Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Freezing terminal #1

Open
mimvdb opened this issue Apr 24, 2022 · 1 comment
Open

Freezing terminal #1

mimvdb opened this issue Apr 24, 2022 · 1 comment

Comments

@mimvdb
Copy link

mimvdb commented Apr 24, 2022

Hi, I'm running into an issue trying to train a model using the instructions in the repository. Once the weights and biases (wandb) choice comes up, the process seems to not do anything, not even responding to SIGINT. It also does not seem to be training the model, since the system resources are barely used.

See the attached log freeze.log.

I had hoped to use this model as reference point to measure the performance of specific models for Duckietown. If there is a pre-trained model available that would also be great.

@kaland313
Copy link
Owner

kaland313 commented Apr 25, 2022

Hi, as a quick and dirty fix, I suggest that you comment/delete the WeightsAndBiasesLogger from the list of loggers in experiments/train-rllib.py:

--- a/experiments/train-rllib.py
+++ b/experiments/train-rllib.py
@@ -99,6 +99,6 @@ tune.run(PPOTrainer,
          checkpoint_score_attr="episode_reward_mean",
          checkpoint_freq=1,
          restore=checkpoint_path,
-         loggers=[CSVLogger, TBXLogger, TensorboardImageLogger, WeightsAndBiasesLogger]
+         loggers=[CSVLogger, TBXLogger, TensorboardImageLogger]
          )

I'm not sure why can't the process continue if you don't use wandb, I'll have a look at it later.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants