Data

Tools

Indexes

Applications

Experiments

Issues

An open API service for providing issue and pull request metadata for open source projects.

GitHub / KevinMusgrave/pytorch-metric-learning issues and pull requests

#748 - Strange about TripletMarginLoss + CrossBatchMemory

Issue - State: open - Opened by DoItWithMe 2 months ago - 5 comments
Labels: bug

#747 - v2.9.0

Pull Request - State: open - Opened by KevinMusgrave 3 months ago

#746 - Fix kwargs not passed to SubCenterArcFaceLoss

Pull Request - State: closed - Opened by marcpaga 4 months ago - 3 comments

#745 - master into dev

Pull Request - State: closed - Opened by KevinMusgrave 4 months ago

#744 - Some losses can't handle large vales in `targets`/`labels`

Issue - State: closed - Opened by v-iashin 4 months ago - 2 comments
Labels: possible bug?

#743 - indices_tuple: Add assertion that each pair should be either positive or negative

Pull Request - State: closed - Opened by lucamarini22 4 months ago - 6 comments

#742 - Multi-labels

Issue - State: closed - Opened by lucamarini22 4 months ago - 3 comments

#741 - k-NN clusters as labels or pairs?

Issue - State: closed - Opened by stanleyjs 6 months ago - 1 comment

#740 - Subcenter Arcface not dropping non-dominant class centers?

Issue - State: closed - Opened by SNeugber 6 months ago - 2 comments

#739 - This error occurs when I use MultiSupConLoss

Issue - State: open - Opened by luoyq6 7 months ago - 1 comment

#737 - Allow string for `device` argument of AccuracyCalculator

Issue - State: open - Opened by KevinMusgrave 7 months ago - 1 comment
Labels: enhancement

#736 - v2.8.1

Pull Request - State: closed - Opened by KevinMusgrave 8 months ago

#735 - v2.8.0

Pull Request - State: closed - Opened by KevinMusgrave 8 months ago

#734 - Support for dictionaries

Issue - State: closed - Opened by ir2718 8 months ago - 2 comments

#733 - Utility for finding the best threshold?

Issue - State: open - Opened by ir2718 8 months ago - 3 comments

#732 - the question of DistributedLossWrapper

Issue - State: closed - Opened by hukongzhi 8 months ago - 3 comments

#731 - Error using DynamicSoftMarginLoss

Issue - State: open - Opened by inakierregueab 9 months ago - 2 comments
Labels: bug

#730 - should faiss be a requirement?

Issue - State: open - Opened by sammlapp 9 months ago - 1 comment

#729 - master into dev

Pull Request - State: closed - Opened by KevinMusgrave 9 months ago

#728 - v2.7.0

Pull Request - State: closed - Opened by KevinMusgrave 9 months ago

#727 - Numpy requirement

Issue - State: closed - Opened by pchampio 9 months ago - 2 comments

#726 - AM-softmax loss

Issue - State: open - Opened by sofpya 9 months ago - 1 comment
Labels: new algorithm request

#725 - Possible "improvement" of MPerClassSampler?

Issue - State: open - Opened by ir2718 9 months ago - 1 comment

#724 - [`feat`] Addition of popular image retrieval benchmark datasets

Pull Request - State: closed - Opened by ir2718 9 months ago - 4 comments

#723 - Bug in MultiSimilarityMiner?

Issue - State: open - Opened by kazimpal87 9 months ago - 6 comments

#722 - Addition of popular benchmark datasets

Issue - State: open - Opened by ir2718 10 months ago - 3 comments

#721 - [`feat`] Implementing SmoothAP loss

Pull Request - State: closed - Opened by ir2718 10 months ago - 3 comments

#720 - [`feat`] Implementing threshold consistent margin loss

Pull Request - State: closed - Opened by ir2718 10 months ago - 3 comments

#719 - Master update to musgrave

Pull Request - State: closed - Opened by frouse4531 10 months ago - 1 comment

#718 - Can MatchFinder/InferenceModel return distances for all classes?

