torchmetrics

torchmetrics copied to clipboard

[Refactor] Classification 6/n

1

## What does this PR do? Continues work on classification refactor https://github.com/Lightning-AI/metrics/issues/1001 Prior work: * https://github.com/Lightning-AI/metrics/pull/1159 * https://github.com/Lightning-AI/metrics/pull/1151 * https://github.com/Lightning-AI/metrics/pull/1145 * https://github.com/Lightning-AI/metrics/pull/1143 * https://github.com/Lightning-AI/metrics/pull/1054 This PR takes care of `precision_recall_curve`,...

SkafteNicki

Reduction on full image not required?

1

https://github.com/Lightning-AI/metrics/blob/ff61c482e5157b43e647565fa0020a4ead6e9d61/torchmetrics/functional/image/ssim.py#L191-L194 To get the full SSIM image to be returned I have to change to `reduction=None` Example: imports and dummy images ```py import torch from torchmetrics import StructuralSimilarityIndexMeasure as SSIM...

mattclifford1

Refactor/classification 8

1

## What does this PR do? Continues work on classification refactor https://github.com/Lightning-AI/metrics/issues/1001 Prior work: * https://github.com/Lightning-AI/metrics/pull/1167 * https://github.com/Lightning-AI/metrics/pull/1163 * https://github.com/Lightning-AI/metrics/pull/1159 * https://github.com/Lightning-AI/metrics/pull/1151 * https://github.com/Lightning-AI/metrics/pull/1145 * https://github.com/Lightning-AI/metrics/pull/1143 * https://github.com/Lightning-AI/metrics/pull/1054 This PR...

SkafteNicki

Removed micro average doc string from auroc

1

## What does this PR do? Removes docstring mistake. Average `micro` is not available for auroc. ## PR review Anyone in the community is free to review the PR once...

lukasfolle

Support for 2D tensors for pearson and spearman correlations

1

Why constraint the pearson correlation to a single dimension? With a 2D tensor, the pearson correlation can be computed simultaneously for every column. Simply need to replace the `.sum()` and...

DomInvivo

Add Area Under PR- curve metric

1

## 🚀 Feature There doesn't seem to be an easy way to calculate the AUPR using torchmetrics. If anybody has the recommended way, please comment below. ### Motivation The area...

rabinadk1

Wrong Calculation of Mean Average Precision

3

## 🐛 Bug Wrong calculation of mean Average Precision (mAP). I get ```py {'map': tensor(1.), 'map_50': tensor(1.), 'map_75': tensor(-1), 'map_small': tensor(1.), 'map_medium': tensor(-1.), 'map_large': tensor(-1.), 'mar_1': tensor(1.), 'mar_10': tensor(1.), 'mar_100':...

RostyslavUA

dist_sync_fn seems dysfunctional

1

## 🐛 Bug `dist_sync_fn` is a documented kwarg of class Metric and suggests to pass an alternative to `torch.distributed.all_gather`. I suppose, the intend is to allow use of distributed contexts...

yellowdolphin

Add Normalized Discounted Cumulative Gain (nDCG) Score for Classification

4

## 🚀 Feature Normalized Discounted Cumulative Gain (nDCG) score for classification. ### Motivation We are using torchmetrics nDCG score in multi-label classification. ([Reference](https://github.com/ASUS-AICS/LibMultiLabel/blob/master/libmultilabel/nn/metrics.py)) ```python3 from torchmetrics import RetrievalNormalizedDCG ``` At...

JamesLYC88

Refactor BERTScore tests

7

## What does this PR do? Fixes #1110 Following on this PR, I'd also suggest a refactor of `BERTScore` itself as its implementation is a bit obscure. ## Before submitting...

stancld