mamba VS Zappa

> how the recall can grow unbounded with no tradeoff

this? https://github.com/state-spaces/mamba/issues/175

If you want to learn this stuff as a computer engineer, you can read the code here [0]. I find the math quite helpful.

[0]: https://github.com/state-spaces/mamba

This model was discussed recently: https://news.ycombinator.com/item?id=38522428 It's a new kind of ML model architecture that can be used instead of a transformer in LLMs.

See also the original repo from the paper: https://github.com/state-spaces/mamba

[2] - https://github.com/state-spaces/mamba

Out of curiosity, does anyone feel as though there's any benefit to linking to reddit when we can link to whatever the link is? I for one do not click the link and read discussion on reddit - if I wanted that sort of discussion, I would browse there, not HN.

The library is compatible with any auto-regressive model, not transformers. To prove our point we integrated Mamba, a new state-space model architecture, to the library. Try it out!

I ran the NanoGPT of Karparthy replacing Self-Attention with Mamba on his TinyShakespeare Dataset and within 5 minutes it started spitting out the following:

You might have come across the paper Mamba paper in the last days, which was the first attempt at scaling up state space models to 2.8B parameters to work on language data.

If people aren't familiar, there's a similar project for Python that's fantastic: https://github.com/zappa/Zappa

Chalice was a consequence, a reaction from AWS to the release of (Zappa Framework)[https://github.com/zappa/Zappa] that provide a very good alternative to migrate very quickly a Django/Flask or any WSGI compliant solution in Python.

Use Zappa https://github.com/zappa/Zappa and host as a Lambda, simple setup and deployment, Lambda only costs when processing requests, no servers to mess around with

One of my favorite options that is probably the most cost-effective is to deploy using a 'severless' model on AWS Lambda using zappa which supports deploying Python webapps to AWS in this way. Zappa also makes it super easy to deploy in just a couple commands! The README includes instructions for everything you might need, including handling sensitive information like your database passwords, running django management commands, setting up DNS, etc.

Deployment works nicely with Docker. I often use AWS AppRunner because it's really easy and just scales. Some people use AWS Lambda with Zappa but I don't recommend it unless you really want to spend less than $15 a month. You will probably need Django Storages to save uploads to an S3 bucket. At some stage you might want to put a CloudFront distribution in front of everything but the configuration of the caching behaviour might be a bit confusing when you do it the first time.

It doesn't have to be an either-or! I have several apps in production that were developed on Django or Flask, and deployed to Lambda using Zappa.

If you need a relational DB, you can use AWS Aurora or RDS and use cloud functions ('lambda' in AWS) that you can invoke with HTTP to process the document first. Zappa will do a lot of the configuration for you if you go that route.

Lambda + API gateway, this library bundles a Django application into a lambda https://github.com/zappa/Zappa . 1 million free invokes from aws, scale to zero, plugs into your RDS

I use Zappa, it just schedules a frequent execution of the lambda: https://github.com/zappa/Zappa#keeping-the-server-warm