Question about keras implementation while dealing Y

kaelzhang / DA-RNN-in-Tensorflow-2-and-PyTorch

A Tensorflow 2 (Keras) implementation of DA-RNN (A Dual-Stage Attention-Based Recurrent Neural Network for Time Series Prediction, arXiv:1704.02971)

MIT License

29 stars 12 forks source link

Open juanguang opened 1 year ago

juanguang commented 1 year ago

Hi, I wonder did you still include y_T here? I dont see the code which abandan y_T should line 439 be: Y = inputs[:, T-1, -self.y_dim:] so that it matches the dimension you wrote in the comment? https://github.com/kaelzhang/DA-RNN-in-Tensorflow-2-and-PyTorch/blob/1a30c3dffd7e8556c775480faa500d84fed5334c/da_rnn/keras/model.py#L435

kaelzhang commented 1 year ago

Please refer to the implementation of pytorch. Actually I don't have time to figure out the issue for tf implementation, but pytorch one is ok.