mamba VS pip

> 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.

Whenever you are working on a Python project that has external dependencies installed with pip, it is strongly recommended to first create a virtual environment.

Unfortunately that feature is easy to break: https://github.com/pypa/pip/issues/9644

I think I did my part https://github.com/pypa/pip/issues/6409

$ sudo pip install y$ sudo pip install yaml WARNING: pip is being invoked by an old script wrapper. This will fail in a future version of pip. Please see https://github.com/pypa/pip/issues/5599 for advice on fixing the underlying issue. To avoid this problem you can invoke Python with '-m pip' instead of running pip directly. ERROR: Could not find a version that satisfies the requirement yaml (from versions: none) ERROR: No matching distribution found for yaml

What are you implying will happen?

Using the build-in tools, you can save the exact versions of dependencies (i.e. a lock file) using "pip freeze >dependencies.txt". This should give you the exact same set of packages in two years' time.

If you want to be even more sure, you can also store hashes in the lock file. This has to be generated by a separate tools at the moment [1][2] but can be consumed by the built-in tools [3], so "pip install -r requirements.txt" is still all you need in two years' time.

[1] https://github.com/pypa/pip/issues/4732

[2] https://pip-tools.readthedocs.io/en/latest/#using-hashes

[3] https://pip.pypa.io/en/stable/topics/secure-installs/#hash-c...

Here’s the issue, https://github.com/pypa/pip/issues/11664. I think the idea would be to have some file/json description of environment that could be passed to pip to allow it to fully cross compile. They are open to supporting it just needs contributor to be found to implement it and go through review/discussion.

To be fair the only alternative is fixing Python, and even then you still would have to wait a good 5 years at least for all the old Python versions to dwindle.

It doesn't look like the fixing effort is progressing very quickly: https://github.com/pypa/pip/issues/8606

To their credit, at least they didn't close it "works as intended" which I imagine a lot of projects would.

Another good benchmark to trying to resolve apache-airflow[all]==1.10.13 using the state of PyPi on 2020-12-02, I give instructions here on how to reproduce that workflow: https://github.com/pypa/pip/issues/11836. Including a benchmark how how many extra packages your resolver should visit.