behzadanksu / rl-attack

Adversarial Example Attacks on Policy Learners
MIT License
39 stars 7 forks source link

NotImplementedError: #1

Closed Semizenbora closed 5 years ago

behzadanksu commented 5 years ago

We noticed this issue just a few days ago. It is probably due to recent architectural updates in both OpenAI Baselines and Cleverhans. We are looking into this and will post an update soon. In the meantime, you can probably run your experiments by reverting to older releases of TF, Baselines, and Cleverhans.

behzadanksu commented 5 years ago

@Semizenbora You may. Alternatively, you can use the new release of RL-Attack, which should be compatible with the latest versions of Cleverhans and OpenAI Gym.