lsorber

8 ms·

1.
by lsorber 1y ago
For those who want to dive deeper, here’s a 300 LOC implementation of GRPO in pure NumPy: https://github.com/superlinear-ai/microGRPO The implementation learns to play Battleship in about 2000 steps, pretty neat!
2.
Show HN: RAGLite – A Python package for the unhobbling of RAG(github.com)
19 points by lsorber 2y ago|0 comments
3.
by lsorber 2y ago
The name ‘late chunking’ is indeed somewhat of a misnomer in the sense that the technique does not partition documents into document chunks. What it actually does is to pool token embeddings (of a large context) into say sentence embeddings
4.
by lsorber 2y ago
You don’t have to reduce a long context to a single embedding vector. Instead, you can compute the token embeddings of a long context and then pool those into say sentence embeddings. The benefit is that each sentence’s embedding is informe
5.
by lsorber 6y ago
Did you even need the D, wouldn't a PI controller be sufficient?
6.
by lsorber 6y ago
If TSMC buys its lithography machines, why should it even get any credit for 5nm at all?
7.
by lsorber 6y ago
Could you give an example of an unsolved riddle from linguistics?
8.
by lsorber 6y ago
In my opinion, the best solution to these issues is to: 1. Declare numbers as numbers in the configuration language. E.g. "decimal(1e1000)". 2. Parse declared numbers with a lossless format like Python's decimal.Decimal. 3. L
9.
by lsorber 6y ago
Where's the data that says Moore's law no longer holds? I see comments and articles asserting this but everytime with evidence. The data that I do find certainly still suggests Moore's law is doing fine.
10.
by lsorber 6y ago
Sounds great until the client realises they can hire someone else who does charge by the hour, saving them a massive 100k - 10k = 90k compared to your proposition.
11.
by lsorber 6y ago
Huh, makes a pretty big difference for us. We were using pandas' built-in to_parquet though, which seems to suffer from some overhead.
12.
by lsorber 6y ago
Are you sure about that? It depends on how Cloudflare defines what a cold start is. It might well include the initial loading of your code, with imports and init.
13.
by lsorber 6y ago
Have you benchmarked this against pickling those data files? In our experience, parquet's overhead isn't worth it for smaller data files.
14.
by lsorber 6y ago
Looks neat. Are you considering a flake8 extension like bandit for easy adoption (in CI and in VS Code)?
15.
by lsorber 6y ago
As a counterexample: we did actually run out of file descriptors on Lambda by not closing file descriptors.
16.
by lsorber 6y ago
Can you give an example of a case outside of the happy path?
17.
by lsorber 6y ago
My experience has been the opposite: building Docker images is much easier with conda than it is with pip. With conda you can start from miniconda3, copy an environment.yml, and then conda create it. With pip, you might need to take additio
18.
by lsorber 6y ago
I sympathize with your views in that a formalisation of meaning in life would seem to take away from it somewhat. However, your response also comes across as putting a stake in the ground and thereby closing your mind as to what the utility
19.
by lsorber 6y ago
Dynaconf [1], though I think this article's approach in combination with pydantic is better. [1] https://dynaconf.readthedocs.io/en/latest/
20.
by lsorber 6y ago
In what sense has Intel vertically integrated design and fab according to you?
21.
by lsorber 7y ago
Nocedal has some papers in this direction.
22.
by lsorber 7y ago
I think you missed the 'easily'.
23.
by lsorber 7y ago
What is your preferred format for digesting content like this?
24.
by lsorber 7y ago
> Any project with CI should recreate the environment a lot. That's a very inefficient way to run your CI, with conda and pip alike. Instead, you could build your environment once in a Docker image and use that as your build image.
25.
by lsorber 7y ago
Don't forget conda ;-)
26.
by lsorber 7y ago
It's not fast on large environment.ymls, but recreating your environment isn't something you do daily so it's not that big of a deal for us.
27.
by lsorber 7y ago
And even with limited APM, the response time can still be superhuman. Bit disappointed that they didn't add some human-level latency to the agents' actions.
28.
by lsorber 7y ago
Feels recognisable. We've outgrown Lambda and moved to Fargate. But now we're wondering whether we should dive into K8s. Worth the investment in your opinion?
29.
by lsorber 7y ago
What should we be publishing to?
30.
by lsorber 7y ago
Thanks, hadn't seen this before! I wonder why they didn't have just one median function and control its behaviour with keyword arguments instead of creating 4 different median functions.

More ›