seungeunrho / minimalRL

Implementations of basic RL algorithms with minimal lines of codes! (pytorch based)
MIT License
2.84k stars 457 forks source link

Wrong formula for calc-target in SAC? #63

Open BeFranke opened 1 month ago

BeFranke commented 1 month ago

See https://github.com/seungeunrho/minimalRL/blob/c8efed8481e3cd40e9739cfde220a55522555b57/sac.py#L127C1-L127C54

Shouldn't the formula be target = r + gamma * (1 - done) * (min_q + entropy)?