:js#ddlmZddlmZddlmZmZmZmZddl Z ddl m Z ddl m Z ddl mZddlmZdd lmZdd lmZmZdd lmZesd gZ dd edede jfdZGddeZy))Sequence)deepcopy)AnyOptionalUnioncastN)apply_to_collection)Tensor) ModuleList)Metric)_MATPLOTLIB_AVAILABLE)_AX_TYPE_PLOT_OUT_TYPE) WrapperMetriczBootStrapper.plotsizesampling_strategyreturncL|dk(retjjd}|j|f}tj|j |j dS|dk(r+tjtj||dStd) zResample a tensor along its first dimension with replacement. Args: size: number of samples sampling_strategy: the strategy to use for sampling, either ``'poisson'`` or ``'multinomial'`` Returns: resampled tensor poissonrdim multinomialT) num_samples replacementzUnknown sampling strategy) torch distributionsPoissonsamplearangerepeat_interleavelongrones ValueError)rrpns x/home/mcse/projects/srt_converter/srt-converter-venv/lib/python3.12/site-packages/torchmetrics/wrappers/bootstrapping.py_bootstrap_samplerr( sI%    ' ' * HHdW ||D!33AFFH!3DDM)  D!1tQUVV 0 11ceZdZUdZdZeeed< ddede deded ee e e fd ed e d ed dffd Zded ed dfdZd ee e ffdZded ed effd Zdfd Z ddee e ee fdeed efdZxZS) BootStrapperaUsing `Turn a Metric into a Bootstrapped`_. That can automate the process of getting confidence intervals for metric values. This wrapper class basically keeps multiple copies of the same base metric in memory and whenever ``update`` or ``forward`` is called, all input tensors are resampled (with replacement) along the first dimension. Args: base_metric: base metric class to wrap num_bootstraps: number of copies to make of the base metric for bootstrapping mean: if ``True`` return the mean of the bootstraps std: if ``True`` return the standard deviation of the bootstraps quantile: if given, returns the quantile of the bootstraps. Can only be used with pytorch version 1.6 or higher raw: if ``True``, return all bootstrapped values sampling_strategy: Determines how to produce bootstrapped samplings. Either ``'poisson'`` or ``multinomial``. If ``'possion'`` is chosen, the number of times each sample will be included in the bootstrap will be given by :math:`n\sim Poisson(\lambda=1)`, which approximates the true bootstrap distribution when the number of samples is large. If ``'multinomial'`` is chosen, we will apply true bootstrapping at the batch level to approximate bootstrapping over the hole dataset. kwargs: Additional keyword arguments, see :ref:`Metric kwargs` for more info. Example:: >>> from pprint import pprint >>> from torch import randint >>> from torchmetrics.wrappers import BootStrapper >>> from torchmetrics.classification import MulticlassAccuracy >>> base_metric = MulticlassAccuracy(num_classes=5, average='micro') >>> bootstrap = BootStrapper(base_metric, num_bootstraps=20) >>> bootstrap.update(randint(5, (20,)), randint(5, (20,))) >>> output = bootstrap.compute() >>> pprint(output) {'mean': tensor(0.2089), 'std': tensor(0.0772)} Tfull_state_updateN base_metricnum_bootstrapsmeanstdquantilerawrkwargsrc Lt |di|t|tst d|t t |D cgc] } t|c} |_||_ ||_ ||_ ||_ ||_ d} || vrt d| d|||_ycc} w)NzKExpected base metric to be an instance of torchmetrics.Metric but received )rrz5Expected argument ``sampling_strategy`` to be one of z but received )super__init__ isinstancer r$r rangermetricsr.r/r0r1r2r) selfr-r.r/r0r1r2rr3_allowed_sampling __class__s r'r7zBootStrapper.__init__]s "6"+v.]^i]jk "%BW"XQ8K#8"XY ,   5 $4 4GHXGY !2 35 "3#YsB!argsct|tjt}t|tjt}t|dkDr|d}nC+DDDZDZ[^^_c_j_jkJ!Q&*4u?Q?QWX`jkH,VU\\5CUCU[\dnoJ $DLL  $ $h =* =  >r)c tj|jDcgc] }tt|j "c}d}i}|j r|j d|d<|jr|jd|d<|j#tj||j|d<|jr||d<|Scc}w)zCompute the bootstrapped metric values. Always returns a dict of tensors, which can contain the following keys: ``mean``, ``std``, ``quantile`` and ``raw`` depending on how the class was initialized. rrr/r0r1r2) rstackr:rr computer/r0r1r2)r;m computed_vals output_dicts r'rSzBootStrapper.computes  $U1T&!_%<%<%>$U[\]  99"/"4"4"4";K  88!.!2!2q!2!9K  == $&+nn]DMM&RK # 88!.K %Vs%Cc*tt| |i|S)z9Use the original forward method of the base metric class.)r6rforward)r;r?r3r>s r'rXzBootStrapper.forwards]D14B6BBr)c|jD]"}tt|}|j$t| y)z#Reset the state of the base metric.N)r:rr resetr6)r;rTr>s r'rZzBootStrapper.resets5 AVQA GGI   r)valaxc&|j||S)aPlot 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 >>> # Example plotting a single value >>> import torch >>> from torchmetrics.wrappers import BootStrapper >>> from torchmetrics.regression import MeanSquaredError >>> metric = BootStrapper(MeanSquaredError(), num_bootstraps=20) >>> metric.update(torch.randn(100,), torch.randn(100,)) >>> fig_, ax_ = metric.plot() .. plot:: :scale: 75 >>> # Example plotting multiple values >>> import torch >>> from torchmetrics.wrappers import BootStrapper >>> from torchmetrics.regression import MeanSquaredError >>> metric = BootStrapper(MeanSquaredError(), num_bootstraps=20) >>> values = [ ] >>> for _ in range(3): ... values.append(metric(torch.randn(100,), torch.randn(100,))) >>> fig_, ax_ = metric.plot(values) )_plot)r;r[r\s r'plotzBootStrapper.plotsTzz#r""r)) TTNFr)rN)NN)__name__ __module__ __qualname____doc__r,rbool__annotations__r intrfloatr strrr7rJdictrSrXrZrrrr_ __classcell__)r>s@r'r+r+7s*!F)-x~, !37!*333 3  3 5/0 3333 3B>C>3>4>.c6k*&CSCCCCC_c*#E&(6*:":;<*#IQRZI[*# *#r)r+)r)collections.abcrcopyrtypingrrrrrlightning_utilitiesr r torch.nnr torchmetrics.metricr torchmetrics.utilities.importsr torchmetrics.utilities.plotrrtorchmetrics.wrappers.abstractr__doctest_skip__rgrir(r+r5r)r'rvsm%-- 3&@@8+, '2 22 \\2.f#=f#r)