profile
viewpoint
Naoki Kobayashi knao124 FiNC Tokyo, Japan

knao124/Bond 0

A Swift binding framework

knao124/Clustering-with-Deep-learning 0

Generic implementation for clustering with deep learning : representation learning (DNN) + clustering

knao124/factory_girl_rails 0

Factory Girl ♥ Rails

knao124/Gasyori100knock 0

画像処理100本ノックして画像処理を画像処理して画像処理するためのもの For Japanese, English and Chinese

knao124/grape-active_model_serializers 0

User active_model_serializers with Grape

knao124/keras-retinanet 0

Keras implementation of RetinaNet object detection.

knao124/py-faster-rcnn 0

Faster R-CNN (Python implementation) -- see https://github.com/ShaoqingRen/faster_rcnn for the official MATLAB version

knao124/sagemaker-python-sdk 0

A library for training and deploying machine learning models on Amazon SageMaker

startedxiangli13/circle-loss

started time in 12 days

startedTinyZeaMays/CircleLoss

started time in 12 days

issue openedKevinMusgrave/pytorch-metric-learning

Is there any filtering functionalities on triplet sampling?

I'd like to ask a question about how to implement sampling with this package in the following situations.

Now I have the following data. - samples with the label "A" - samples with the label "B". - samples with the label "C". - samples with the label "not A".

"not A" means "B or C, or an unknown label other than A~C".

The sampling I want to implement is when a sample labeled "A" is an anchor, only the sample labeled "not A" is used for the negative. In other words, I don't want to use samples labeled "not A" as anchor. I want to use them only as negative.

In this case, how should I implement it using pytorch-metric-learning?

created time in 13 days

startedKevinMusgrave/pytorch-metric-learning

started time in 18 days

startedpfnet/pytorch-pfn-extras

started time in 18 days

more