torch_enhance.metrics¶

torch_enhance.metrics.mae(y_pred: torch.Tensor, y_true: torch.Tensor) → torch.Tensor[source]¶

Mean absolute error (MAE) metric

Parameters:	y_pred (torch.Tensor) – Super-Resolved image tensor y_true (torch.Tensor) – High Resolution image tensor
Returns:	Mean absolute error between y_true and y_pred
Return type:	torch.Tensor

torch_enhance.metrics.mse(y_pred: torch.Tensor, y_true: torch.Tensor) → torch.Tensor[source]¶

Mean squared error (MSE) metric

Parameters:	y_pred (torch.Tensor) – Super-Resolved image tensor y_true (torch.Tensor) – High Resolution image tensor
Returns:	Mean squared error between y_true and y_pred
Return type:	torch.Tensor

torch_enhance.metrics.psnr(y_pred: torch.Tensor, y_true: torch.Tensor) → torch.Tensor[source]¶

Peak-signal-noise ratio (PSNR) metric

Parameters:	y_pred (torch.Tensor) – Super-Resolved image tensor y_true (torch.Tensor) – High Resolution image tensor
Returns:	Peak-signal-noise-ratio between y_true and y_pred
Return type:	torch.Tensor