Jing-Cheng Pang

icon indicating a website link https://www.lamda.nju.edu.cn/pangjc icon indicating an email [email protected]

icon company Nanjing University icon location NanJing, Jiangsu, China icon location Ph.D. student at Nanjing University. Interested in reinforcement learning.

Results 5 comments of Jing-Cheng Pang

TypeError: cuda() got an unexpected keyword argument 'device_id'

Just remove device_id from model.cuda() with args.gpu remaining.

KeyError '_agent_name' with ex_elfpy.py

It's seemed that adding 'b' before '_agent_name' will be ok.

Is there any existing implement of training with different type RL models?

Here is a toy example for 2 players loading different models. You can check it at ./rlpytorch/model_loader.py https://github.com/lafmdp/ELF

train MiniRTS is slow - no information when will end

Maybe you can control it at ./rlpytorch/runner/single_process.py

Installation issue related to mjrl

Running this command `git config --global url."https://".insteadOf git://` could solve the problem for me.