Issue - State: open - Opened by vemchance 10 months ago - 3 comments

#717 - KeyError in load_model()

Issue - State: open - Opened by ManuelZ 11 months ago - 2 comments

#715 - Evaluating big datasets on GPU

Issue - State: closed - Opened by ir2718 11 months ago - 3 comments

#714 - Corrected the paper reference in the CosFace loss

Pull Request - State: closed - Opened by elisim 11 months ago - 1 comment

#712 - Wrapping NTXentLoss in DistributedLossWrapper makes loss extremely high

Issue - State: closed - Opened by irowberryFS 11 months ago - 8 comments

#711 - Confused about SelfSupervisedLoss and labeling with dataset of pairs

Issue - State: closed - Opened by irowberryFS 11 months ago - 6 comments

#710 - Wrapping the loss with DistributedDataParallel

Issue - State: closed - Opened by elisim 12 months ago - 1 comment

#708 - Quadruplet Miners

Issue - State: open - Opened by Mahendra114027 12 months ago - 1 comment
Labels: enhancement

#706 - UnboundLocalError: local variable 'all_q' referenced before assignment

Issue - State: closed - Opened by zxlml about 1 year ago - 1 comment

#705 - Another ntxent loss question

Issue - State: closed - Opened by Stas384 about 1 year ago - 4 comments

#704 - v2.6.1

Pull Request - State: closed - Opened by KevinMusgrave about 1 year ago

#703 - v2.6.0

Pull Request - State: closed - Opened by KevinMusgrave about 1 year ago

#702 - DistributedLossWrapper enhancements

Pull Request - State: closed - Opened by elisim about 1 year ago - 4 comments

#701 - arguments names are different in forward in when using the DDP wrapper

Issue - State: closed - Opened by elisim about 1 year ago - 1 comment

#700 - Distributed training not using all available GPUs on large dataset

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

#699 - InfoNCE errors

Issue - State: closed - Opened by davidireland3 about 1 year ago - 4 comments

#697 - Allow additional parameters in `compute_mat`

Issue - State: open - Opened by celsofranssa about 1 year ago - 10 comments
Labels: enhancement

#696 - NTXentLoss, normalize issue.

Issue - State: closed - Opened by jzhanghzau about 1 year ago - 3 comments

#694 - DDP and Faiss TemporaryMemoryBuffer error

Issue - State: closed - Opened by vemchance over 1 year ago - 4 comments

#692 - v2.5.0

Pull Request - State: closed - Opened by KevinMusgrave over 1 year ago

#690 - Model overfitting, smooth triplet margin loss

Issue - State: open - Opened by taaresh7 over 1 year ago - 2 comments

#687 - CASIA Face Recognition

Issue - State: open - Opened by gauravkuppa over 1 year ago - 4 comments

#686 - InfoNCE of pre-defined anchor, positive and negative pairs

Issue - State: closed - Opened by alaaj27 over 1 year ago - 1 comment

#685 - The split of the Stanford Cars dataset.

Issue - State: closed - Opened by ppanzx over 1 year ago - 1 comment

#684 - Bug in MPerClassSampler: It does not select all classes and samples in each epoch of training

Issue - State: open - Opened by Jalilnkh over 1 year ago - 1 comment
Labels: enhancement

#683 - Neural representation similarities

Issue - State: open - Opened by domenicoMuscill0 over 1 year ago - 7 comments

#682 - Normalised embeddings when using a combination of miners and losses.

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

#681 - Contrastive loss using only indices_tuple without labels.

Issue - State: closed - Opened by fdila over 1 year ago - 2 comments
Labels: documentation

#680 - Consider importing submodules in top level __init__ and using lazy importing

Issue - State: open - Opened by KevinMusgrave over 1 year ago
Labels: enhancement

#679 - master into dev

Pull Request - State: closed - Opened by KevinMusgrave over 1 year ago

#678 - v2.4.1: use `license_files` in setup.py

Pull Request - State: closed - Opened by KevinMusgrave over 1 year ago

#677 - v2.4.0

Pull Request - State: closed - Opened by KevinMusgrave over 1 year ago

#676 - add introduction of PNP Loss

Pull Request - State: closed - Opened by interestingzhuo over 1 year ago - 3 comments

#675 - NDCGLoss

Issue - State: open - Opened by celsofranssa over 1 year ago
Labels: new algorithm request

#674 - While updating conda(installing current version), I found some erros, how to solve it?

