Spijkervet/SimCLR

Issues

• Deprecation warning noise on each run (torchvision from pretrained)

  #48 opened 8 months ago by jojje
  0

• Code explanation in gather.py

  #33 opened 4 years ago by littleredxh
  1

• The loss doesn't decrease when using multi nodes.

  #30 opened 4 years ago by lailvlong
  10

• Multi GPU support is ... ornamental?

  #37 opened 3 years ago by mbabadi
  7

• Checkpoint at epoch 0 will be covered in the end when using DDP

  #45 opened 2 years ago by yankungou
  0

• No upscale in image augmentation?

  #44 opened 2 years ago by vedantroy
  0

• Issue about the NT_Xent loss

  #43 opened 2 years ago by 980202006
  1

• Slightly different strength parameter in CIFAR-10 setting

  #16 opened 3 years ago by nzw0301
  1

• Some questions about the projection head.

  #40 opened 3 years ago by ToaTao
  0

• Loss decerases slowly in LFW dataset, Is it normal?

  #39 opened 3 years ago by muyuuuu
  0

• Question about multi-gpu training

  #38 opened 3 years ago by Allfu
  2

• Hi spijkervet, when youreproduce the experiment on CIFAR, what is the image size and batch size you set?

  #24 opened 3 years ago by LanXiaoPang613
  1

• normalize is not used in data augmentation

  #25 opened 3 years ago by steermomo
  1

• Hello, If I want to use vgg16, how can I modify it?

  #31 opened 3 years ago by dmndxld
  1

• Features computed in def inference(loader, simclr_model, device) in linear_evaluation.py are changing everytime if we run the script again

  #36 opened 3 years ago by yang-ruixin
  0

• query regarding usage of TPU

  #11 opened 3 years ago by ambekarsameer96
  2

• train on Imagenet

  #14 opened 3 years ago by beyondpzk
  4

• Is simclr's package model trained on cifar-10?

  #32 opened 3 years ago by MiladInk
  1

• Mistakes in linear_evaluation.py

  #35 opened 3 years ago by q642936448
  0

• I got some problems when I was testing

  #28 opened 4 years ago by dmndxld
  2

• Maybe a small bug in LARS implementation

  #27 opened 4 years ago by Big-Brother-Pikachu
  0

• How to train a model in the same terminal when I used DDP training？

  #23 opened 4 years ago by szhang963
  3

• CIFAR10 accuracy

  #12 opened 4 years ago by MrChenFeng
  3

• What is world_size in this case? NT_Xent seems to ask this as a default argument but it hasn't been provided in the code?

  #22 opened 4 years ago by ajaykrishnan23
  2

• Loss function is giving different results from tensorflow implementation

  #18 opened 4 years ago by samedii
  2

• Logistic Regression Issue without "eval()"

  #19 opened 4 years ago by jaeohwoo
  1

• Why L2 norm then CosSim?

  #15 opened 4 years ago by BoPang1996
  5

• inter-gpu communication

  #13 opened 4 years ago by ShijianXu
  2

• Mismatching resnet50 model for CIFAR-10 experiments

  #8 opened 4 years ago by alegonz
  1

• torch version

  #10 opened 5 years ago by fourk-ty
  2

• Issue loading mask_correlated_samples

  #7 opened 5 years ago by Kabongosalomon
  1

• Error when loading the datasets

  #6 opened 5 years ago by Kabongosalomon
  1

• About the Cifar-10 accuracy

  #4 opened 5 years ago by jimmy-dq
  3

• Small refactoring of your loss

  #5 opened 5 years ago by lescientifik
  1

• Shouldn't this line be `args.epochs` instead of `len(loader)`?

  #3 opened 5 years ago by nlml
  1