NEDSMasker# class pretrain.models.NEDSMasker(mask_ratio, modalities=['spikes', 'choice', 'reward', 'stimulus_contrast', 'whisker_motion_energy', 'wheel_speed', 'right_paw_speed', 'left_paw_speed', 'licking_rate'], mask_types=None)[source]# Bases: torch.nn.modules.module.Module forward(input, generator=None)[source]# generator: draw from it instead of the global RNG, to hold the masks fixed across epochs (the validation pass seeds one per epoch).