Issue - State: closed - Opened by SC-PIONEER over 1 year ago - 2 comments
Labels: pip/conda

#673 - PNPLoss

Issue - State: closed - Opened by celsofranssa over 1 year ago - 15 comments
Labels: documentation

#672 - Use the trained model as a classifier?

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

#671 - Add MultipleNegativesRankingLoss

Pull Request - State: closed - Opened by mcschmitz almost 2 years ago - 3 comments

#670 - Default normalization in distances is counterintuitive (or wrong)

Issue - State: open - Opened by mcschmitz almost 2 years ago - 1 comment
Labels: bug, documentation

#669 - Hello , Is this package able to train a multi-label dataset with one-hot encoding?

Issue - State: open - Opened by Mahiro2211 almost 2 years ago - 2 comments
Labels: enhancement

#668 - fix test_histogram_loss

Pull Request - State: closed - Opened by GaetanLepage almost 2 years ago - 1 comment

#667 - Multi-SupConLoss

Pull Request - State: open - Opened by PhyseChan almost 2 years ago

#666 - What should be embedding size in ArcFace Loss

Issue - State: closed - Opened by ashray21 almost 2 years ago - 4 comments
Labels: question

#664 - Add MultipleNegativesRanking Loss

Issue - State: closed - Opened by mcschmitz almost 2 years ago - 2 comments
Labels: new algorithm request

#663 - Suggestion on loss function and other combinations

Issue - State: closed - Opened by jainrahulsethi almost 2 years ago - 5 comments
Labels: question

#661 - Few-shot learning

Issue - State: closed - Opened by westbalon almost 2 years ago - 1 comment
Labels: question

#660 - Fix PNP loss to make it work with negatives without related positives

Pull Request - State: closed - Opened by Puzer almost 2 years ago - 3 comments

#659 - Dynamic soft margin loss & Ranked List Loss

Pull Request - State: closed - Opened by domenicoMuscill0 almost 2 years ago - 5 comments

#658 - Multi-label version of SupConLoss

Issue - State: open - Opened by penguinwang96825 almost 2 years ago - 3 comments
Labels: new algorithm request

#657 - how to train using classfier without trainer

Issue - State: closed - Opened by sofpya almost 2 years ago - 1 comment
Labels: question

#656 - how implement combining trunk model and embedding model without pytorch metric learning-based trainer and tester

Issue - State: closed - Opened by sofpya almost 2 years ago - 15 comments
Labels: question

#653 - v2.3.0

Pull Request - State: closed - Opened by KevinMusgrave about 2 years ago

#652 - error while training

Issue - State: closed - Opened by pheersaquoff about 2 years ago - 5 comments
Labels: question

#651 - Histogram loss

Pull Request - State: closed - Opened by domenicoMuscill0 about 2 years ago - 5 comments

#650 - Adding details to NTXentLoss documentation

Pull Request - State: closed - Opened by stompsjo about 2 years ago - 1 comment

#649 - Reduction type for NTXentLoss should be "element"

Issue - State: closed - Opened by KevinMusgrave about 2 years ago - 1 comment
Labels: enhancement

#648 - Combining two loss functions

Issue - State: closed - Opened by HarryP23 about 2 years ago - 1 comment
Labels: question

#647 - Can MatchFinder implement the query library?

Issue - State: closed - Opened by ZH-UCAS about 2 years ago - 1 comment
Labels: question

#646 - Simplify reducers & mixin

Pull Request - State: open - Opened by domenicoMuscill0 about 2 years ago

#645 - Sample triplets obtained with TripletMarginMiner to reduce memory usage

Issue - State: closed - Opened by nefagi-01 about 2 years ago - 1 comment
Labels: question

#644 - Fix typo

Pull Request - State: closed - Opened by bot66 about 2 years ago - 1 comment

#643 - Fixed issues #491, #430 and #602

Pull Request - State: open - Opened by domenicoMuscill0 about 2 years ago - 3 comments

#642 - distributed training bug fix, to maintain same order for mem queue fo…

Pull Request - State: open - Opened by zhaoyuac09 about 2 years ago - 16 comments

#641 - Add information for contributors

Issue - State: open - Opened by domenicoMuscill0 about 2 years ago
Labels: documentation

#640 - Make `to_device` support list inputs

Issue - State: open - Opened by KevinMusgrave about 2 years ago
Labels: enhancement