facebookresearch/moco issues and pull requests

#154 - maybe wrong optimizer

Issue - State: open - Opened by poonam2308 about 1 year ago

#153 - Where can I find the code?

Issue - State: open - Opened by SHIMURA0 about 1 year ago

#152 - Where can I find moco code?

Issue - State: closed - Opened by Ashespt about 1 year ago - 2 comments

#151 - Adding Code of Conduct file

Pull Request - State: open - Opened by facebook-github-bot about 1 year ago

#150 - Adding Contributing file

Pull Request - State: open - Opened by facebook-github-bot about 1 year ago - 1 comment
Labels: CLA Signed

#150 - Adding Contributing file

Pull Request - State: open - Opened by facebook-github-bot about 1 year ago

#149 - Unable to open the compressed file

Issue - State: open - Opened by PeterJiang122 about 1 year ago

#148 - About using the model dict

Issue - State: open - Opened by Dennis-404 over 1 year ago - 1 comment

#147 - About License

Issue - State: open - Opened by myalos over 1 year ago

#146 - The size of the dictionary

Issue - State: closed - Opened by bftan1949 over 1 year ago - 2 comments

#145 - Why is labels = zeros(N) set to zero?

Issue - State: open - Opened by Dorothea5 almost 2 years ago - 3 comments

#143 - Question about queue dimension

Issue - State: open - Opened by yeomjh00 about 2 years ago

#142 - why pretrain from encoder_q?

Issue - State: open - Opened by hu5tao about 2 years ago - 1 comment

#141 - More idiomatic param update

Pull Request - State: open - Opened by eellison about 2 years ago
Labels: CLA Signed

#140 - Can you tell me dataset structure and how images are named in the dataset

Issue - State: open - Opened by shenyehui over 2 years ago - 1 comment

#139 - Concerns about feature dimensionality in MoCo self-training

Issue - State: open - Opened by LALBJ over 2 years ago

#138 - What is the label format of the cifar-10 dataset?

Issue - State: open - Opened by wangfuguii over 2 years ago - 1 comment

#137 - what information is leaked due to intra-batch communication?

Issue - State: open - Opened by hackeryhw over 2 years ago - 2 comments

#136 - About training

Issue - State: open - Opened by chencn2020 over 2 years ago - 2 comments

#135 - Relicense MoCo under MIT

Pull Request - State: closed - Opened by subramen over 2 years ago - 4 comments
Labels: CLA Signed, Merged

#134 - How to load the Hyperparameters without command line code Argument Parser?

Issue - State: open - Opened by arohi2bujji almost 3 years ago

#133 - One question about single GPU

Issue - State: open - Opened by Sangh0 almost 3 years ago - 2 comments

#132 - Low Accuracy

Issue - State: open - Opened by dbsdmlgus50 almost 3 years ago

#131 - Issue with batch size

Issue - State: open - Opened by kkannan8291 about 3 years ago - 1 comment

#130 - Why labels are all zeros, should first columns of labels be ones?

Issue - State: closed - Opened by FrankWork about 3 years ago - 4 comments

#129 - Question about the queue for key encoder

Issue - State: open - Opened by Rohit8y over 3 years ago - 2 comments

#128 - Issue about dequeue_and_enqueue

Issue - State: closed - Opened by qishibo over 3 years ago - 3 comments

#127 - Question about transfering to COCO with Mocov1 and Mocov2 checkpoint

Issue - State: open - Opened by mZhenz over 3 years ago

#126 - cannot reproduce the results of moco-v2

Issue - State: open - Opened by shaoming20798 over 3 years ago - 2 comments

#125 - question about the momentum encoder

Issue - State: open - Opened by muaz1994 over 3 years ago - 1 comment

#124 - about change the base model

Issue - State: open - Opened by zhou-rui1 over 3 years ago

#123 - Minor fix for accuracy calculation

Pull Request - State: open - Opened by chingisooinar over 3 years ago - 2 comments
Labels: CLA Signed

#122 - The size of an embedding is 1000, how would i change it?

