torcheval

Reviewed By: jermenkoo

Differential Revision: D68317779

fbshipit-source-id: 36e00f1766bbe0baf7138320f3092bfafe65ae2d

Author: generatedunixname89002005287564

docs/source/ext/fbcode.py

@@ -1,5 +1,4 @@
 #!/usr/bin/env python3
-# -*- coding: utf-8 -*-
 # Copyright (c) Meta Platforms, Inc. and affiliates.
 # All rights reserved.
 #

docs/update_docs.py

@@ -8,10 +8,9 @@
 import ast
 import os
 from pathlib import Path
-from typing import List, Tuple
 
 
-def get_submodule_vars(filename) -> Tuple[str, List[str]]:
+def get_submodule_vars(filename) -> tuple[str, list[str]]:
     """This function reads the init files which are inside of the metrics/ and functional/ subdirectories.
     Each subdirectory within these two folders are associated with a domain of metrics, e.g. classification or regression.
     This function assumes that each of the init files has two variables defined.
@@ -27,7 +26,7 @@ def get_submodule_vars(filename) -> Tuple[str, List[str]]:
         doc_name (str): defined above
         all_modules (List[str]): defined above
     """
-    with open(filename, "r") as file:
+    with open(filename) as file:
         tree = ast.parse(file.read())
 
     doc_name = None

setup.py

@@ -10,8 +10,6 @@
 
 from datetime import date
 
-from typing import List
-
 from setuptools import find_packages, setup
 from torcheval import __version__
 
@@ -20,7 +18,7 @@ def current_path(file_name: str) -> str:
     return os.path.abspath(os.path.join(__file__, os.path.pardir, file_name))
 
 
-def read_requirements(file_name: str) -> List[str]:
+def read_requirements(file_name: str) -> list[str]:
     with open(current_path(file_name), encoding="utf8") as f:
         return f.read().strip().split()
 

tests/metrics/aggregation/test_cat.py

@@ -6,7 +6,6 @@
 
 # pyre-strict
 
-from typing import List
 
 import torch
 from torcheval.metrics import Cat
@@ -19,7 +18,7 @@
 
 class TestCat(MetricClassTester):
     def _test_cat_class_with_input(
-        self, input_val_tensors: List[torch.Tensor], dim: int = 0
+        self, input_val_tensors: list[torch.Tensor], dim: int = 0
     ) -> None:
         self.run_class_implementation_tests(
             metric=Cat(),

tests/metrics/aggregation/test_cov.py

@@ -6,15 +6,14 @@
 
 # pyre-strict
 
-from typing import List
 
 import torch
 from torcheval.metrics import Covariance
 from torcheval.utils.test_utils.metric_class_tester import MetricClassTester
 
 
 class TestCovariance(MetricClassTester):
-    def _test_covariance_with_input(self, batching: List[int]) -> None:
+    def _test_covariance_with_input(self, batching: list[int]) -> None:
         gen = torch.Generator()
         gen.manual_seed(3)
         X = torch.randn(sum(batching), 4, generator=gen)

tests/metrics/aggregation/test_mean.py

@@ -6,7 +6,6 @@
 
 # pyre-strict
 
-from typing import List, Union
 
 import numpy as np
 import torch
@@ -77,8 +76,8 @@ def test_mean_class_update_input_valid_weight(self) -> None:
         ]
 
         def _compute_result(
-            update_value: List[torch.Tensor],
-            update_weight: List[Union[float, int, torch.Tensor]],
+            update_value: list[torch.Tensor],
+            update_weight: list[float | int | torch.Tensor],
         ) -> torch.Tensor:
             weighted_sum = 0.0
             weights = 0.0

tests/metrics/aggregation/test_sum.py

@@ -6,7 +6,6 @@
 
 # pyre-strict
 
-from typing import List, Union
 
 import torch
 from torcheval.metrics import Sum
@@ -68,8 +67,8 @@ def test_sum_class_update_input_valid_weight(self) -> None:
         ]
 
         def _compute_result(
-            update_inputs: List[torch.Tensor],
-            update_weights: List[Union[float, torch.Tensor]],
+            update_inputs: list[torch.Tensor],
+            update_weights: list[float | torch.Tensor],
         ) -> torch.Tensor:
             weighted_sum = torch.tensor(0.0, dtype=torch.float64)
             for v, w in zip(update_inputs, update_weights):

tests/metrics/aggregation/test_throughput.py

@@ -7,7 +7,6 @@
 # pyre-strict
 
 import random
-from typing import List
 
 from torcheval.metrics import Throughput
 from torcheval.utils.test_utils.metric_class_tester import (
@@ -20,8 +19,8 @@
 class TestThroughput(MetricClassTester):
     def _test_throughput_class_with_input(
         self,
-        num_processed: List[int],
-        elapsed_time_sec: List[float],
+        num_processed: list[int],
+        elapsed_time_sec: list[float],
     ) -> None:
         num_individual_update = NUM_TOTAL_UPDATES // NUM_PROCESSES
         expected_num_total = sum(num_processed)

tests/metrics/audio/test_fad.py

@@ -7,7 +7,6 @@
 # pyre-strict
 
 import unittest
-from typing import Optional
 
 import numpy as np
 import torch
@@ -39,7 +38,7 @@ def gen_sine_wave(
     freq: float = 600,
     length_seconds: float = 6,
     sample_rate: int = 16_000,
-    std_dev: Optional[float] = None,
+    std_dev: float | None = None,
 ) -> torch.Tensor:
     """Creates sine wave of the specified frequency, sample_rate and length."""
     t = np.linspace(0, length_seconds, int(length_seconds * sample_rate))
@@ -53,7 +52,7 @@ def gen_sine_wave(
     return torch.from_numpy(np.asarray(2**15 * samples, dtype=np.int16)).float()
 
 
-def gen_fad_test_batch(num_files: int, std_dev: Optional[float]) -> torch.Tensor:
+def gen_fad_test_batch(num_files: int, std_dev: float | None) -> torch.Tensor:
     """Creates a tensor representing a batch of sine waves with optional
     gaussian noise added in."""
     frequencies = np.linspace(100, 1000, num_files).tolist()

tests/metrics/classification/test_auprc.py

@@ -8,7 +8,6 @@
 
 # pyre-ignore-all-errors[56]: Pyre was not able to infer the type of argument
 
-from typing import Optional, Tuple
 
 import numpy as np
 
@@ -427,7 +426,7 @@ def _get_sklearn_equivalent(
         self,
         input: torch.Tensor,
         target: torch.Tensor,
-        average: Optional[str] = "macro",
+        average: str | None = "macro",
         device: str = "cpu",
     ) -> torch.Tensor:
         # Convert input/target to sklearn style inputs
@@ -445,7 +444,7 @@ def _get_sklearn_equivalent(
 
     def _get_rand_inputs_multilabel(
         self, num_updates: int, num_labels: int, batch_size: int
-    ) -> Tuple[torch.Tensor, torch.Tensor]:
+    ) -> tuple[torch.Tensor, torch.Tensor]:
         input = torch.rand(size=[num_updates, batch_size, num_labels])
         targets = torch.randint(
             low=0, high=2, size=[num_updates, batch_size, num_labels]
@@ -457,7 +456,7 @@ def _check_against_sklearn(
         input: torch.Tensor,
         target: torch.Tensor,
         num_labels: int,
-        average: Optional[str] = "macro",
+        average: str | None = "macro",
         num_updates: int = NUM_TOTAL_UPDATES,
     ) -> None:
         device = "cuda" if torch.cuda.is_available() else "cpu"