AISmolDataEnvs launches with 5,000-plus reinforcement-learning tasks
The release announcement describes verifiable tasks for improving small models in code and data science, with environments, evaluations and training all open source.
TLDR
SmolDataEnvs’ release announcement introduces 5,000-plus verifiable reinforcement-learning tasks aimed at improving small models in code and data science. It says the environments, evaluations and training are completely open source.
SmolDataEnvs launches with 5,000-plus reinforcement-learning tasks
The release announcement describes verifiable tasks for improving small models in code and data science, with environments, evaluations and training all open source.
