torchmetrics

torchmetrics copied to clipboard

Accuracy outputs conflicting results in a binary classification setting

3

## 🐛 Bug `Accuracy` (functional implementation; most likely the module as well) gives conflicting results. `StatScores` (functional) does give expected results when I follow the docs. As so many input...

ThScheeve

`BinnedPrecisionRecall` not working for multi-dimensional arrays

4

## 🐛 Bug Binned Precision Recall Curve not working expected with multi-class and multi-dimensional input/target ### To Reproduce Steps to reproduce the behavior: 1. Go to 'torchmetrics/classification/binned_precision_recall.py' 2. See line...

omerferhatt

Stat_Scores does not behave as expected

3

## 🐛 Bug torchmetrics.functional.stat_scores gives out wrong values ### To Reproduce Steps to reproduce the behavior: ```py preds = torch.tensor([1, 0, 0, 1]) target = torch.tensor([1, 1, 0, 0]) stat_scores(preds,target)...

durvesh8

Classification metrics consistency

4

## 🚀 Feature First of all, thank you very much for this awesome project. It helps a lot evaluating deep learning models on standard and out-of-the-box metrics for several application...

mayeroa

`ignore_index` for F1 doesn't behave as expected.

6

## 🐛 Bug F1 doesn't ignore indices properly. ### To Reproduce Run the following code. ```python import torch from torchmetrics import F1 f1 = F1(ignore_index=0) f1(torch.tensor([1, 1, 1, 1, 2,...

yukw777

## 🚀 Feature Hi in the sklearn API there is a `multi_class` parameters which can be set to ‘ovr or ‘ovo'. But it seems that this is not yet implemented...

chris-tkinter

FBeta binary average

8

## 🚀 Feature Have fbeta binary metric following sklearn ### Motivation This is used and is the default in sklearn. It can be misleading if people use it without thinking....

zippeurfou

Behaviour with average=`sample`, mdmc_average=`samplewise` seems confusing.

7

## 🐛 Bug Computing F1 with `average="samples"` and `mdmc_average=samplewise` seems to take into consideration all samples, including those that are ignored. Not sure if this is a bug or I...

AndreiSiliUiPath

Support float targets for classification metrics

6

## 🚀 Feature Support float targets (possibly softened one-hot encodings) for classification metrics. ### Motivation Techniques such as mixup or label smoothing use one-hot encoded targets that are then converted...

ethanwharris

Make Dice a ModuleMetric similar to IoU

5

## 🚀 Feature Make Dice a ModuleMetric similar to IoU ### Motivation Currently, the usability of Dice (functional metric) is no on par with that of (modular metric) IoU. ###...

seannz