o .wid*@sddlmZddlmZmZmZddlZddlmZddlm Z ddl m Z m Z ddl mZddlmZdd lmZmZesAd gZ dd ed ededededede ddefddZGdddeZdS))Sequence)AnyOptionalUnionN)Tensor)Literal)_mean_squared_error_update+_normalized_root_mean_squared_error_compute)Metric)_MATPLOTLIB_AVAILABLE)_AX_TYPE_PLOT_OUT_TYPE#NormalizedRootMeanSquaredError.plotmeanmin_valmax_valmean_valvar_valtarget_squaredtotal normalizationrrangestdl2returncCst|dkr*|dkr|dS|dkr|d|dS|dkr"|dS|dkr*|dS|d|d|d|d|d|df\}}} } } } tdt|D]r} || || || || || || f\}}}}}}| |}| | |||}| d|| | }| || ||||d7} |d|||}|||||||d7}| |}t||}t||}| |}qK|dkr|S|dkr||S|dkr||S|S)z_In the case of multiple devices we need to aggregate the statistics from the different devices.rrrrr)lenrtorchminmaxsqrt)rrrrrrr min_val_1 max_val_1 mean_val_1 var_val_1target_squared_1total_1i min_val_2 max_val_2 mean_val_2 var_val_2target_squared_2total_2r_tempvarr2Z/home/ubuntu/sommelier/.venv/lib/python3.10/site-packages/torchmetrics/regression/nrmse.py_final_aggregation!sR        r4c seZdZUdZdZeed<dZeed<dZeed<dZ e ed<e ed <e ed <e ed <e ed <e ed <e ed<e ed<  d$de dde deddffdd Zde de ddfddZde fddZ d%d eee ee fd!eedefd"d#ZZS)&NormalizedRootMeanSquaredErroraCalculates the `Normalized Root Mean Squared Error`_ (NRMSE) also know as scatter index. The metric is defined as: .. math:: \text{NRMSE} = \frac{\text{RMSE}}{\text{denom}} where RMSE is the root mean squared error and `denom` is the normalization factor. The normalization factor can be either be the mean, range, standard deviation or L2 norm of the target, which can be set using the `normalization` argument. As input to ``forward`` and ``update`` the metric accepts the following input: - ``preds`` (:class:`~torch.Tensor`): Predictions from model - ``target`` (:class:`~torch.Tensor`): Ground truth values As output of ``forward`` and ``compute`` the metric returns the following output: - ``nrmse`` (:class:`~torch.Tensor`): A tensor with the mean squared error Args: normalization: type of normalization to be applied. Choose from "mean", "range", "std", "l2" which corresponds to normalizing the RMSE by the mean of the target, the range of the target, the standard deviation of the target or the L2 norm of the target. num_outputs: Number of outputs in multioutput setting kwargs: Additional keyword arguments, see :ref:`Metric kwargs` for more info. Example:: Single output normalized root mean squared error computation: >>> import torch >>> from torchmetrics import NormalizedRootMeanSquaredError >>> target = torch.tensor([2.5, 5.0, 4.0, 8.0]) >>> preds = torch.tensor([3.0, 5.0, 2.5, 7.0]) >>> nrmse = NormalizedRootMeanSquaredError(normalization="mean") >>> nrmse(preds, target) tensor(0.1919) >>> nrmse = NormalizedRootMeanSquaredError(normalization="range") >>> nrmse(preds, target) tensor(0.1701) Example:: Multioutput normalized root mean squared error computation: >>> import torch >>> from torchmetrics import NormalizedRootMeanSquaredError >>> preds = torch.tensor([[0., 1], [2, 3], [4, 5], [6, 7]]) >>> target = torch.tensor([[0., 1], [3, 3], [4, 5], [8, 9]]) >>> nrmse = NormalizedRootMeanSquaredError(num_outputs=2) >>> nrmse(preds, target) tensor([0.2981, 0.2222]) Tis_differentiableFhigher_is_betterfull_state_updategplot_lower_boundsum_squared_errorrrrrrrrrrr num_outputskwargsrNc stjdi||dvrtd|||_t|tr |dks'td|||_|jdt |dd|jdt |dd|jd t d t |jdd|jd t d  t |jdd|jd t |jdd|jd t |jdd|jdt |jdddS)NrzRArgument `normalization` should be either 'mean', 'range', 'std' or 'l2', but got rz6Expected num_outputs to be a positive integer but got r:sum)defaultdist_reduce_fxrrInfrrrrr2) super__init__ ValueErrorr isinstanceintr; add_staterzerosfloatones)selfrr;r< __class__r2r3rBs  "z'NormalizedRootMeanSquaredError.__init__predstargetcCst|||j\}}|j|7_|jdkr|dn|}t|jddj|j|_t |j ddj|j |_ |j |dj dd7_ |j|j|j dd|j|}|j|7_||||jj dd}||_|j|7_dS)zjUpdate state with predictions and targets. See `mean_squared_error_update` for details. rrdimrN)rr;r:viewrminimumr valuesrmaximumr!rrr=rrr)rJrMrNr:num_obsnew_meannew_varr2r2r3updates"z%NormalizedRootMeanSquaredError.updatec Cs|jdkr |jdks|jdkr3|jjdkr3t|j|j|j|j|j|j |j d}|j j dd}n-|j dkr<|j}n!|j dkrH|j|j}n|j dkrWt |j|j }nt |j}|j }t|j||S)z[Computes NRMSE over state. See `mean_squared_error_compute` for details. r)rrrrrrrrrPrrr)r;rnumelndimr4rrrrrrsqueezer=rr"r r:)rJdenomrr2r2r3computes(.    z&NormalizedRootMeanSquaredError.computevalaxcCs |||S)aGPlot a single or multiple values from the metric. Args: val: Either a single result from calling `metric.forward` or `metric.compute` or a list of these results. If no value is provided, will automatically call `metric.compute` and plot that result. ax: An matplotlib axis object. If provided will add plot to that axis Returns: Figure and Axes object Raises: ModuleNotFoundError: If `matplotlib` is not installed .. plot:: :scale: 75 >>> from torch import randn >>> # Example plotting a single value >>> from torchmetrics.regression import NormalizedRootMeanSquaredError >>> metric = NormalizedRootMeanSquaredError() >>> metric.update(randn(10,), randn(10,)) >>> fig_, ax_ = metric.plot() .. plot:: :scale: 75 >>> from torch import randn >>> # Example plotting multiple values >>> from torchmetrics.regression import NormalizedRootMeanSquaredError >>> metric = NormalizedRootMeanSquaredError() >>> values = [] >>> for _ in range(10): ... values.append(metric(randn(10,), randn(10,))) >>> fig, ax = metric.plot(values) )_plot)rJr_r`r2r2r3plots (r)rr)NN)__name__ __module__ __qualname____doc__r6bool__annotations__r7r8r9rHrrrErrBrYr^rrrr r rb __classcell__r2r2rKr3r5_sD  6   r5)r)collections.abcrtypingrrrrrtyping_extensionsr(torchmetrics.functional.regression.nrmserr torchmetrics.metricr torchmetrics.utilities.importsr torchmetrics.utilities.plotr r __doctest_skip__r4r5r2r2r2r3s<       >