google-research/simclr

Issues

• Is there a way to load the savedModel as a keras model ?

  #219 opened 7 months ago by capmac18
  0

• Error when Finetuning: WARNING:absl:Importing a function (__inference___call___15879) with ops with custom gradients. Will likely fail if a gradient is requested.

  #190 opened 3 years ago by aynesss
  2

• Command about fine-tune and semi-supervised learning on tf2

  #161 opened 3 years ago by lucasliunju
  2

• Finetune on custom dataset locally using TF2

  #211 opened 2 years ago by vsk-phi
  1

• Reproduce the results of SimCLR_v2 Finetune_10pct checkpoints

  #215 opened 2 years ago by BierOne
  1

• where are "EMA" decay and memory buffer used in this simclr work?

  #167 opened 3 years ago by fortisaqua
  4

• Cannot find the pretrained model for cifar-10 in simclrv1

  #218 opened a year ago by RTRodoshi
  0

• Regarding the situation where positive samples in a batch_size are treated as negative samples.

  #216 opened 2 years ago by YanziWang-dot
  1

• Is training on multiple GPUs possible?

  #186 opened 2 years ago by lukasugar
  2

• Loss, entropy, accuracy trends

  #188 opened 3 years ago by slala2121
  5

• Is the trained projection head available?

  #199 opened 3 years ago by lkshrsch
  7

• Reported imagenet results (e.g., Table 6) base on the validation set?

  #214 opened 2 years ago by BierOne
  2

• ProjectionHead setting in tf2

  #213 opened 2 years ago by geanshu
  0

• Linear Fine-tuning Setting

  #212 opened 2 years ago by ekurtulus
  0

• A quick question about the definition of SimCLR

  #210 opened 2 years ago by HelloWorldLTY
  2

• simclr on medical dataset gives error

  #187 opened 3 years ago by fatmacelebi
  7

• Spikes in training loss

  #189 opened 3 years ago by slala2121
  1

• About the training epochs

  #206 opened 2 years ago by utjune
  2

• layer freezing during pretraining

  #207 opened 2 years ago by HIM003
  0

• Why use bn in projection head?

  #205 opened 2 years ago by shesung
  0

• Is it possible to use gradient accumulation to counter small GPU memory?

  #203 opened 2 years ago by ChrisJWest
  3

• Dividing contrastive loss by number of replicas -- correct?

  #202 opened 2 years ago by piotrmwojcik
  0

• '_UserObject' object has no attribute 'summary'

  #196 opened 3 years ago by aynesss
  4

• Question about defining the contrastive loss on zi’s rather than hi’s

  #198 opened 3 years ago by garynlfd
  0

• Fine-tuning the model on a custom dataset?

  #194 opened 3 years ago by ImSo3K
  0

• when using the saved_model format for storing pretrained model, model loading causes many unnecessary warnings.

  #193 opened 3 years ago by aynesss
  0

• WARNING:absl:Importing a function (__inference_sync_batch_normalization_13_layer_call_and_return_conditional_losses_20965) with ops with custom gradients. Will likely fail if a gradient is requested.

  #195 opened 3 years ago by aynesss
  0

• Load Model from Checkpoints After Pre-training for Fine-Tuning Issue

  #192 opened 3 years ago by fatmacelebi
  0

• Models trained only with self supervision

  #185 opened 3 years ago by MorenoLaQuatra
  2

• Dataset Split for Linear Evaluation and Fine Tuning

  #184 opened 3 years ago by vamsi231297
  1

• Issue in Finetune V2 simclr on Cifar10

  #182 opened 3 years ago by AlexMaOLS
  2

• Two potential Bugs

  #183 opened 3 years ago by AlexMaOLS
  0

• tf2 saved_models with no projection_head output

  #181 opened 3 years ago by Maryeon
  1

• Saving model after finetuning on colab tf2

  #177 opened 3 years ago by junhuplim
  0

• Dataset of supervised baseline

  #174 opened 3 years ago by aynesss
  1

• Why do we still need a dataset with LABELS? Isn't that Self-supervised learning supposed to work without labels?

  #176 opened 3 years ago by myagmur01
  1

• A baseline supervised model is trained using random initialization.

  #175 opened 3 years ago by aynesss
  1

• supervised baseline was intialized with weight = None or "Imagenet"

  #173 opened 3 years ago by aynesss
  2

• Trainable=False or Trainable = "True" in Finetune

  #172 opened 3 years ago by aynesss
  1

• Data augmentation details

  #171 opened 3 years ago by ines321
  0

• Are data augmentation operations required at the finetuning level?

  #170 opened 3 years ago by ines321
  2

• Could you clarify `with proper learning rate scaling across batch size` in "Intriguing Properties of Contrastive Losses"?

  #166 opened 3 years ago by nzw0301
  4

• large number question

  #165 opened 3 years ago by yinciki
  3

• Transfer learning

  #163 opened 3 years ago by fjsaezm
  4

• Error in Generalized contrastive loss

  #162 opened 3 years ago by aakash-saboo
  5

• Execution in Google Colab

  #157 opened 3 years ago by fjsaezm
  0

• Error about tf2

  #160 opened 3 years ago by lucasliunju
  2

• Why knowledge distillation improves the test accuracy?

  #159 opened 3 years ago by OutstanderWang
  1

• Constrastive loss input format in the code

  #158 opened 3 years ago by Darthholi
  1

• Unknown command line flag 'variable_schema'

  #156 opened 3 years ago by liuzheyu1998
  2