https://jax-ml.github.io/scaling-book/gpus/

Compute-Optimal Scaling for Value-Based Deep RL https://arxiv.org/abs/2508.14881 https://www.alphaxiv.org/ru/overview/2508.14881v1

ComputerRL: Scaling End-to-End Online Reinforcement Learning for Computer Use Agents https://arxiv.org/abs/2508.14040…

Сегодня в 22:00 мск (через час) будем презентовать нашу статью Quartet: Native FP4 Training Can Be Optimal for Large Language Models на GPU MODE. Залетайте.

OS-R1: Agentic Operating System Kernel Tuning with Reinforcement Learning https://arxiv.org/abs/2508.12551 https://www.alphaxiv.org/ru/overview/2508.12551v1

https://github.com/tiny-tpu-v2/tiny-tpu