Merge branch 'master' into ter

CHANGELOG.md

@@ -33,6 +33,9 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - Scalar metrics will now consistently have additional dimensions squeezed ([#622](https://github.com/PyTorchLightning/metrics/pull/622))
 
 
+- Metrics having third party dependencies removed from global import ([#463](https://github.com/PyTorchLightning/metrics/pull/463))
+
+
 - `BLEUScore` now expects untokenized input to stay consistent with all the other text metrics ([#640](https://github.com/PyTorchLightning/metrics/pull/640))
 
 

docs/source/references/functional.rst

@@ -14,7 +14,7 @@ Audio Metrics
 pesq [func]
 ~~~~~~~~~~~
 
-.. autofunction:: torchmetrics.functional.pesq
+.. autofunction:: torchmetrics.functional.audio.pesq.pesq
 
 
 pit [func]
@@ -55,7 +55,7 @@ snr [func]
 stoi [func]
 ~~~~~~~~~~~
 
-.. autofunction:: torchmetrics.functional.stoi
+.. autofunction:: torchmetrics.functional.audio.stoi.stoi
     :noindex:
 
 
@@ -433,7 +433,7 @@ Text
 bert_score [func]
 ~~~~~~~~~~~~~~~~~~
 
-.. autofunction:: torchmetrics.functional.bert_score
+.. autofunction:: torchmetrics.functional.text.bert.bert_score
 
 bleu_score [func]
 ~~~~~~~~~~~~~~~~~
@@ -462,7 +462,7 @@ match_error_rate [func]
 rouge_score [func]
 ~~~~~~~~~~~~~~~~~~
 
-.. autofunction:: torchmetrics.functional.rouge_score
+.. autofunction:: torchmetrics.functional.text.rouge.rouge_score
     :noindex:
 
 sacre_bleu_score [func]

docs/source/references/modules.rst

@@ -77,7 +77,7 @@ the metric will be computed over the ``time`` dimension.
 PESQ
 ~~~~
 
-.. autoclass:: torchmetrics.PESQ
+.. autoclass:: torchmetrics.audio.pesq.PESQ
 
 PIT
 ~~~
@@ -112,7 +112,7 @@ SNR
 STOI
 ~~~~
 
-.. autoclass:: torchmetrics.STOI
+.. autoclass:: torchmetrics.audio.stoi.STOI
     :noindex:
 
 
@@ -369,25 +369,25 @@ learning algorithms such as `Generative Adverserial Networks (GANs) <https://en.
 FID
 ~~~
 
-.. autoclass:: torchmetrics.FID
+.. autoclass:: torchmetrics.image.fid.FID
     :noindex:
 
 IS
 ~~
 
-.. autoclass:: torchmetrics.IS
+.. autoclass:: torchmetrics.image.inception.IS
     :noindex:
 
 KID
 ~~~
 
-.. autoclass:: torchmetrics.KID
+.. autoclass:: torchmetrics.image.kid.KID
     :noindex:
 
 LPIPS
 ~~~~~
 
-.. autoclass:: torchmetrics.LPIPS
+.. autoclass:: torchmetrics.image.lpip_similarity.LPIPS
     :noindex:
 
 PSNR
@@ -411,7 +411,7 @@ Object detection metrics can be used to evaluate the predicted detections with g
 MAP
 ~~~
 
-.. autoclass:: torchmetrics.MAP
+.. autoclass:: torchmetrics.detection.map.MAP
     :noindex:
 
 ******************
@@ -613,7 +613,7 @@ Text
 BERTScore
 ~~~~~~~~~~
 
-.. autoclass:: torchmetrics.BERTScore
+.. autoclass:: torchmetrics.text.bert.BERTScore
     :noindex:
 
 BLEUScore
@@ -643,7 +643,7 @@ MatchErrorRate
 ROUGEScore
 ~~~~~~~~~~
 
-.. autoclass:: torchmetrics.ROUGEScore
+.. autoclass:: torchmetrics.text.rouge.ROUGEScore
     :noindex:
 
 SacreBLEUScore

tests/audio/test_pesq.py

@@ -21,8 +21,8 @@
 
 from tests.helpers import seed_all
 from tests.helpers.testers import MetricTester
-from torchmetrics.audio import PESQ
-from torchmetrics.functional import pesq
+from torchmetrics.audio.pesq import PESQ
+from torchmetrics.functional.audio.pesq import pesq
 from torchmetrics.utilities.imports import _TORCH_GREATER_EQUAL_1_6
 
 seed_all(42)

tests/audio/test_stoi.py

@@ -21,8 +21,8 @@
 
 from tests.helpers import seed_all
 from tests.helpers.testers import MetricTester
-from torchmetrics.audio import STOI
-from torchmetrics.functional import stoi
+from torchmetrics.audio.stoi import STOI
+from torchmetrics.functional.audio.stoi import stoi
 from torchmetrics.utilities.imports import _TORCH_GREATER_EQUAL_1_6
 
 seed_all(42)

tests/image/test_fid.py

@@ -37,7 +37,7 @@ def generate_cov(n):
 
     scipy_res = scipy_sqrtm((cov1 @ cov2).numpy()).real
     tm_res = sqrtm(cov1 @ cov2)
-    assert torch.allclose(torch.tensor(scipy_res).float(), tm_res, atol=1e-3)
+    assert torch.allclose(torch.tensor(scipy_res).float().trace(), tm_res.trace())
 
 
 @pytest.mark.skipif(not _TORCH_FIDELITY_AVAILABLE, reason="test requires torch-fidelity")

tests/text/test_bertscore.py

@@ -7,8 +7,8 @@
 import torch.distributed as dist
 import torch.multiprocessing as mp
 
-from torchmetrics.functional import bert_score as metrics_bert_score
-from torchmetrics.text import BERTScore
+from torchmetrics.functional.text.bert import bert_score as metrics_bert_score
+from torchmetrics.text.bert import BERTScore
 from torchmetrics.utilities.imports import _BERTSCORE_AVAILABLE
 
 if _BERTSCORE_AVAILABLE:

tests/wrappers/test_minmax.py

@@ -65,7 +65,8 @@ class TestMinMaxWrapper(MetricTester):
 
     atol = 1e-6
 
-    @pytest.mark.parametrize("ddp", [True, False])
+    # TODO: fix ddp=True case, difference in how compare function works and wrapper metric
+    @pytest.mark.parametrize("ddp", [False])
     def test_minmax_wrapper(self, preds, target, base_metric, ddp):
         self.run_class_metric_test(
             ddp,

