Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Similarity Learning lacks a framework. So we built one

blog.qdrant.tech

Post reply on HN
Print view View original

11–20 of 25 posts

Re: Similarity Learning lacks a framework. So we built one

#11

Post by monatis »

dmos62 wrote: post #6

I realise this is an overly-broad question, but any insight into what's the state-of-art in Similarity Learning for article-type text? More specifically, I'm interested in deriving distances between writing style, arguing style, etc.

There's study here: http://cs230.stanford.edu/projects_spring_2021/reports/57.pd...

Basically, you can collect text from different authors, then you can use authors names as labels to train a similarity learning with it. My suggestion would be finetune a Transformer model with a specific head and an ArcFace loss.

Re: Similarity Learning lacks a framework. So we built one

#12

Post by kacperlukawski »

artex_xh wrote: post #10

There is one https://github.com/jina-ai/finetuner pretty well-designed and also gives SOTA performance from its docs

Unfortunately, it's no longer open source, but requires using their cloud.

"From 0.5.0, Finetuner computing is hosted on Jina Cloud. THe last local version is 0.4.1, one can install it via pip or check out git tags/releases here."

Re: Similarity Learning lacks a framework. So we built one

#13

Post by generall »

artex_xh wrote: post #10

There is one https://github.com/jina-ai/finetuner pretty well-designed and also gives SOTA performance from its docs

Starting from 0.5.0 finetuner is no longer an open-source.

> From 0.5.0, Finetuner computing is hosted on Jina Cloud. The last local version is 0.4.1, one can install it via pip or check out git tags/releases here.

But there are some cool ideas implemented there as well, I encourage you to try both!

Re: Similarity Learning lacks a framework. So we built one

#14

Post by artex_xh »

kacperlukawski wrote: post #12
artex_xh wrote: post #10

There is one https://github.com/jina-ai/finetuner pretty well-designed and also gives SOTA performance from its docs

Unfortunately, it's no longer open source, but requires using their cloud. "From 0.5.0, Finetuner computing is hosted on Jina Cloud. THe last local version is 0.4.1, one can install it via pip or check out git tags/releases here."

but does being "opensource" naturally make a software good in quality & performance? In the end, people try to solve a problem right?

Re: Similarity Learning lacks a framework. So we built one

#15

Post by artex_xh »

artex_xh wrote: post #14
kacperlukawski wrote: post #12

Earlier quoted context omitted.

Unfortunately, it's no longer open source, but requires using their cloud. "From 0.5.0, Finetuner computing is hosted on Jina Cloud. THe last local version is 0.4.1, one can install it via pip or check out git tags/releases here."

but does being "opensource" naturally make a software good in quality & performance? In the end, people try to solve a problem right?

anyway, congrats on the release :P and I upvoted! so we are good.

Re: Similarity Learning lacks a framework. So we built one

#16

Post by kacperlukawski »

artex_xh wrote: post #14
kacperlukawski wrote: post #12

Earlier quoted context omitted.

Unfortunately, it's no longer open source, but requires using their cloud. "From 0.5.0, Finetuner computing is hosted on Jina Cloud. THe last local version is 0.4.1, one can install it via pip or check out git tags/releases here."

but does being "opensource" naturally make a software good in quality & performance? In the end, people try to solve a problem right?

Of course, solving a problem is the most important thing at the end of the day. However, there are some data privacy constraints you may sometimes need to fulfil, and sending the data to an external cloud, managed by 3rd party is not an option.

Moreover, fine-tuning might be just one of the applications of neural networks in the organization, and you may already have some pipelines built to train them, so it should be also unified.

And more importantly, Jina's finetuner gives you some pretrained models to choose from, while Quaterion is PyTorch Lightning based, so you can easily integrate it if you already use PyTorch and have the flexibility to fine-tune any custom network as well.

Re: Similarity Learning lacks a framework. So we built one

#18

Post by dimitry12 »

flaviojuvenal wrote: post #17

Not a full featured framework, but pytorch-metric-learning has data loaders, lossess, etc. to facilitate similarity learning: https://github.com/KevinMusgrave/pytorch-metric-learning Disclaimer: I've made some contributions to it.

Seconding this. PML is high quality, active, and well documented.

Re: Similarity Learning lacks a framework. So we built one

#19

Post by generall »

dimitry12 wrote: post #18
flaviojuvenal wrote: post #17

Not a full featured framework, but pytorch-metric-learning has data loaders, lossess, etc. to facilitate similarity learning: https://github.com/KevinMusgrave/pytorch-metric-learning Disclaimer: I've made some contributions to it.

Seconding this. PML is high quality, active, and well documented.

PML is a great collection of implementations, but not the best framework. Also you can use PML with Quaterion: https://github.com/qdrant/quaterion/blob/master/examples/tra...

Re: Similarity Learning lacks a framework. So we built one

#20

Post by usgroup »

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?

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?

Post reply on HN