Last week in tech news, DeepMind to control nuclear fusion with reinforcement learning, the 4M model from Google on how to reduce carbon emission owing to machine learning, Yann LeCun's world model on next generation autonomous intelligence, several new checkpoints of T5, new library on image sketching, Anomalib providing new benchmark for DNN models and the Textless NLP library from Meta AI.

Last week in tech news, we have the release of under the hud of Meta research center, Open AI again with their newly trained GPT-3 model, google's new language dialog model namely "LaMDA", integration of Stable-baselines3 into hugging-face, EvoGym —a platform for the benchmarking of reinforcement learning models and more, and a library on Transformer Reinforcement Learning.

Last week in tech news, we have DeepETA from Uber, a new trained model from OPEN AI, one new platform on reinforcement learning, one new python library for deep learning on sensitive data, coming up new documentation from Gym, a comprehensive PhD thesis on neural differential equations and a new library TorchMetrics for metric categorization in pytorch.