TE
TechEcho
Home24h TopNewestBestAskShowJobs
GitHubTwitter
Home

TechEcho

A tech news platform built with Next.js, providing global tech news and discussions.

GitHubTwitter

Home

HomeNewestBestAskShowJobs

Resources

HackerNews APIOriginal HackerNewsNext.js

© 2025 TechEcho. All rights reserved.

Similarity Learning lacks a framework. So we built one

59 pointsby generallalmost 3 years ago

9 comments

flaviojuvenalalmost 3 years ago
Not a full featured framework, but pytorch-metric-learning has data loaders, lossess, etc. to facilitate similarity learning: <a href="https:&#x2F;&#x2F;github.com&#x2F;KevinMusgrave&#x2F;pytorch-metric-learning" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;KevinMusgrave&#x2F;pytorch-metric-learning</a><p>Disclaimer: I&#x27;ve made some contributions to it.
评论 #32085907 未加载
dmos62almost 3 years ago
Found the wiki article more useful in describing what Similarity Learning and Metric Learning are: <a href="https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;Similarity_learning" rel="nofollow">https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;Similarity_learning</a>
fzliualmost 3 years ago
Great article. I&#x27;ve been working in and around this space since 2014, and I think similarity learning, vector search, and embedding management will be a core part of future applications that leverage ML.<p>I recently built a similarity search application that recommends new Pinterest users channels to follow based on liked images using Milvus (<a href="https:&#x2F;&#x2F;github.com&#x2F;milvus-io&#x2F;milvus" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;milvus-io&#x2F;milvus</a>) as a backend. Similarity learning is a huge part of it, and I&#x27;m glad more and more tools like Quaterion are being released to help make this kind of tech ubiquitous.
nocturnal_ptalmost 3 years ago
Is it somehow connected to Qdrant similarity search engine? Is there a default integration for it?
评论 #32083427 未加载
usgroupalmost 3 years ago
I’m familiar with metric learning within the Mahalanobis family for kNN oriented applications . I’m not getting what use cases this framework targets? Is it custom image search type stuff which may benefit from fine tuning?<p>What is a realistic minimum viable dataset for an approach like this? When is it not advisable? How does it compare to other more basic approaches?
评论 #32086843 未加载
评论 #32087161 未加载
vervezalmost 3 years ago
Very cool. Can you comment on how this compares with tensorflow similarity? <a href="https:&#x2F;&#x2F;blog.tensorflow.org&#x2F;2021&#x2F;09&#x2F;introducing-tensorflow-similarity.html" rel="nofollow">https:&#x2F;&#x2F;blog.tensorflow.org&#x2F;2021&#x2F;09&#x2F;introducing-tensorflow-s...</a>
评论 #32083804 未加载
dmos62almost 3 years ago
I realise this is an overly-broad question, but any insight into what&#x27;s the state-of-art in Similarity Learning for article-type text?<p>More specifically, I&#x27;m interested in deriving distances between writing style, arguing style, etc.
评论 #32083921 未加载
评论 #32083617 未加载
artex_xhalmost 3 years ago
There is one <a href="https:&#x2F;&#x2F;github.com&#x2F;jina-ai&#x2F;finetuner" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;jina-ai&#x2F;finetuner</a> pretty well-designed and also gives SOTA performance from its docs
评论 #32083957 未加载
评论 #32083943 未加载
binbagalmost 3 years ago
The title is written in a clickbait format. So I had to point it out.