HNHacker News
TopNewBestAskShowJobs

Narew

126 karma · joined July 12, 2017

submissionscomments
Narew··on Scientists discover laser light can cast a shadow
It's changing the crystal not the structure but the state. Ruby crystal are gained medium for laser https://en.wikipedia.org/wiki/Ruby_laser you can pumped them with light.
Narew··on Nukefix: Simulating your nuclear weapons program
There was also this site to simulate area affected : https://www.nukethis.com/
Narew··on Vulkan Documentation
Vulkan and DirectX12 are really lower level than OpenGL. The API of Metal is high level and more beginner friendly. I really hope to have a Metal like API on top of Vulkan because currently vulkan can't replace OpenGL for beginner and simple visualisation but Metal can.
Narew··on Pretendo: A free and open source server replacement for the 3DS and Wii U
in the FAQ : Will Pretendo support the Wii/Switch?

> The Wii already has custom servers provided by Wiimmfi. We currently do not wish to target the Switch as it is both paid and completely different to Nintendo Network.

Narew··on Stable Diffusion in C/C++
Unfortunatly its not that simple expecially for NVIDIA driver and cuda install. That's why we usually use conda that can handle cuda install but even with that some time it work flawlessly and some time not.
Narew··on Keras Core: Keras for TensorFlow, Jax, and PyTorch
Same question for the decorator tf.function ?
Narew··on Keras Core: Keras for TensorFlow, Jax, and PyTorch
Yes I understand why they do the move (they want to attract pytorch user). What's the benefit for the user instead of directly using pytorch for example ? I see we can maybe use tpu by switching to jax etc...

PS: sorry I'm a bit salty by my user experience of Keras.

Narew··on Keras Core: Keras for TensorFlow, Jax, and PyTorch
Keras was already that some years ago. It supported tensorflow, theano, mxnet if my memory is right. And then they ditched everything for tensorflow. At the time it was really hard to use keras without calling backend directly for lots of optimisation, unsupported feature on they API etc... This make the use of Keras not agnostic at all. What's different now ?
Narew··on Meta Open-Sources Computer Vision Foundation Model DINOv2
Some of these models are really open source. For example segment anything is MIT licensed.
Narew··on Running JavaScript in Rust with Deno: Experimenting with Deno's Rust Crates
Maybe a naive question but why using Deno instead of v8 directly for javascript interpreter/compiler/runtime, it can also be called from rust.
Narew··on ONNX Runtime merges WebGPU backend
The ONNX format have been develop by Microsoft and Faceboof and is really well supported by pytorch as it was develop for exchange between Pytorch and Caffe2. The ONNX Runtime is use by lots of software for inference only to run on windows without having to manage specific vendor provider. (fi, WinML runtime and ONNX runtime is the same code) TVM is also use but currently I have the impression it's more widely use in embed device.

ONNX Runtime enable also provider specific inference code like TensorRT, CUDA, CoreML etc... without having you to change your code

Narew··on Codon: A Python compiler if you have a need for C/C++ speed
I have also seen some use case where python is use as a DSL to generate optimised C++/GPU code. For example Pythran can be use like that. This generated code will be use somewhere else like existing C++ codebase.
Narew··on Stochastic gradient descent written in SQL
There was also this neural network module in redis if you want to do training and inference in a redid DB. (quite old btw) https://github.com/antirez/neural-redis
Narew··on Show HN: Kuboble.com – Minimalistic sliding pieces puzzle game
It's really close to "Ricochet Robot" in its principle. https://boardgamegeek.com/boardgame/51/ricochet-robots Really fun game
Narew··on AMD Alveo V70 AI inference accelerator card
It's for inference only not training. In this use case, there is lots of device that's not Nvidia. For server you have Google tpu, for more close to public there is the Apple Neural Engine for example.
Narew··on The hype around esports is fading as investors and sponsors dry up
Maybe I'm too conservative but I don't think Counter-strike or fps in general could make on a wide broadcast. (It still people killing other people) We would have more chance on politically acceptable game like (Trackmania, Rocket League, ...) but they have smaller audience in general.
Narew··on C++ audio mixing library design
Doesn't you will have the same issue, a C-like interface. Maybe I missed something, but Rust and ffi mean you use a C interface from C++ ? The only C++<->Rust I know is either using C intermediate code or using cxx.rs but the support is limited ?
Narew··on AITemplate, a revolutionary new inference engine by Meta AI
How did AITemplate performance to state of art inference engine like tvm or onnx runtime ? Did AITemplate optimize/quantify network?

Edit: link for TVM https://tvm.apache.org/

Narew··on Fractal Generation: A Rustacean’s Prespective (2021)
that coincidence :). I just implemented this in c++ using shader yesterday. The code is probably not so clean as the goal was to learn how to use shaders. but the code is quite simple as it does only this. https://github.com/edmBernard/bg-generation-mandelbrot
Narew··on Safety: A comparaison between Rust, C++ and Go
btw, fzf is written in go ^^
Narew··on Light-shrinking material lets ordinary microscope see in super resolution
probably not, speckle is random and they use several images to reconstruct the final result.
Narew··on Light-shrinking material lets ordinary microscope see in super resolution
> As light passes through, its wavelengths shorten and scatter to generate a series of random high-resolution speckled patterns.

Speckle allow random illumination with small resolution. They reconstruct several images with differente speckle pattern to obtain better resolution on the object

Narew··on Flashlight: Fast and flexible machine learning in C++
btw, swift for tensorflow have been discontinued https://github.com/tensorflow/swift
Narew··on Flashlight: Fast and flexible machine learning in C++
Flashlight vs torch (pytorch). ^^
Narew··on Ask HN: What is the best enterprise software you use every day?
I used to have same opinion when I used only JIRA. But on my new work we use all JIRA, Stash, Crucible, Bamboo etc ... and the way all of these tools work together is really powerfull
Narew··on World’s largest steam locomotive rolls through Chicago area
We must not hate the laws that protect our privacy. IMHO. But I'm agree with the fact it can demand more works to be compliant than just block EU.
Narew··on Show HN: Datafix: connecting people with unclean data to people who clean data
But did we really want to share our unclean data to 3rd party ?
Narew··on Programmer Migration Patterns
This migration pattern remember me an other blog post https://erikbern.com/2017/03/15/the-eigenvector-of-why-we-mo...
Narew··on Text-based Template Engine for C++
see also inja[1] a c++ template engine inspired by jinja that use nlohmann/json

[1] : https://github.com/pantor/inja

Narew··on FizzBuzz in ten languages
For short solution (code golf), there is this challenge in PPCG : https://codegolf.stackexchange.com/questions/58615/1-2-fizz-...
← PreviousPage 2 of 3Next →