search

eugenevinitsky / sequential_social_dilemma_games

Repo for reproduction of sequential social dilemmas
MIT License
380 stars 134 forks source link

Visualization using A3C agents in harvest env #210

Open farinazAH opened 1 year ago

farinazAH commented 1 year ago

Hello everybody. After learning 5 agents in harvest environment using A3C algorithm and for example baseline method. How can I save the movie of a test run using saved model and checkpoints? I tried to use render_checkpoint.py file in visualization directory. This file can create movies using ppo algorithm without any problem but it has errors using A3C algorithm. Regards

farinazAH commented 9 months ago

Hello,I uased the versions in the requirement file introduced in  internetcoffeephone fork.

On یکشنبه, نوامبر ۵, ۲۰۲۳ at ۳:۲۵, Alexis @.***> wrote:

Hello Farinaz, just wondering if you have ever had any dependency issues when running the setup? The main issue we are having is that ray 0.8.5 is no longer available (and "from ray.rllib.agents.registry import get_agent_class" does not work in newer versions of ray like 2.8). We tried to get ray 0.8.5 from the internet but then it uses a lot of older versions of packages that are no longer available, like tensorflow 1.13 that uses tf.config (also does not exist in tensorflow 2.)

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you are subscribed to this thread.Message ID: @.***>

tungts1101 commented 6 months ago

Hi @farinazAH, I am trying to follow the README to set up the baseline model training in run_scipts/train.py but I got the error ModuleNotFoundError. How did you do to work around this issue?