrain with multigpu

  #107 opened 4 years ago by 652994331
  3

• Optimal Learning Rate and Training Steps for Large Batch Size

  #129 opened 3 years ago by robinsongh381
  0

• About the Electra paper

  #128 opened 3 years ago by lgdgodv
  0

• NumPy Import Error

  #126 opened 3 years ago by tommybean
  2

• Train electra with another tokenizer

  #127 opened 3 years ago by upskyy
  2

• no module named tensorflow.contrib

  #122 opened 4 years ago by sprajagopal
  1

• Question regarding TrainingsData/Validation Data split

  #125 opened 3 years ago by samuelgoodall
  0

• Electra Vocabulary

  #123 opened 4 years ago by avinashsai
  1

• A possible mistake in the FLOPs calculation of attn_output_layer_norm in the file flops_computation.py

  #121 opened 4 years ago by hrheru2021
  0

• Adversarially training

  #117 opened 4 years ago by mshislam
  2

• ELECTRA-base fine tuned on MNLI

  #120 opened 4 years ago by ngoquanghuy99
  1

• what should i do to extract the electra discriminator

  #119 opened 4 years ago by mmx1997
  2

• The exact English pretraining data and Chinese pretraining data that are exact same to the BERT paper's pretraining data.

  #118 opened 4 years ago by guotong1988
  0

• Training on TPU got stuck

  #112 opened 4 years ago by stefan-it
  5

• It is possible do Fine-tuning on text summarization task?

  #116 opened 4 years ago by LuRuz
  0

• Typo in BERT Large flops computation

  #115 opened 4 years ago by lucadiliello
  0

• Electra-small embedding size

  #105 opened 4 years ago by YovaKem
  2

• What is ExampleBuilder and why?

  #114 opened 4 years ago by miyamonz
  0

• Exporting to SavedModel?

  #108 opened 4 years ago by artmatsak
  2

• change the dataset to MIDI

  #113 opened 4 years ago by ashBaliga
  0

• finetune models have relative bad performance when using my own base level pretrain models

  #111 opened 4 years ago by 652994331
  0

• How to pretrain electra model from my own pretrain-model

  #103 opened 4 years ago by 652994331
  0

• Question: Same Batchsize on different TPU sizes

  #89 opened 4 years ago by PhilipMay
  2

• The unbalance between original tokens and replaced tokens.

  #109 opened 4 years ago by allanchen95
  0

• Restoring ELECTRA-Small checkpoint into HuggingFace transformers model doesn't work properly

  #94 opened 4 years ago by DevKretov
  4

• ERROR:tensorflow: Failed to close session after error.Other threads may hang.

  #102 opened 4 years ago by etetteh
  5

• ValueError: Tensor conversion requested dtype string for Tensor with dtype float32: <tf.Tensor 'args_0:0' shape=() dtype=float32>

  #101 opened 4 years ago by etetteh
  0

• Can you share models trained with all weights tied?

  #100 opened 4 years ago by YovaKem
  0

• about tagging task

  #99 opened 4 years ago by LastRyan
  0

• Question about expected results

  #98 opened 4 years ago by richarddwang
  1

• How to Change Embedding Size of the Model?

  #95 opened 4 years ago by FeryET
  0

• Do you apply wnli trick ? If so, can you open the code ?

  #93 opened 4 years ago by RyanHuangNLP
  0

• UnicodeDecodeError: 'utf-8' codec can't decode byte 0xd7 in position 0: invalid continuation byte (while running build_openwebtext_pretraining_dataset.py )

  #90 opened 4 years ago by elyorman
  0