Marco Cerliani questions

Marco Cerliani answers

• Combining CNN and bidirectional LSTM
• Keras multioutput custom loss with intermediate layers output
• How to prepare the inputs in Keras implementation of Wavenet for time-series prediction
• Calculating Fscore for each epoch using keras (not batch-wise)
• InvalidArgumentError: indices[24,0] = 335 is not in [0, 304) [[{{node user-embedding-mlp_1/GatherV2}}]]
• LightGBM early stopping with custom eval function and built-in loss function
• Saving normalization values in Keras model
• tensorflow TypeError: cannot unpack non-iterable float object
• Creating a multi-channel network: 'Concatenate' object has no attribute 'shape'
• Unable to transform string column to categorical matrix using Keras and Sklearn