:jg"fddlmZddlZddlmZddlmZddlmZddlm Z ddl m Z  dde d e d eed d ed deedddf dZ ddedede d e d ed deeeeff dZ ddeded eed deeddeedef dZ ddedede d e d eed d ed deeddefdZy))OptionalN)Tensor)Literal)_segmentation_inputs_format)rank_zero_warn) _safe_divide num_classesinclude_backgroundaveragemicromacroweightednone input_formatone-hotindexmixedaggregation_level samplewiseglobalreturnc t|tr|dkrtd|dt|tstd|dgd}|||vrtd|d|d|d vrtd |d|d vrtd |y) z%Validate the arguments of the metric.rzDExpected argument `num_classes` must be a positive integer, but got .zBExpected argument `include_background` must be a boolean, but got r Nz)Expected argument `average` to be one of z or None, but got rzSExpected argument `input_format` to be one of 'one-hot', 'index', 'mixed', but got rzSExpected argument `aggregation_level` to be one of `samplewise`, `global`, but got ) isinstanceint ValueErrorbool)r r r rrallowed_averages ~/home/mcse/projects/srt_converter/srt-converter-venv/lib/python3.12/site-packages/torchmetrics/functional/segmentation/dice.py_dice_score_validate_argsr#s k3 ';!+;_`k_llmnoo ($ /]^p]qqrstt>> from torch import randint >>> from torchmetrics.functional.segmentation import dice_score >>> _ = torch.manual_seed(42) >>> preds = randint(0, 2, (4, 5, 16, 16)) # 4 samples, 5 classes, 16x16 prediction >>> target = randint(0, 2, (4, 5, 16, 16)) # 4 samples, 5 classes, 16x16 target >>> # dice score micro averaged over all classes >>> dice_score(preds, target, num_classes=5, average="micro") tensor([0.4842, 0.4968, 0.5053, 0.4902]) >>> # dice score per sample and class >>> dice_score(preds, target, num_classes=5, average="none") tensor([[0.4724, 0.5185, 0.4710, 0.5062, 0.4500], [0.4571, 0.4980, 0.5191, 0.4380, 0.5649], [0.5428, 0.4904, 0.5358, 0.4830, 0.4724], [0.4715, 0.4925, 0.4797, 0.5267, 0.4788]]) >>> # global dice score over all samples with macro averaging >>> dice_score(preds, target, num_classes=5, average="macro", aggregation_level="global") tensor([0.4942]) Example (with index tensors): >>> from torch import randint >>> from torchmetrics.functional.segmentation import dice_score >>> _ = torch.manual_seed(42) >>> preds = randint(0, 5, (4, 16, 16)) # 4 samples, 5 classes, 16x16 prediction >>> target = randint(0, 5, (4, 16, 16)) # 4 samples, 5 classes, 16x16 target >>> # dice score micro averaged over all classes >>> dice_score(preds, target, num_classes=5, average="micro", input_format="index") tensor([0.2031, 0.1914, 0.2266, 0.1641]) >>> # dice score per sample and class >>> dice_score(preds, target, num_classes=5, average="none", input_format="index") tensor([[0.1731, 0.1667, 0.2400, 0.2424, 0.1947], [0.2245, 0.2247, 0.2321, 0.1132, 0.1682], [0.2500, 0.2476, 0.1887, 0.1818, 0.2718], [0.1308, 0.1800, 0.1980, 0.1607, 0.1522]]) >>> # global dice score over all samples with macro averaging >>> dice_score(preds, target, num_classes=5, average="macro", aggregation_level="global", input_format="index") tensor([0.1965]) r zdice_score metric currently defaults to `average=micro`, but will change to`average=macro` in the v1.9 release. If you've explicitly set this parameter, you can ignore this warning.)rr6)r UserWarningr#r7rF) r%r&r r r rrr4r5r6s r" dice_scorerIisdF' U   k+=w Vgh&8 Ugiu&v#I{G y+wRcmt uur$)r rr)r)r rN)Trrr)typingrr.rtyping_extensionsr*torchmetrics.functional.segmentation.utilsrtorchmetrics.utilitiesrtorchmetrics.utilities.computerrr r#tupler7rFrIr$r"rQs %R17 HO9BCO    gBC D 56   (> ?@    <:C + + ++ + 56 +  666 !" +0HOCO $ AAAgBC DA (> ?@ A f  A  AL $GN9BCOLv Lv LvLv Lv gBC D Lv 56 Lv (> ?@Lv Lvr$