search

drawbridge / keras-mmoe

A TensorFlow Keras implementation of "Modeling Task Relationships in Multi-task Learning with Multi-gate Mixture-of-Experts" (KDD 2018)
MIT License
681 stars 217 forks source link

got an error when run the demo #1

Closed yogixRush closed 5 years ago

yogixRush commented 5 years ago

Training data shape = (10000, 100) Validation data shape = (1000, 100) Test data shape = (1000, 100) Traceback (most recent call last): File "synthetic_demo.py", line 152, in main() File "synthetic_demo.py", line 110, in main )(input_layer) File "/home/chenjw/anaconda3/lib/python3.6/site-packages/keras/engine/topology.py", line 603, in call output = self.call(inputs, **kwargs) File "/home/chenjw/Project/WTX/mmoe/keras-mmoe-master/mmoe.py", line 175, in call expert_outputs = K.bias_add(x=expert_outputs, bias=self.expert_bias) File "/home/chenjw/anaconda3/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py", line 3551, in bias_add if len(bias_shape) != 1 and len(bias_shape) != ndim(x) - 1: TypeError: unsupported operand type(s) for -: 'NoneType' and 'int'

alvin319 commented 5 years ago

Hi @yogixRush! Thanks for catching this issue, I've updated the requirements.txt and tested it myself. It seems to work fine. Let me know if this also works for you. I'll close this ticket for now and feel free to re-open it if you run into similar issues.