Issue - State: open - Opened by algoteam5 over 3 years ago

#121 - Config issue about 'Res5ROIHeadsExtraNorm'

Issue - State: closed - Opened by Kevinz-code over 3 years ago - 1 comment

#120 - Report a mistake

Issue - State: open - Opened by AntonotnaWang over 3 years ago

#119 - with official pretrained model, i train a classification for 100 epochs, but find nothing record-file.

Issue - State: open - Opened by guangminglftg over 3 years ago - 1 comment

#118 - Algorithm question

Issue - State: closed - Opened by slala2121 almost 4 years ago

#117 - 'RGB' for MoCo, 'BGR' for Detectron2

Issue - State: closed - Opened by fwtan almost 4 years ago - 2 comments

#116 - eval in training part of linear classification code

Issue - State: open - Opened by shuuchen almost 4 years ago - 1 comment

#115 - requirements of the packages?

Issue - State: open - Opened by zjuKeLiu almost 4 years ago

#114 - Error while inferences from the trained detectron2 model on image

Issue - State: open - Opened by letdivedeep almost 4 years ago - 1 comment

#113 - moco stepwise lr scheduler (bug?)

Issue - State: closed - Opened by abaqusstudent almost 4 years ago - 2 comments

#112 - Training an Pretrained model on object detection task on single GPU

Issue - State: closed - Opened by letdivedeep about 4 years ago

#111 - About the batch size

Issue - State: open - Opened by iamimage about 4 years ago

#110 - nvm

Issue - State: closed - Opened by YellowPig-zp about 4 years ago

#109 - Algorithm 1 in MoCo Paper

Issue - State: open - Opened by xuehaouwa about 4 years ago - 1 comment

#101 - Positive key indicators are always zeros

Issue - State: closed - Opened by AlekseySh about 4 years ago - 2 comments

#100 - enqueue fixed.

Pull Request - State: open - Opened by sargdsra about 4 years ago
Labels: CLA Signed

#99 - enqueue fixed.

Pull Request - State: closed - Opened by sargdsra about 4 years ago - 1 comment

#98 - Questions on scaling the learning rate

Issue - State: open - Opened by DianCh about 4 years ago

#97 - Multi-node

Issue - State: open - Opened by techmonsterwang about 4 years ago - 3 comments

#96 - What is the training dataset of the moco pretrained models?

Issue - State: closed - Opened by EdyQiu about 4 years ago

#95 - Start Train Linear Classfication Task . view size is not compatible

Issue - State: closed - Opened by wscjxky over 4 years ago - 1 comment

#94 - A question about the encoder

Issue - State: open - Opened by pamela0530 over 4 years ago

#93 - Same-category sample in the negative queue

Issue - State: open - Opened by triangleCZH over 4 years ago - 1 comment

#92 - questions about downstream tasks

Issue - State: open - Opened by tangyuelm over 4 years ago

#91 - BN

Issue - State: closed - Opened by ghost over 4 years ago - 1 comment

#90 - about the lr of main_lincls.py

Issue - State: open - Opened by TheEighthDay over 4 years ago

#89 - Unable to load pre-trained Moco V2 model

Issue - State: open - Opened by andreesteva over 4 years ago

#88 - Loading a pre-trained model on single GPU?

Issue - State: closed - Opened by linzhiqiu over 4 years ago - 2 comments

#87 - one question about the implementation

Issue - State: open - Opened by shuuchen over 4 years ago - 1 comment

#86 - Question regarding the paper

Issue - State: open - Opened by ivanpanshin over 4 years ago

#85 - anyone apply moco to new dataset(not imagenet ) to get a better result

Issue - State: open - Opened by qingzew over 4 years ago - 2 comments

#84 - The pretrain loading doesn't need "module."

Issue - State: open - Opened by linzhlalala over 4 years ago

#83 - Checkpointing on multiple nodes

Issue - State: open - Opened by yoavmiron98 over 4 years ago

#82 - Use moco pretrained weights for detection

Issue - State: closed - Opened by Epiphqny over 4 years ago - 1 comment

#81 - How to use MOCO for a segmentation task？

Issue - State: open - Opened by NeuZhangQiang over 4 years ago - 1 comment

#80 - Can group normalization (GN) be used to avoid cheating?

