Trouble replicating results on agent_visualizer.ipynb on colab

[tpprymjmdr]

Hi,

I am trying to replicate the results by running the colab notebook

https://colab.research.google.com/github/google/dopamine/blob/master/dopamine/colab/agent_visualizer.ipynb

I have modified the following in order to run the code:

1. uninstalled tensorflow 2.x and installed tensorflow 1.15
2. added the argument use_legacy_checkpoint=False in the function example_viz_lib.run()

The graph for reward is generated correctly, but not the graph for the reward distribution:

The log after running example_viz_lib.run() is here.

Could you please take a look and let me know what the issue is?

[GoingMyWay]

Many ipynb files are old I think.

[psc-g]

apologies for this. we realize some of the colab files have gone stale, but are working on migrating to TF2 (and will fix up the colab notebooks as part of this).

@tpprymjmdr: i just tried running that notebook as is (but setting it to install tensorflow==1.15) and it seemed to work for me. would you mind re-running again? [image: image.png]

On Thu, Jul 23, 2020 at 7:55 AM Alexander [email protected] wrote:

Many ipynb files are old I think.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/google/dopamine/issues/143#issuecomment-662964233, or unsubscribe https://github.com/notifications/unsubscribe-auth/AE3CCMMSKYNZAS5V4BHADFLR5AQKFANCNFSM4OIURF4Q .

[psc-g]

we just pushed some fixed and it seems to work from my tests. can you try it again? you may have to reset your runtime to clear its cache.