l-m-z

6 ms·

1.
by l-m-z 1y ago
Hibiki is an auto-regressive model with temperature based sampling so very similar to a LLM, generations are "random" and you can make them deterministic by fixing the RNG seed.
2.
by l-m-z 2y ago
Hi swyx, laurent from kyutai here. We actually used the online demo at moshi.chat for the live event (the original demo), so same quantization. We updated the weights on the online version since then to add support for more emotions but we
3.
Moshi: A voice AI that can display emotions(twitter.com)
26 points by l-m-z 2y ago|1 comments
4.
by l-m-z 2y ago
Candle dev here, we also support training/backdrop! We certainly focus on optimizing inference performance but hopefully that should improve the training efficiency too.
5.
Show HN: MetaVoice-1B on Candle, local text-to-speech in Rust with GPU support(twitter.com)
3 points by l-m-z 3y ago|0 comments
6.
Fast inference for text models using Rust(github.com)
3 points by l-m-z 3y ago|0 comments
7.
by l-m-z 3y ago
The tensors should be Send and Sync so can be manipulated from multiple threads, the underlying data is protected by a RWLock to guard against data races. The heavy operations such as matrix multiplication will be run on multiple cores even
8.
by l-m-z 3y ago
Another random (self) plug for a rust version, this uses the candle ML library we've been working on for the last month and can be run in the browser. https://laurentmazare.github.io/candle-llama2/index.html The n
9.
An OCaml port of min-DALLE(github.com)
1 points by l-m-z 4y ago|0 comments
10.
Show HN: Stable Diffusion implementation in Rust/libtorch(github.com)
4 points by l-m-z 4y ago|0 comments
11.
Using Python and OCaml in the same Jupyter notebook(blog.janestreet.com)
165 points by l-m-z 7y ago|34 comments
12.
by l-m-z 7y ago
This is from yesterday's (June the 15th) workshop on self-supervised learning at ICML. The video of this talk can be seen here: https://www.facebook.com/icml.imls/videos/2030095370631729/ (not sure if th
13.
Playing Atari Games with OCaml and Deep Reinforcement Learning(blog.janestreet.com)
8 points by l-m-z 8y ago|0 comments
14.
by l-m-z 8y ago
(I'm obviously biased as being the author of ocaml-torch) Using functional programming is not of much help for size mismatches, the current bindings don't even use the type system to check the number of dimensions - although that
15.
by l-m-z 8y ago
Bindings author here. This should support most of the PyTorch ops - this is done by automatically generating the binding code as there are more than a thousand of these. Most of the models available in torchvision should also be there (with