leonardovvla / multi-agent-cooperation-learning

This is a project based on OpenAI's multi-agent-emergence-environments (Emergent Tool Use from Multi-Agent Autocurricula, Baker et al.), and uses its framework to build a new environment (MACL) where developers can deploy their own experiments in order to study good ways to foster cooperation intelligence in machines. This work was done during a Semester Project at ETH Zürich and I hope it can serve as additional material to support engineers and researchers in using OpenAI's original code to develop their own projects. For more information, a report is attached below under the name "Cooperation Learning in multi-agent systems".
10 stars 3 forks source link

Could you specify steps of running the code #1

Open gaoyuankidult opened 3 years ago

gaoyuankidult commented 3 years ago

Dear Leonardo Albuquerque

Could you specify in the README file of how to run your code?

BitastaB commented 3 years ago

I could train it by running Training/ma_run.py. I think he also mentions it in either the project report or in the code documentation. Although I am not able to test it yet :(

leonardovvla commented 3 years ago

Hi everyone! BitastaB is correct: by running python ma_run.py in the terminal, you can run a training sesison with the ppo parameters specified in ma_run.py.

pythlang commented 3 years ago

I'm unable to run it; I'm getting 'cannot import' or 'no module named' errors even though they are there. Why can't I use the modules? Do you have any advice for me to try?

alan303138 commented 2 years ago

How can i test it? Missing file test_policy_ppo.py

SpinLeo commented 2 years ago

How can i test it? Missing file test_policy_ppo.py

兄弟,我也去跑了你的代码,救我!!!跑完那个ma_run.py后也没可视化界面啊,那还需要改什么地方,我是在虚拟机上面跑的

alan303138 commented 2 years ago

How can i test it? Missing file test_policy_ppo.py

兄弟,我也去跑了你的代码,救我!!!跑完那个ma_run.py后也没可视化界面啊,那还需要改什么地方,我是在虚拟机上面跑的

我是win10跑的,所以有些依賴項會搬來搬去,其他檔案都可以照原作者,我只多加了\spinup\utils\test_policy_PPO.py這個檔案,還有修改spinup.utils.logx.restore_tf_graph 其他應該都一樣。 你這邊是訓練完要test了嗎? openAi<<這個repo執行 python bin/examine.py examples/hide_and_seek_quadrant.jsonnet examples/hide_and_seek_quadrant.npz 有辦案法跑出可視化MuJoCo引擎嗎? 有沒有報什麼錯誤?

SpinLeo commented 2 years ago

How can i test it? Missing file test_policy_ppo.py

兄弟,我也去跑了你的代码,救我!!!跑完那个ma_run.py后也没可视化界面啊,那还需要改什么地方,我是在虚拟机上面跑的

我是win10跑的,所以有些依賴項會搬來搬去,其他檔案都可以照原作者,我只多加了\spinup\utils\test_policy_PPO.py這個檔案,還有修改spinup.utils.logx.restore_tf_graph 其他應該都一樣。 你這邊是訓練完要test了嗎? openAi<<這個repo執行 python bin/examine.py examples/hide_and_seek_quadrant.jsonnet examples/hide_and_seek_quadrant.npz 有辦案法跑出可視化MuJoCo引擎嗎? 有沒有報什麼錯誤?

我去跑了一下,就是在那个multi-agent-emergence-environments-master里面跑,用了你的命令跑出来报错 TypeError: join() argument must be str or bytes , not 'tuple' ,我现在也不知道改怎么训练,只是在现在这个项目下跑了那个ma_run.py文件。 你的意思是说这个cooperation-learning这个项目要和multi-agent-emergence-environments-这个项目一起跑吗?

SpinLeo commented 2 years ago

How can i test it? Missing file test_policy_ppo.py

兄弟,我也去跑了你的代码,救我!!!跑完那个ma_run.py后也没可视化界面啊,那还需要改什么地方,我是在虚拟机上面跑的

我是win10跑的,所以有些依賴項會搬來搬去,其他檔案都可以照原作者,我只多加了\spinup\utils\test_policy_PPO.py這個檔案,還有修改spinup.utils.logx.restore_tf_graph 其他應該都一樣。 你這邊是訓練完要test了嗎? openAi<<這個repo執行 python bin/examine.py examples/hide_and_seek_quadrant.jsonnet examples/hide_and_seek_quadrant.npz 有辦案法跑出可視化MuJoCo引擎嗎? 有沒有報什麼錯誤?

有一些问题: File "/home/ddr/Downloads/multi-agent-emergence-environments/mujoco-worldgen/mujoco_worldgen/util/envs/flexible_load.py", line 22, in get_function result = getattr(import(module_path, fromlist=(function_name,)), function_name) ModuleNotFoundError: No module named 'mae_envs'

alan303138 commented 2 years ago

How can i test it? Missing file test_policy_ppo.py

兄弟,我也去跑了你的代码,救我!!!跑完那个ma_run.py后也没可视化界面啊,那还需要改什么地方,我是在虚拟机上面跑的

我是win10跑的,所以有些依賴項會搬來搬去,其他檔案都可以照原作者,我只多加了\spinup\utils\test_policy_PPO.py這個檔案,還有修改spinup.utils.logx.restore_tf_graph 其他應該都一樣。 你這邊是訓練完要test了嗎? openAi<<這個repo執行 python bin/examine.py examples/hide_and_seek_quadrant.jsonnet examples/hide_and_seek_quadrant.npz 有辦案法跑出可視化MuJoCo引擎嗎? 有沒有報什麼錯誤?

有一些问题: File "/home/ddr/Downloads/multi-agent-emergence-environments/mujoco-worldgen/mujoco_worldgen/util/envs/flexible_load.py", line 22, in get_function result = getattr(import(module_path, fromlist=(function_name,)), function_name) ModuleNotFoundError: No module named 'mae_envs'

multi-agent-emergence-environments mujoco-worldgen 這兩個是主要的,先要可以跑multi-agent-emergence-environments,再把一些裡面我改的一些覆蓋檔案蓋過去就行了 在目錄下pip install -e 可以把mea_env、ma_policy裝到pip裡面

SpinLeo commented 2 years ago

How can i test it? Missing file test_policy_ppo.py

兄弟,我也去跑了你的代码,救我!!!跑完那个ma_run.py后也没可视化界面啊,那还需要改什么地方,我是在虚拟机上面跑的

我是win10跑的,所以有些依賴項會搬來搬去,其他檔案都可以照原作者,我只多加了\spinup\utils\test_policy_PPO.py這個檔案,還有修改spinup.utils.logx.restore_tf_graph 其他應該都一樣。 你這邊是訓練完要test了嗎? openAi<<這個repo執行 python bin/examine.py examples/hide_and_seek_quadrant.jsonnet examples/hide_and_seek_quadrant.npz 有辦案法跑出可視化MuJoCo引擎嗎? 有沒有報什麼錯誤?

有一些问题: File "/home/ddr/Downloads/multi-agent-emergence-environments/mujoco-worldgen/mujoco_worldgen/util/envs/flexible_load.py", line 22, in get_function result = getattr(import(module_path, fromlist=(function_name,)), function_name) ModuleNotFoundError: No module named 'mae_envs'

multi-agent-emergence-environments mujoco-worldgen 這兩個是主要的,先要可以跑multi-agent-emergence-environments,再把一些裡面我改的一些覆蓋檔案蓋過去就行了 在目錄下pip install -e 可以把mea_env、ma_policy裝到pip裡面

感谢老哥,成功跑通了,呜呜呜,环境安装了一个周终于能用了,感激!