Pytorch-Metric-Learning

Visit Tool

pytorch-metric-learning is an open-source PyTorch library that simplifies deep metric learning. It offers modular, flexible, and extensible components for various metric learning tasks.

Claim this tool

1View

At a glance

Pricing

Open Source

Free tier

Yes

API

Yes

Skill level

Technical

About

What is pytorch-metric-learning?

pytorch-metric-learning is a comprehensive PyTorch library designed to make deep metric learning accessible and easy to implement. It provides a wide array of modules that can be used independently or combined for a complete train/test workflow, including various loss functions, miners, distances, reducers, and regularizers. The library supports unsupervised and self-supervised learning, with wrappers like SelfSupervisedLoss and features for MoCo-style self-supervision. It also includes a Datasets module for easy access to common datasets such as CUB200 and Stanford Online Products, along with trainers and testers for streamlined model development and evaluation. Its modular design allows for high customizability and integration into existing PyTorch projects.

Best used for

Ideal for developers and data scientists who need to implement deep metric learning in their PyTorch applications, customize loss functions with specific distances and reducers, and evaluate the accuracy of embedding spaces. Especially valuable for researchers and practitioners working on tasks like image retrieval and face recognition.

Common actions

implement metric learning

train deep learning models

evaluate model embeddings

customize loss functions

collaborationautomated workflowopen-sourcedeepfakelow-code/no-codeface swappinggithub copilot"AI Agents"workflows

Capabilities

Key features

Modular loss functions
Hard sample mining
Customizable distances
Self-supervised learning
Dataset module
Trainers and testers

Target Audience

developerdata scientist

Integrations

Not yet documented

Pricing & Plans

Open Source

Free

FAQs

What PyTorch versions are compatible with pytorch-metric-learning?

pytorch-metric-learning versions >= v0.9.90 require PyTorch >= 1.6. Earlier versions (< v0.9.90) were tested with PyTorch >= 1.2, but do not have a strict version requirement. It's recommended to use a recent PyTorch version for full compatibility and features.

How can I install pytorch-metric-learning with evaluation and logging capabilities?

To install with evaluation and logging, use `pip install pytorch-metric-learning[with-hooks]` for GPU support (includes faiss-gpu, record-keeper, tensorboard) or `pip install pytorch-metric-learning[with-hooks-cpu]` for CPU-only (includes faiss-cpu, record-keeper, tensorboard).

Does pytorch-metric-learning support self-supervised learning?

Yes, the library provides a `SelfSupervisedLoss` wrapper for self-supervised learning. It also includes features like `CrossBatchMemory` to implement momentum encoder queues, enabling MoCo-style self-supervision with any tuple loss and miner.

Trending

Subcategories trending in AI Agents & Automation

Chatbots & Conversational AI General-Purpose Agents Workflow Agents Personal Assistants RAG & Document AI Voice Agents

Trending

Explore

Browse AI tools by category

Content & Design Productivity & Business Coding & Development AI Agents & Automation Research & Education Wellness & Lifestyle Career Development Marketing & Growth Data & Analytics Customer Support & CX Finance E-commerce