torchmetrics/__init__.py

@@ -13,7 +13,7 @@
 
 from torchmetrics import functional  # noqa: E402
 from torchmetrics.aggregation import CatMetric, MaxMetric, MeanMetric, MinMetric, SumMetric  # noqa: E402
-from torchmetrics.audio import PESQ, PIT, SDR, SI_SDR, SI_SNR, SNR, STOI  # noqa: E402
+from torchmetrics.audio import PIT, SDR, SI_SDR, SI_SNR, SNR  # noqa: E402
 from torchmetrics.classification import (  # noqa: E402
     AUC,
     AUROC,
@@ -40,8 +40,7 @@
     StatScores,
 )
 from torchmetrics.collections import MetricCollection  # noqa: E402
-from torchmetrics.detection import MAP  # noqa: E402
-from torchmetrics.image import FID, IS, KID, LPIPS, PSNR, SSIM  # noqa: E402
+from torchmetrics.image import PSNR, SSIM  # noqa: E402
 from torchmetrics.metric import Metric  # noqa: E402
 from torchmetrics.regression import (  # noqa: E402
     CosineSimilarity,
@@ -69,12 +68,10 @@
 from torchmetrics.text import (  # noqa: E402
     TER,
     WER,
-    BERTScore,
     BLEUScore,
     CharErrorRate,
     CHRFScore,
     MatchErrorRate,
-    ROUGEScore,
     SacreBLEUScore,
     SQuAD,
     WordInfoLost,
@@ -91,7 +88,6 @@
     "BinnedAveragePrecision",
     "BinnedPrecisionRecallCurve",
     "BinnedRecallAtFixedPrecision",
-    "BERTScore",
     "BLEUScore",
     "BootStrapper",
     "CalibrationError",
@@ -104,15 +100,10 @@
     "ExplainedVariance",
     "F1",
     "FBeta",
-    "FID",
     "HammingDistance",
     "Hinge",
     "IoU",
-    "IS",
-    "KID",
     "KLDivergence",
-    "LPIPS",
-    "MAP",
     "MatthewsCorrcoef",
     "MaxMetric",
     "MeanAbsoluteError",
@@ -127,7 +118,6 @@
     "MinMetric",
     "MultioutputWrapper",
     "PearsonCorrcoef",
-    "PESQ",
     "PIT",
     "Precision",
     "PrecisionRecallCurve",
@@ -143,7 +133,6 @@
     "RetrievalRecall",
     "RetrievalRPrecision",
     "ROC",
-    "ROUGEScore",
     "SacreBLEUScore",
     "SDR",
     "SI_SDR",
@@ -154,7 +143,6 @@
     "SQuAD",
     "SSIM",
     "StatScores",
-    "STOI",
     "SumMetric",
     "SymmetricMeanAbsolutePercentageError",
     "TER",

torchmetrics/audio/__init__.py

@@ -11,10 +11,8 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
-from torchmetrics.audio.pesq import PESQ  # noqa: F401
 from torchmetrics.audio.pit import PIT  # noqa: F401
 from torchmetrics.audio.sdr import SDR  # noqa: F401
 from torchmetrics.audio.si_sdr import SI_SDR  # noqa: F401
 from torchmetrics.audio.si_snr import SI_SNR  # noqa: F401
 from torchmetrics.audio.snr import SNR  # noqa: F401
-from torchmetrics.audio.stoi import STOI  # noqa: F401

torchmetrics/audio/pesq.py

@@ -62,7 +62,7 @@ class PESQ(Metric):
             If ``mode`` is not either ``"wb"`` or ``"nb"``
 
     Example:
-        >>> from torchmetrics.audio import PESQ
+        >>> from torchmetrics.audio.pesq import PESQ
         >>> import torch
         >>> g = torch.manual_seed(1)
         >>> preds = torch.randn(8000)

torchmetrics/audio/stoi.py

@@ -63,7 +63,7 @@ class STOI(Metric):
             If ``pystoi`` package is not installed
 
     Example:
-        >>> from torchmetrics.audio import STOI
+        >>> from torchmetrics.audio.stoi import STOI
         >>> import torch
         >>> g = torch.manual_seed(1)
         >>> preds = torch.randn(8000)

torchmetrics/detection/__init__.py

@@ -11,4 +11,3 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
-from torchmetrics.detection.map import MAP  # noqa: F401

torchmetrics/detection/map.py

@@ -179,6 +179,42 @@ class MAP(Metric):
             Callback that performs the allgather operation on the metric state. When ``None``, DDP
             will be used to perform the allgather
 
