Model: lstm Number of layers: 3 Learning rate: 0.0001 Batch size: 2048 Training data shape: x = torch.Size([12000000, 10, 7]) y = torch.Size([12000000]) Validation data shape: x = torch.Size([1200000, 10, 7]) y = torch.Size([1200000]) Test data shape: x = torch.Size([1200000, 10, 7]) y = torch.Size([1200000]) Trainable parameters: 48625 Optimizer: Adam ( Parameter Group 0 amsgrad: False betas: (0.9, 0.999) eps: 1e-08 lr: 0.0001 weight_decay: 0.0 )