Commits

Breadcrumbs

History for

onv2.4.0

User selector

All users

Datepicker

All time

Commit History

Commits on Jan 28, 2020

• Default save steps 50 to 500 in all scripts

  LysandreJik

  committed
  335dd5e

  Copy full SHA for 335dd5e

• [run_lm_finetuning] GPT2 tokenizer doesn't have a pad_token

  Show description for 6b4c3ee
  julien-c

  committed
  6b4c3ee

  Copy full SHA for 6b4c3ee

Commits on Jan 21, 2020

• Line-by-line text dataset (including padding)

  julien-c

  committed
  1a8e87b

  Copy full SHA for 1a8e87b

• change order

  julien-c

  committed
  b94cf7f

  Copy full SHA for b94cf7f

• Easier to not support this, as it could be confusing

  Show description for 2eaa8b6
  julien-c

  committed
  2eaa8b6

  Copy full SHA for 2eaa8b6

• make style

  julien-c

  committed
  801aaa5

  Copy full SHA for 801aaa5

• [run_lm_finetuning] Train from scratch

  julien-c

  committed
  56d4ba8

  Copy full SHA for 56d4ba8

Commits on Jan 7, 2020

• spelling correction (#2434)

  

  orena1

  authored and

  LysandreJik

  committed
  43114b8

  Copy full SHA for 43114b8

• Fix error with global step in run_lm_finetuning.py

  LysandreJik

  committed
  27c1b65

  Copy full SHA for 27c1b65

Commits on Jan 6, 2020

• GPU text generation: mMoved the encoded_prompt to correct device

  

  alberduris

  authored and

  LysandreJik

  committed
  81d6841

  Copy full SHA for 81d6841

• Moved the encoded_prompts to correct device

  

  alberduris

  authored and

  LysandreJik

  committed
  dd4df80

  Copy full SHA for dd4df80

Commits on Jan 1, 2020

• [run_lm_finetuning] mask_tokens: document types

  julien-c

  committed
  629b22a

  Copy full SHA for 629b22a

Commits on Dec 22, 2019

• Update comments mentioning Python 2.

  aaugustin

  committed
  d6eaf4e

  Copy full SHA for d6eaf4e

• Remove __future__ imports.

  aaugustin

  committed
  c824d15

  Copy full SHA for c824d15

• Fix E266 flake8 warning (x90).

  aaugustin

  committed
  fa2ccbc

  Copy full SHA for fa2ccbc

• Fix E722 flake8 warnings (x26).

  aaugustin

  committed
  631be27

  Copy full SHA for 631be27

• Sort imports with isort.

  Show description for 158e82e
  aaugustin

  committed
  158e82e

  Copy full SHA for 158e82e

Commits on Dec 21, 2019

• Reformat source code with black.

  Show description for fa84ae2
  aaugustin

  committed
  fa84ae2

  Copy full SHA for fa84ae2

• Merge pull request #2130 from huggingface/ignored-index-coherence

  Show description for 6f68d55
  thomwolf

  authored
  6f68d55

  Copy full SHA for 6f68d55

Commits on Dec 19, 2019

• [doc] Param name consistency

  julien-c

  committed
  a5a06a8

  Copy full SHA for a5a06a8

• Minor/basic text fixes (#2229)

  Show description for 1718fb9
  

  aidankierans

  authored and

  julien-c

  committed
  1718fb9

  Copy full SHA for 1718fb9

Commits on Dec 10, 2019

• Correct index in script

  LysandreJik

  committed
  b72f9d3

  Copy full SHA for b72f9d3

Commits on Dec 9, 2019

• Remove unnecessary epoch variable

  

  bilal2vec

  authored and

  LysandreJik

  committed
  79526f8

  Copy full SHA for 79526f8

• Add functionality to continue training from last saved global_step

  

  bilal2vec

  authored and

  LysandreJik

  committed
  9626e04

  Copy full SHA for 9626e04

• Stop saving current epoch

  

  bilal2vec

  authored and

  LysandreJik

  committed
  2d73591

  Copy full SHA for 2d73591

• Use saved optimizer and scheduler states if available

  

  bilal2vec

  authored and

  LysandreJik

  committed
  0eb973b

  Copy full SHA for 0eb973b

• Save tokenizer after each epoch to be able to resume training from a checkpoint

  

  bilal2vec

  authored and

  LysandreJik

  committed
  a03fcf5

  Copy full SHA for a03fcf5

• Save optimizer state, scheduler state and current epoch

  

  bilal2vec

  authored and

  LysandreJik

  committed
  f71b1bb

  Copy full SHA for f71b1bb

Commits on Dec 4, 2019

• fix #1991

  thomwolf

  committed
  5bfcd04

  Copy full SHA for 5bfcd04

Commits on Dec 3, 2019

• Use full dataset for eval (SequentialSampler in Distributed setting)

  VictorSanh

  committed
  48cbf26

  Copy full SHA for 48cbf26

Commits on Nov 29, 2019

• Added Camembert to available models

  

  maxvidal

  authored and

  julien-c

  committed
  b0ee7c7

  Copy full SHA for b0ee7c7

Commits on Nov 25, 2019

• resize model when special tokenizer present

  

  iedmrc

  authored and

  LysandreJik

  committed
  aa92a18

  Copy full SHA for aa92a18

• fix #1894

  LysandreJik

  committed
  7485cae

  Copy full SHA for 7485cae

Commits on Nov 14, 2019

• Merge pull request #1804 from ronakice/master

  Show description for 9629e2c
  thomwolf

  authored
  9629e2c

  Copy full SHA for 9629e2c

• Merge pull request #1832 from huggingface/memory-leak-schedulers

  Show description for df99f8c
  thomwolf

  authored
  df99f8c

  Copy full SHA for df99f8c

Pagination

Previous

Next