Lysandre: Running 30B models in colab was cert...| Robocrunch AI

Running 30B models in colab was certainly a challenge, but OPT models by @MetaAI from @stephenroller and team were a great test case for the feature! https://t.co/TWZTRH89T4
— Lysandre (@LysandreJik) May 12, 2022

https://huggingface.co/facebook/opt-30b

@MetaAI's OPT models come in flavors: from 125M to 175B params. Models up to the 30B variant are freely accessible, Accelerate v0.8 breaks t ...

Shared by Hugging Face at 5/12/2022 0-29

http://arxiv.org/abs/2202.08906

Interested in using sparse expert models, but find they are unstable, hard to design or don’t fine-tune well? We address these key issues and t ...

Shared by Barret Zoph at 2/21/2022 1-4

https://github.com/rwightman/efficientdet-pytorch

It's been a few moons since I updated my #PyTorch EfficientDet impl, but finally got around to running an EfficientNetV2 backbone experiment. I ...

Shared by Ross Wightman at 11/17/2021 23-174

At Hugging Face we have crazy engineers like @MorganFunto who are building racks of M1 Mac minis The goal is to try and build the most efficien ...

Shared by Julien Chaumond at 5/20/2022 1-18

https://github.com/jtimonen/odemodeling

The ODE model sampling was done with @mcmc_stan, and the odemodeling R package for making all this easy is available at https://t.co/luTHOLVvJ9. ...

Shared by Aki Vehtari at 5/19/2022 1-4

https://github.com/bigscience-workshop/bigscience/blob/master/train/tr11-176B-ml/chronicles-prequel.md#choosing-the-fastest-3d-topology

Doing dozens of benchmarks to choose the TP/PP/DP/MBS topology and NHEADS for the fastest throughput for the final training model @BigscienceW p ...

Shared by Stas Bekman at 3/4/2022 0-3

768