Model: lstm Number of layers: 1 Learning rate: 0.0004 Batch size: 128 Training data shape: x = torch.Size([6000000, 10, 7]) y = torch.Size([6000000]) Validation data shape: x = torch.Size([600000, 10, 7]) y = torch.Size([600000]) Test data shape: x = torch.Size([600000, 10, 7]) y = torch.Size([600000]) Trainable parameters: 10993 Optimizer: Adam ( Parameter Group 0 amsgrad: False betas: (0.9, 0.999) eps: 1e-08 lr: 0.0004 weight_decay: 0.0 )