search

farizrahman4u / qlearning4k

Q-learning for Keras
MIT License
385 stars 84 forks source link

The layer "sequential_1 has multiple inbound nodes, with different input shapes. Hence the notion of "input shape" is ill-defined for the layer. Use `get_input_shape_at(node_index)` instead. #22

Open naouss80 opened 7 years ago

naouss80 commented 7 years ago

Hi I seem to be running $ python test_catch.py and here is what I seem to get

Epoch 1000/1000 | Loss 0.0104 | Epsilon 0.10 | Win count 793 Traceback (most recent call last): File "test_catch.py", line 21, in agent.play(catch) File "C:\Users\user\AppData\Local\Continuum\Anaconda3\lib\site-packages\qlearning4k-0.0.1-py3.6.egg\qlearning4k\agent.py", line 105, in play File "C:\Users\user\AppData\Local\Continuum\Anaconda3\lib\site-packages\qlearning4k-0.0.1-py3.6.egg\qlearning4k\agent.py", line 38, in check_game_compatibility File "C:\Users\user\AppData\Local\Continuum\Anaconda3\lib\site-packages\keras\engine\topology.py", line 1028, in input_shape ' has multiple inbound nodes, ' AttributeError: The layer "sequential_1 has multiple inbound nodes, with different input shapes. Hence the notion of "input shape" is ill-defined for the layer. Use get_input_shape_at(node_index) instead.

jasperr commented 7 years ago

I have seen the same error

gaara100 commented 7 years ago

I'm having the same issue.

jkempson commented 7 years ago

Same. Looks like something changed in Keras?

sdgroeve commented 7 years ago

same here....

vftens commented 7 years ago

The same error. Use Keras==1.2.2 tensorflow==1.3.0 tensorflow-gpu==1.3.0 to avoid such error