dopamine
dopamine copied to clipboard
google
Dopamine is a research framework for fast prototyping of reinforcement learning algorithms.
It seems that the method *is_valid_transition* of OutOfGraphReplayBuffer is not checking if the stacked images are coming from another __truncated__ trajectory, in which case the index is invalid. It only...
Hi. The dependencies should be changed (particularly the `jax` version shouldn't be the latest one `0.4.13`). Related [issue ](https://github.com/deepmind/optax/issues/533) in another repo.  Tested in [colab](https://colab.research.google.com/github/google/dopamine/blob/master/dopamine/colab/jax_agent_visualizer.ipynb)
Cartpole example https://colab.research.google.com/github/google/dopamine/blob/master/dopamine/colab/cartpole.ipynb Is pretty much out of date... Using it's code on C51, it throws this error on my local machine: ``` ~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~ RuntimeError: in user code: File "/home/arre/Code/Syncd/exchange_downloader/snek/.venv/lib/python3.11/site-packages/dopamine/discrete_domains/gym_lib.py",...
Hi, i tried to study the colab sample linked from: https://github.com/google/dopamine/tree/master/dopamine/jax This colab sample: https://colab.research.google.com/github/google/dopamine/blob/master/dopamine/colab/jax_agent_visualizer.ipynb First issue is easy to fix, you need additional: ``` !pip install gym[accept-rom-license] !pip install...
I've been following the ReDo paper ("The Dormant Neuron Phenomenon in Deep Reinforcement Learning", https://arxiv.org/pdf/2302.12902.pdf) which describes the procedure to determine the dormant neurons in terms of their normalized scores....
First, thanks for this awesome codebase, it helped me a lot :) I have three questions 1. Based on #147 and the [white paper](https://arxiv.org/abs/1812.06110), the results in the [baseline folder](https://github.com/google/dopamine/tree/master/baselines)...
in gin.configuration "text_summary" is not found I'd ask what the version of tensorflow
In the TF version of DQN, the value of `epsilon_decay_period` is set to 1M steps (see [here](https://github.com/google/dopamine/blob/a2753dae222c75ae991758d4110a84bc01c3215f/dopamine/agents/dqn/configs/dqn_nature.gin#L15)), and for Rainbow, the value is set to 250k steps (see [here](https://github.com/google/dopamine/blob/a2753dae222c75ae991758d4110a84bc01c3215f/dopamine/agents/rainbow/configs/rainbow.gin#L18)). However,...
