Issue #348: len is not well defined for symbolic Tensors

[rckitson]

This only addresses the first half of Issue #348 , but was able to run examples of DQN and DDPG with:

• tensorflow 1.15.0
• keras 2.3.1

[msat59]

This only addresses the first half of Issue #348 , but was able to run examples of DQN and DDPG with:

• tensorflow 1.15.0
• keras 2.3.1

I faced the same error with tensorflow 1.15.0. I fixed manually the dqn.py line 108:

original code: if hasattr(model.output, '__len__') and len(model.output) > 1:

changed to 108 if hasattr(model.output, '__len__') and model.output.shape[0] > 1:

[dave7895]

With this fix I get the error TypeError: '>' not supported between instances of 'NoneType' and 'int'.