Issue - State: open - Opened by eezywu over 4 years ago - 3 comments

#79 - Does MOCO collapses under simpler augmentation?

Issue - State: open - Opened by a411919924 over 4 years ago - 4 comments

#78 - Bad result when combining moco and ResNet-IBN.

Issue - State: open - Opened by tengteng95 over 4 years ago

#77 - Question regarding the code in main_lincls.py

Issue - State: open - Opened by zhangdan8962 over 4 years ago

#76 - Instance segmentation on CoCo

Issue - State: closed - Opened by 1144181135 over 4 years ago - 1 comment

#75 - Training on multi-gpu

Issue - State: open - Opened by bhat-prashant over 4 years ago

#74 - Why the lr in main_lincls is 30?

Issue - State: closed - Opened by lzyhha almost 5 years ago - 1 comment

#73 - ValueError: Decompressed Data Too Large and acc is strangely low

Issue - State: open - Opened by feufhd almost 5 years ago

#72 - How to set specific gpus if I do not have 8 gpus?

Issue - State: open - Opened by xuChenSJTU almost 5 years ago - 4 comments

#71 - The final accuracy rate during evaluation is too low

Issue - State: closed - Opened by 781458112 almost 5 years ago - 3 comments

#70 - Buffer of BN in EMA update

Issue - State: closed - Opened by LeeDoYup almost 5 years ago - 2 comments

#69 - train my own dataset

Issue - State: closed - Opened by sunset326 almost 5 years ago - 2 comments

#68 - RuntimeError: inputs must be on unique devices when python main.py

Issue - State: closed - Opened by sunset326 almost 5 years ago - 1 comment

#67 - Is shuffle batch norm tied with momentum contrast training?

Issue - State: closed - Opened by dongyaoli10x about 5 years ago - 6 comments

#66 - Are there results with other normalizations?

Issue - State: closed - Opened by LeeDoYup about 5 years ago - 6 comments

#65 - Reproducing semantic segmentation results

Issue - State: open - Opened by Neabfi about 5 years ago

#64 - Update README.md on Colab notebook

Pull Request - State: closed - Opened by KaimingHe about 5 years ago - 1 comment
Labels: CLA Signed, Merged

#63 - Difference between the encoder_q and the encoder_k.

Issue - State: open - Opened by RuoyiDu about 5 years ago - 1 comment

#62 - Negative samples

Issue - State: open - Opened by zhangdan8962 about 5 years ago - 3 comments

#61 - AssertionError: Default process group is not initialized

Issue - State: open - Opened by upupbo about 5 years ago - 4 comments

#60 - sharing traning log

Issue - State: open - Opened by ROBOTICSENGINEER about 5 years ago

#59 - Is that possible to report pretraining checkpoint model?

Issue - State: open - Opened by miaozhang0525 about 5 years ago

#58 - RuntimeError: No rendezvous handler for ://

Issue - State: closed - Opened by ming1993li about 5 years ago - 4 comments

#57 - AttributeError: 'Tensor' object has no attribute 'T'

Issue - State: closed - Opened by chinglamchoi about 5 years ago - 2 comments

#56 - Validation step missing in main_moco.py?

Issue - State: closed - Opened by ankuPRK about 5 years ago - 1 comment

#55 - Code and settings of semantic segmentation task

Issue - State: closed - Opened by yshinya6 about 5 years ago - 5 comments

#54 - torch.multiprocessing.spawn got error "process 0 terminated with exit code 1"

Issue - State: closed - Opened by miaozhang0525 about 5 years ago - 1 comment

#47 - Can't use one GPU?

Issue - State: closed - Opened by HymEric about 5 years ago - 8 comments

#46 - Can synchronized batch norm (SyncBN) be used to avoid cheating? Is shuffling BN a must?

Issue - State: closed - Opened by w-hc about 5 years ago - 3 comments

GitHub / facebookresearch/moco issues and pull requests