feat: CrossQ algorithm — SAC without target networks#533

Merged

kengz merged 38 commits intomasterfrom

feat/improvements-roadmap

Mar 4, 2026

Commits on Feb 28, 2026

refactor: framework improvements — PPO, nets, memory, plotting, utils

kengz
and
claude
committed
feat: CrossQ algorithm — SAC without target networks + BatchRenorm

kengz
and
claude
committed
feat: CrossQ benchmark specs for all environments

kengz
and
claude
committed
docs: CrossQ benchmark results, plots, and documentation

kengz
and
claude
committed
fix: update dstack configs for 0.20.x compatibility

kengz
and
claude
committed
feat: CrossQ improvement specs for ⚠️ envs (InvPend, InvDblPend, Hopper)

kengz
and
claude
committed
docs: update CrossQ classic/box2d benchmark scores and plots

kengz
and
claude
committed

Commits on Mar 1, 2026

Commits on Mar 2, 2026

Commits on Mar 3, 2026