egnehots

9 ms·

31.
Fast Development in Rust, Part One(blog.sdf.com)
3 points by egnehots 3y ago|0 comments
32.
by egnehots 3y ago
Yes, that's a hard issue for arbitrary/user-provided HTML pages. But with templates under your control, the context is different. Your designers do not have to trust Chrome; they can preview, tweak using print media queries, and p
33.
Ella: Equip Diffusion Models with LLM for Enhanced Semantic Alignment(ella-diffusion.github.io)
5 points by egnehots 3y ago|0 comments
34.
by egnehots 3y ago
The main issue is conflating templating and pdf generation. Using html to pdf solutions allow to do the templating in html, where it is pretty much a solved issue. And as many said, headless chrome is a robust html to pdf solution, even tho
35.
Luminal: Fast ML in Rust through graph compilation(github.com)
2 points by egnehots 3y ago|0 comments
36.
by egnehots 3y ago
What about clojure? IntelliJ & VS Code are also popular IDEs for it. It also support datastructure literals vec [1 2 3], map {:a 1 :b 2} which are mostly considered as helpful for beginners since it's closer to what they are used t
37.
by egnehots 3y ago
it's based on the JVM which uses UTF-16 internally for strings
38.
by egnehots 3y ago
from their doc, Reasons to use Pingora: - Security is your top priority: Pingora is a more memory safe alternative for services that are written in C/C++. - Your service is performance-sensitive: Pingora is fast and efficient. - Your s
39.
Why the Future of Cloud Deployments Is Unikernels (2023)(kraft.cloud)
12 points by egnehots 3y ago|5 comments
40.
by egnehots 3y ago
also: https://unikraft.org/docs/concepts
41.
by egnehots 3y ago
for devops, datascience and ml stuff Python is a good contender
42.
Extending Rust's Effect System(blog.yoshuawuyts.com)
151 points by egnehots 3y ago|24 comments
43.
PyTorch 2.2: FlashAttention-v2 integration, AOTInductor(pytorch.org)
15 points by egnehots 3y ago|1 comments
44.
One PID to Lock Them All: Finding the Source of the Lock in Postgres(crunchydata.com)
2 points by egnehots 3y ago|0 comments
45.
Vera: Vector-Based Random Matrix Adaptation(dkopi.github.io)
1 points by egnehots 3y ago|3 comments
46.
by egnehots 3y ago
VeRA makes LoRA ~10x more parameter efficient while retaining the same performance. It's somewhat like a recursive LoRA scheme, where the LoRA A and B matrices are also decomposed using two small trainable vector parameters.
47.
Stable Code 3B: Coding on the Edge(stability.ai)
315 points by egnehots 3y ago|140 comments
48.
Debugging a Bit-Flip Error(sillycross.github.io)
1 points by egnehots 3y ago|0 comments
49.
by egnehots 3y ago
There was recently a very interesting overview of the different distributed PostgreSQL architectures: https://www.crunchydata.com/blog/an-overview-of-distributed-...
50.
Amazon lays off 500 Twitch employees, hundreds more at MGM and Prime Video(arstechnica.com)
46 points by egnehots 3y ago|57 comments
51.
by egnehots 3y ago
No, it can indeed be vastly simpler to handle auth. But you have to consider many complex aspects to provide a simple solution: - What are your real use cases (authentication, authorization, delegation?) - What is your threat model? (avoidi
52.
by egnehots 3y ago
there are pros and cons to both verb-first and subject-first commit styles. verb-first commits can be organised in a few categories. since they are few commit verbs (fix, add, refact..), always use the same verb prefix for bug, features mes
53.
by egnehots 3y ago
interesting but not very practical? let's take the max_wal_size parameter for ex. for most systems it's much simpler to set the max_wal_size to a value high enough that its value does not put pressure for checkpoints under load. s
54.
by egnehots 3y ago
Yes, but on the other hand, people who don't care can create huge images (especially in Node.js and Java ecosystems) that slow down CI and have a large attack surface.
55.
Beyond Backpropagation, Automatic Differentiation in Rust(getcode.substack.com)
4 points by egnehots 3y ago|0 comments
56.
by egnehots 3y ago
This paper combines knowledge distillation and efficient attention mechanisms. => It works (still efficient, lower cost). Not an unexpected result, but to their credit, they established a new benchmark to test these combinations. KD+Long
57.
by egnehots 3y ago
I'm concerned that many might miss the nuance and subtlety of the main issue: - Nicotine IS really addictive when distributed through common uses (smoking, vaping). Yes, if the absorption is progressive and in small dosages, the addi
58.
by egnehots 3y ago
The tokio ecosystem is also a world of its own. Tokio internals are plentiful and more complex than in other runtimes. But the common difficulty is choosing what to use for an http server. - hyper? just an http library but bring your own bo
59.
by egnehots 3y ago
As a Rust user, I have to say that it was an incredibly confusing feature. Not because of the technical decisions behind async/await, but because of the async ecosystem, especially the runtimes. Picture a Rust user when the feature cam
60.
Making code reusable is not an art, it's a job(acko.net)
1 points by egnehots 3y ago|0 comments

More ›