SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1912.00385 · 2019

The Group Loss for Deep Metric Learning

arXiv · PDF · Open in the Atlas

Code that ran

We lifted 18 functions out of this paper's own repositories and ran 12 of them in a sandbox. "Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results.

RepositoryRoleRan
dvl-tum/group_loss canonical 1 of 1
laurinwagner/grouploss_plus pwc_unofficial 11 of 17
FunctionStatusWhere it lives
bn_inception Ran laurinwagner/grouploss_plus/net/inception_bn.py
code served (permissive licence) · get_code("650c1166736654f2")
conv1x1 Ran laurinwagner/grouploss_plus/net/resnet.py
code served (permissive licence) · get_code("d9def42110729a85")
conv3x3 Ran laurinwagner/grouploss_plus/net/resnet.py
code served (permissive licence) · get_code("160bb14bd76201b4")
dynamics Ran laurinwagner/grouploss_plus/dynamics.py
code served (permissive licence) · get_code("0fa98ac2194783f3")
get_W_gt Ran laurinwagner/grouploss_plus/gtg.py
code served (permissive licence) · get_code("5efaf143a9a72186")
get_sim_pairs Ran laurinwagner/grouploss_plus/gtg.py
code served (permissive licence) · get_code("d574dd05dd75020c")
make_embedding_layer Ran laurinwagner/grouploss_plus/net/embed.py
code served (permissive licence) · get_code("4e716cb135e5eb7c")
predict_batchwise Ran laurinwagner/grouploss_plus/utils.py
code served (permissive licence) · get_code("21bfe6f420cc452c")
predict_batchwise_inshop Ran laurinwagner/grouploss_plus/utils.py
code served (permissive licence) · get_code("b8c440cba313c634")
re_rank Ran laurinwagner/grouploss_plus/utils.py
code served (permissive licence) · get_code("d477cdb6933c981b")
rnd Ran dvl-tum/group_loss/train_finetune.py
code served (permissive licence) · get_code("e9ea44d93d8ffa92")
update_probs Ran laurinwagner/grouploss_plus/combine_sampler.py
code served (permissive licence) · get_code("108396eb628e18c6")
densenet121 Not yet run laurinwagner/grouploss_plus/net/densenet.py
code served (permissive licence) · get_code("662688cdde7e85be")
densenet161 Not yet run laurinwagner/grouploss_plus/net/densenet.py
code served (permissive licence) · get_code("a15fc654c671c2bf")
densenet169 Not yet run laurinwagner/grouploss_plus/net/densenet.py
code served (permissive licence) · get_code("2f494ef1c2bb7bab")
embed Not yet run laurinwagner/grouploss_plus/net/embed.py
code served (permissive licence) · get_code("757fc132892039c3")
resnet18 Not yet run laurinwagner/grouploss_plus/net/resnet.py
code served (permissive licence) · get_code("9771a961f7827f37")
thresh_func Not yet run laurinwagner/grouploss_plus/gtg.py
code served (permissive licence) · get_code("a087db29a375ae43")

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Deep metric learning has yielded impressive results in tasks such as clustering and image retrieval by leveraging neural networks to obtain highly discriminative feature embeddings, which can be used to group samples into different classes. Much research has been devoted to the design of smart loss functions or data mining strategies for training such networks. Most methods consider only pairs or triplets of samples within a mini-batch to compute the loss function, which is commonly based on the distance between embeddings. We propose Group Loss, a loss function based on a differentiable label-propagation method that enforces embedding similarity across all samples of a group while promoting, at the same time, low-density regions amongst data points belonging to different groups. Guided by the smoothness assumption that "similar objects should belong to the same group", the proposed loss trains the neural network for a classification task, enforcing a consistent labelling amongst samples within a class. We show state-of-the-art results on clustering and image retrieval on several datasets, and show the potential of our method when combined with other techniques such as ensembles

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("1912.00385")
get_code_for_paper("1912.00385")
have("1912.00385")

Connect an agent — have() is free.