search

DeepX-inc / machina

Control section: Deep Reinforcement Learning framework
MIT License
279 stars 43 forks source link

max_ent in example/run_sac.py is not used #255

Open caprest opened 5 years ago

caprest commented 5 years ago

https://github.com/DeepX-inc/machina/blob/caf2e51859e764876963f8a24982607c0774e3b2/example/run_sac.py#L60-L61 max_ent params in the example seems not used.

https://github.com/DeepX-inc/machina/blob/caf2e51859e764876963f8a24982607c0774e3b2/machina/algos/sac.py#L3 Furthermore, algo/sac.py refers to https://arxiv.org/pdf/1801.01290.pdf as a reference of algorithm but the algorithm uses temperature adjustment technique not used in the paper. Probably adding https://arxiv.org/pdf/1812.05905.pdf is better.

Thanks.

rarilurelo commented 5 years ago

Thanks for pointing out an unused argument and the correct reference! Could you send PR for these problems, if you'd like?