opal
por xingjian-bai
OPAL: a reactive optimizer. A coding agent rewrites the optimizer while your model trains.
Do README
OPAL: a reactive optimizer A PyTorch optimizer wrapper that lets a coding agent change the optimizer while your model trains. Blog post · Paper · alphaXiv · Guide · Start with your agent · Examples Most networks are trained with an optimizer that is fixed before training starts. OPAL (OPtimizer with an Agent in the Loop) turns the optimizer into a program that an agent keeps rewriting during the run. You wrap your PyTorch optimizer in one line. Every so often the agent (Claude Code) reads what the run reports and decides whether to rewrite two small programs: the optimizer program, which…
Trecho do README do projeto no GitHub.
Skills
Um projeto open-source — explore o código e faça o self-hosting a partir do GitHub.