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.

Towards Reproducible Research with PyTorch Hub

109 pointsby rstojalmost 6 years ago

3 comments

kradroyalmost 6 years ago
I love that the tooling for ML experimentation is becoming more mature. Keeping track of hyperparameters, training&#x2F;validation&#x2F;test experiment test set manifests, code state, etc is both extremely crucial and extremely necessary. I can&#x27;t count how many times I&#x27;ve trained a great model only to lose the exact state and be unable to reproduce it. It&#x27;s extremely frustrating. When I found sacred (<a href="https:&#x2F;&#x2F;github.com&#x2F;IDSIA&#x2F;sacred" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;IDSIA&#x2F;sacred</a>) it changed my team&#x27;s workflow in a very positive way. We already have this approach of saving default experiment workbenche images, but formalizing it is much nicer.
评论 #20152859 未加载
samzeralmost 6 years ago
This is like an app store of ml models which is pretty cool. There are a couple of tooling around ML that more of compliments the above than being redundant.<p>-- Open Source --<p>Mlflow : <a href="https:&#x2F;&#x2F;github.com&#x2F;mlflow&#x2F;mlflow" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;mlflow&#x2F;mlflow</a><p>Polyaxon : <a href="https:&#x2F;&#x2F;github.com&#x2F;polyaxon&#x2F;polyaxon" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;polyaxon&#x2F;polyaxon</a><p>Modelchimp(mine): <a href="https:&#x2F;&#x2F;github.com&#x2F;ModelChimp&#x2F;modelchimp" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;ModelChimp&#x2F;modelchimp</a><p>ModelDB : <a href="https:&#x2F;&#x2F;github.com&#x2F;mitdbg&#x2F;modeldb" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;mitdbg&#x2F;modeldb</a><p>Sacred : <a href="https:&#x2F;&#x2F;github.com&#x2F;IDSIA&#x2F;sacred" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;IDSIA&#x2F;sacred</a><p>-- Non open source --<p>Cometml : <a href="https:&#x2F;&#x2F;www.comet.ml&#x2F;" rel="nofollow">https:&#x2F;&#x2F;www.comet.ml&#x2F;</a><p>Weights and Biases : <a href="https:&#x2F;&#x2F;www.wandb.com&#x2F;" rel="nofollow">https:&#x2F;&#x2F;www.wandb.com&#x2F;</a><p>MissinglinkAi : <a href="https:&#x2F;&#x2F;missinglink.ai&#x2F;" rel="nofollow">https:&#x2F;&#x2F;missinglink.ai&#x2F;</a>
panpannaalmost 6 years ago
This will also make things easier for people writing algorithms on top of one of the base models.<p>You start with something simple but fast like resnet18 and once the general aproach works you replace it with something better&#x2F;slower by changing a single line.