lucidrains/byol-pytorch

Issues

• RuntimeError: Only Tensors created explicitly by the user (graph leaves) support the deepcopy protocol at the moment

  #52 opened 4 years ago by wengooooo
  2

• GPU Memory Usage Extremely High

  #94 opened 5 months ago by ClemensSchwarke
  0

• [BUG] in trainer.py

  #93 opened 5 months ago by ClemensSchwarke
  1

• Expected to have finished reduction in the prior iteration before starting a new one.

  #92 opened 8 months ago by DingHuilsp
  1

• duplicated calculation in gradient

  #91 opened a year ago by MaryAhn
  15

• 0.5.6 failing Multiple GPU Error

  #58 opened 4 years ago by shivam-aggarwal
  3

• About target encoder

  #89 opened 2 years ago by ZhaoYuTJPU
  3

• Resnet with pre-trained initialization

  #88 opened 2 years ago by dpetrini
  0

• What does the register_hook do?

  #87 opened 2 years ago by YTHmamba
  0

• How to get embeddings using our own trained model on custom dataset

  #86 opened 2 years ago by mahaling
  0

• How to transfer the trained ckpt to pytorch.pth model?

  #48 opened 4 years ago by knaffe
  5

• The loss has always been around 0.7

  #85 opened 2 years ago by liuzuo-byte
  0

• Need advice using custom dataset

  #60 opened 4 years ago by ramdhan1989
  1

• How to convert CKPT file to PTH file for prediction?

  #84 opened 2 years ago by weikai520
  0

• Increase EMA-parameter during training

  #82 opened 3 years ago by Benjamin-Hansson
  1

• How to cluster/predict images?

  #78 opened 3 years ago by laxmimerit
  1

• loss_fn is not cosine_similarity??

  #83 opened 3 years ago by IQ17
  0

• 【beginner question】The loss should get larger or smaller? Thank you.

  #72 opened 3 years ago by guotong1988
  1

• Why the loss is different from BYOL authors'

  #81 opened 3 years ago by Jing-XING
  2

• BN layer weights and biases are not updated

  #77 opened 3 years ago by kregmi
  1

• Warning: grad and param do not obey the gradient layout contract.

  #74 opened 3 years ago by mohaEs
  3

• For BYOL, cosine similarity of two images should be evaluated by online_encoder or target_encoder?

  #75 opened 3 years ago by vincezengqiang
  1

• EMA Update of bn buffer

  #73 opened 3 years ago by luyvlei
  5

• Intutions regarding setting up number of epochs when training on a novel dataset

  #71 opened 3 years ago by Sino0904
  0

• batch_size does not work, always 2 images are fed into model

  #70 opened 3 years ago by Mollylulu
  0

• grayscale images

  #69 opened 3 years ago by matteoguarrera
  0

• Advice of how to avoid cuda out of memory?

  #68 opened 3 years ago by HyamsG
  0

• Have you run your implementation and compare with the published results?

  #67 opened 3 years ago by wetliu
  0

• erorr about load ckpt

  #66 opened 3 years ago
  0

• Can't load ckpt

  #38 opened 4 years ago by AndrewTal
  4

• How to use the byol train our model?

  #65 opened 3 years ago by zhaowanpeng
  0

• How to train the better backbone with BYOL?

  #49 opened 4 years ago by knaffe
  1

• Is there anybody do the experiment to see whether this code base align the accuracy with the BYOL paper?

  #64 opened 3 years ago by double-fire-0
  0

• Hi, do you have pretrained checkpoints?

  #63 opened 3 years ago by CoinCheung
  0

• Throw error when retrieving an embedding for inferencing one image

  #62 opened 3 years ago by MimiCheng
  5

• Error when use 2 gpus

  #53 opened 4 years ago by superwj1990
  4

• error in image augmentation

  #61 opened 4 years ago by Jackie-LJQ
  0

• Fine-tune VIT model with BYOL method

  #59 opened 4 years ago by khawar-islam
  0

• 0.5.5 version no longer returns projections.

  #57 opened 4 years ago by shivam-aggarwal
  2

• Cannot get embedding

  #56 opened 4 years ago by Yurushia1998
  1

• Validation

  #55 opened 4 years ago by ivanpanshin
  0

• Code error when using torch.nn.DataParallel for multi-gpu: AssertionError: hidden layer avgpool never emitted an output

  #50 opened 4 years ago by SilverUnicorn
  4

• BYOL Pretrained Model

  #51 opened 4 years ago by AhmadM-DL
  1

• BYOL collapses

  #44 opened 4 years ago by mark-selyaeff
  2

• Differences between BYOL and SimSiam

  #46 opened 4 years ago by SilverUnicorn
  2

• how to inference to get embedding with this code？

  #43 opened 4 years ago by sunruina2
  1

• Whether this code is used for self-supervised learning?

  #42 opened 4 years ago by Epiphqny
  2

• example code error with multi-gpu?

  #40 opened 4 years ago by sunruina2
  0

• Training loss decreased and then increased

  #39 opened 4 years ago by easonyang1996
  4

• CKPT load

  #37 opened 4 years ago by ansleepykid
  1