maddy VS pytorch-lightning

It's interesting how there is now

* Maddy: https://github.com/foxcpp/maddy

* Mox: https://github.com/mjl-/mox

* and Stalwart

which all see to aim for more or less the same niche. I wonder if we'll see two of those merge eventually.

Does anybody run maddy (https://github.com/foxcpp/maddy) in production or for personal matters?

Yeah, that's a bummer. If you need the docs, then its here.

I've been using maddy mail server as it's quick and easy to setup for over a year now

Both my home server and VPS have unattended upgrades with email notifications about faliures that override do not disturb on my phone. acme.sh has automatic TLS renewal configured with systemd. Maddy mail server itself runs on Docker that is also set up with systemd to automatically restart and start at reboots. When it comes to updating Maddy i just keep an eye on the release Atom feed and change the docker compose file and restart the container.

I wonder how does it compare to Maddy mail server: https://github.com/foxcpp/maddy

If you want to self hosted, I strongly recomend https://github.com/foxcpp/maddy since it's just a single binary to run.

If you still want to do it on your own, there is Go simple mail server - maddy where you can borrow some code or simply use it for the purpose as a IMAP server (I mentioned above, but dovecot as IMAP probably would be the best)

https://github.com/emersion has great libraries related to emails, for example go-imap. https://github.com/foxcpp/maddy is an email server written in Go that you could look into.

Check ready-to-go maddy-SMTP/IMAP server

It's very easy to get started, right in your Terminal, no fees! No credit card at all.

And there are cloud providers like https://replicate.com/ and https://lightning.ai/ that will let you use your LLM via an API key just like you did with OpenAI if you need that.

You don't need OpenAI - nobody does.

https://see.stanford.edu/Course/CS229 https://lightning.ai/ https://www.youtube.com/watch?v=00s9ireCnCw&t=57s https://towardsdatascience.com/

I've been wondering on what is the recommended method of saving logits/activations using PyTorch Lightning. I've looked at Callbacks, Loggers and ModelHooks but none of the use-cases seem to be for this kind of activity (even if I were to create my own custom variants of each utility). The ModelCheckpoint Callback in its utility makes me feel like custom Callbacks would be the way to go but I'm not quite sure. This closed GitHub issue does address my issue to some extent.

Our codebase for the diffusion models builds heavily on OpenAI's ADM codebase , lucidrains, Stable Diffusion, Lightning and Hugging Face. Thanks for open-sourcing!

For lightweight use, pytorch-lightning is too heavy, and its source code will be very difficult for beginners to read, at least for me.

For the A100 GPU, theoretical performance is the same for FP16/BF16 and both rely on the same number of bits, meaning memory should be the same. However since it's quite newly added to PyTorch, performance seems to still be dependent on underlying operators used (pytorch lightning debugging in progress here).