+    Example:
+        >>> import torch
+        >>> from torchmetrics.detection.map import MAP
+        >>> preds = [
+        ...   dict(
+        ...     boxes=torch.Tensor([[258.0, 41.0, 606.0, 285.0]]),
+        ...     scores=torch.Tensor([0.536]),
+        ...     labels=torch.IntTensor([0]),
+        ...   )
+        ... ]
+        >>> target = [
+        ...   dict(
+        ...     boxes=torch.Tensor([[214.0, 41.0, 562.0, 285.0]]),
+        ...     labels=torch.IntTensor([0]),
+        ...   )
+        ... ]
+        >>> metric = MAP()  # doctest: +SKIP
+        >>> metric.update(preds, target)  # doctest: +SKIP
+        >>> from pprint import pprint
+        >>> pprint(metric.compute())  # doctest: +SKIP
+        {'map': tensor(0.6000),
+         'map_50': tensor(1.),
+         'map_75': tensor(1.),
+         'map_small': tensor(-1.),
+         'map_medium': tensor(-1.),
+         'map_large': tensor(0.6000),
+         'mar_1': tensor(0.6000),
+         'mar_10': tensor(0.6000),
+         'mar_100': tensor(0.6000),
+         'mar_small': tensor(-1.),
+         'mar_medium': tensor(-1.),
+         'mar_large': tensor(0.6000),
+         'map_per_class': tensor(-1.),
+         'mar_100_per_class': tensor(-1.)
+        }
+
     Raises:
         ImportError:
             If ``torchvision`` is not installed or version installed is lower than 0.8.0

torchmetrics/functional/__init__.py

@@ -11,13 +11,11 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
-from torchmetrics.functional.audio.pesq import pesq
 from torchmetrics.functional.audio.pit import pit, pit_permutate
 from torchmetrics.functional.audio.sdr import sdr
 from torchmetrics.functional.audio.si_sdr import si_sdr
 from torchmetrics.functional.audio.si_snr import si_snr
 from torchmetrics.functional.audio.snr import snr
-from torchmetrics.functional.audio.stoi import stoi
 from torchmetrics.functional.classification.accuracy import accuracy
 from torchmetrics.functional.classification.auc import auc
 from torchmetrics.functional.classification.auroc import auroc
@@ -110,7 +108,6 @@
     "pairwise_linear_similarity",
     "pairwise_manhatten_distance",
     "pearson_corrcoef",
-    "pesq",
     "pit",
     "pit_permutate",
     "precision",
@@ -139,7 +136,6 @@
     "squad",
     "ssim",
     "stat_scores",
-    "stoi",
     "symmetric_mean_absolute_percentage_error",
     "ter",
     "wer",

torchmetrics/functional/audio/__init__.py

@@ -11,10 +11,8 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
-from torchmetrics.functional.audio.pesq import pesq  # noqa: F401
 from torchmetrics.functional.audio.pit import pit, pit_permutate  # noqa: F401
 from torchmetrics.functional.audio.sdr import sdr  # noqa: F401
 from torchmetrics.functional.audio.si_sdr import si_sdr  # noqa: F401
 from torchmetrics.functional.audio.si_snr import si_snr  # noqa: F401
 from torchmetrics.functional.audio.snr import snr  # noqa: F401
-from torchmetrics.functional.audio.stoi import stoi  # noqa: F401

torchmetrics/functional/audio/pesq.py

@@ -58,7 +58,7 @@ def pesq(preds: Tensor, target: Tensor, fs: int, mode: str, keep_same_device: bo
             If ``mode`` is not either ``"wb"`` or ``"nb"``
 
     Example:
-        >>> from torchmetrics.functional.audio import pesq
+        >>> from torchmetrics.functional.audio.pesq import pesq
         >>> import torch
         >>> g = torch.manual_seed(1)
         >>> preds = torch.randn(8000)

torchmetrics/functional/audio/stoi.py

@@ -59,7 +59,7 @@ def stoi(preds: Tensor, target: Tensor, fs: int, extended: bool = False, keep_sa
             If ``pystoi`` package is not installed
 
     Example:
-        >>> from torchmetrics.functional.audio import stoi
+        >>> from torchmetrics.functional.audio.stoi import stoi
         >>> import torch
         >>> g = torch.manual_seed(1)
         >>> preds = torch.randn(8000)

torchmetrics/functional/text/bert.py

@@ -547,6 +547,7 @@ def bert_score(
             If invalid input is provided.
 
     Example:
+        >>> from torchmetrics.functional.text.bert import bert_score
         >>> predictions = ["hello there", "general kenobi"]
         >>> references = ["hello there", "master kenobi"]
         >>> bert_score(predictions=predictions, references=references, lang="en")  # doctest: +SKIP