Commits

Branch selector

 master

User selector

All users

Datepicker

All time

Commit History

Commits on Jul 15, 2024

• address https://github.com/lucidrains/byol-pytorch/issues/93

  lucidrains

  committed
  0c3ab54

  Copy full SHA for 0c3ab54

Commits on May 5, 2024

• address https://github.com/lucidrains/byol-pytorch/issues/92

  lucidrains

  committed
  3cb1a5e

  Copy full SHA for 3cb1a5e

Commits on Nov 16, 2023

• add BYOLTrainer for huggingface accelerate distributed training

  lucidrains

  committed
  61621ed

  Copy full SHA for 61621ed

• concat augmented images and only call online encoder once

  lucidrains

  committed
  40565b7

  Copy full SHA for 40565b7

Commits on Nov 15, 2023

• remove some inplace operations

  lucidrains

  committed
  4258651

  Copy full SHA for 4258651

Commits on Oct 9, 2023

• just automatically sync batchnorm if more than one machine detected, allow for overriding with sync_batchnorm keyword argument

  lucidrains

  committed
  f12087b

  Copy full SHA for f12087b

Commits on Apr 6, 2022

• fix simsiam, thanks to @chingisooinar

  lucidrains

  committed
  6717204

  Copy full SHA for 6717204

Commits on Jul 15, 2021

• add more informative error message

  lucidrains

  committed
  caa65d7

  Copy full SHA for caa65d7

Commits on Apr 12, 2021

• default to returning projections, but can be turned off with return_projection = False on forward

  lucidrains

  committed
  2aa84ee

  Copy full SHA for 2aa84ee

Commits on Mar 31, 2021

• 0.5.5

  lucidrains

  authored
  8c08cef

  Copy full SHA for 8c08cef

• Merge pull request #54 from gauenk/patch-1

  Show description for ad094de
  lucidrains

  authored
  ad094de

  Copy full SHA for ad094de

Commits on Mar 30, 2021

• Correcting default behavior

  Show description for fe10272
  gauenk

  authored
  fe10272

  Copy full SHA for fe10272

Commits on Feb 17, 2021

• update

  lucidrains

  authored
  e071db0

  Copy full SHA for e071db0

Commits on Feb 13, 2021

• fix byol

  lucidrains

  committed
  88a9d9d

  Copy full SHA for 88a9d9d

Commits on Feb 6, 2021

• bump

  lucidrains

  committed
  6a20ce8

  Copy full SHA for 6a20ce8

• try fix for multi-gpu, as suggested here https://discuss.pytorch.org/t/register-forward-hook-with-multiple-gpus/12115/10

  lucidrains

  committed
  8b0be48

  Copy full SHA for 8b0be48

Commits on Jan 28, 2021

• make sure sync batchnorm is turned on for distributed training, in light of momentum^2 paper

  lucidrains

  committed
  26f70a0

  Copy full SHA for 26f70a0

Commits on Jan 12, 2021

• store original net on self for access

  lucidrains

  committed
  2d34da1

  Copy full SHA for 2d34da1

Commits on Dec 28, 2020

• update readme

  lucidrains

  committed
  cde298d

  Copy full SHA for cde298d

Commits on Dec 19, 2020

• advertise other repo

  lucidrains

  committed
  3d8a22a

  Copy full SHA for 3d8a22a

Commits on Dec 15, 2020

• update readme

  lucidrains

  committed
  b25b0e4

  Copy full SHA for b25b0e4

• Merge pull request #45 from mark-selyaeff/master

  Show description for 0063171
  lucidrains

  authored
  0063171

  Copy full SHA for 0063171

• allow for returning embeddings and projections with an extra flag on forward

  lucidrains

  committed
  38ddb42

  Copy full SHA for 38ddb42

• Add use_momentum support for Lightning

  mark-selyaeff

  committed
  ff59341

  Copy full SHA for ff59341

Commits on Dec 8, 2020

• Update setup.py

  lucidrains

  authored
  afaad2d

  Copy full SHA for afaad2d

• Merge pull request #41 from umbertov/master

  Show description for 0eefd08
  lucidrains

  authored
  0eefd08

  Copy full SHA for 0eefd08

• edit dependencies

  umbertov

  committed
  3f6aafe

  Copy full SHA for 3f6aafe

• Refactor to use the correct RandomApply

  umbertov

  committed
  387dc40

  Copy full SHA for 387dc40

• Use Torchvision augmentations instead of Kornia, which work with TPUs

  Show description for 18e7ca8
  umbertov

  committed
  18e7ca8

  Copy full SHA for 18e7ca8

Commits on Nov 23, 2020

• cite

  lucidrains

  committed
  0b58f56

  Copy full SHA for 0b58f56

• update readme

  lucidrains

  committed
  0a71b88

  Copy full SHA for 0a71b88

• add option to turn BYOL into SimSiam

  lucidrains

  committed
  816e252

  Copy full SHA for 816e252

• more data

  lucidrains

  committed
  f03d919

  Copy full SHA for f03d919

Commits on Nov 1, 2020

• make sure parameters of target encoder is never updated by default

  lucidrains

  committed
  c914a18

  Copy full SHA for c914a18

Commits on Oct 26, 2020

• update readme

  lucidrains

  committed
  7182100

  Copy full SHA for 7182100

Pagination

Previous

Next