test: Don't check old PyTorch versions

argus/callbacks/lr_schedulers.py

@@ -4,7 +4,6 @@
 """
 from typing import Optional, Callable, Iterable, Any, Union, List
 
-import torch
 from torch.optim import Optimizer
 from torch.optim import lr_scheduler as _scheduler
 
@@ -370,16 +369,12 @@ def __init__(self,
                                   List[Callable[[int], float]]],
                  last_epoch: int = -1,
                  step_on_iteration: bool = False):
-        from distutils.version import LooseVersion
-        if LooseVersion(torch.__version__) >= LooseVersion("1.4.0"):
-            super().__init__(
-                lambda opt: _scheduler.MultiplicativeLR(opt,
-                                                        lr_lambda,
-                                                        last_epoch=last_epoch),
-                step_on_iteration=step_on_iteration
-            )
-        else:
-            raise ImportError("Update torch>=1.4.0 to use 'MultiplicativeLR'")
+        super().__init__(
+            lambda opt: _scheduler.MultiplicativeLR(opt,
+                                                    lr_lambda,
+                                                    last_epoch=last_epoch),
+            step_on_iteration=step_on_iteration
+        )
 
 
 class OneCycleLR(LRScheduler):
@@ -451,23 +446,19 @@ def __init__(self,
                  div_factor: float = 25.,
                  final_div_factor: float = 1e4,
                  last_epoch: int = -1):
-        from distutils.version import LooseVersion
-        if LooseVersion(torch.__version__) >= LooseVersion("1.3.0"):
-            super().__init__(
-                lambda opt: _scheduler.OneCycleLR(opt,
-                                                  max_lr,
-                                                  total_steps=total_steps,
-                                                  epochs=epochs,
-                                                  steps_per_epoch=steps_per_epoch,
-                                                  pct_start=pct_start,
-                                                  anneal_strategy=anneal_strategy,
-                                                  cycle_momentum=cycle_momentum,
-                                                  base_momentum=base_momentum,
-                                                  max_momentum=max_momentum,
-                                                  div_factor=div_factor,
-                                                  final_div_factor=final_div_factor,
-                                                  last_epoch=last_epoch),
-                step_on_iteration=True
-            )
-        else:
-            raise ImportError("Update torch>=1.3.0 to use 'OneCycleLR'")
+        super().__init__(
+            lambda opt: _scheduler.OneCycleLR(opt,
+                                              max_lr,
+                                              total_steps=total_steps,
+                                              epochs=epochs,
+                                              steps_per_epoch=steps_per_epoch,
+                                              pct_start=pct_start,
+                                              anneal_strategy=anneal_strategy,
+                                              cycle_momentum=cycle_momentum,
+                                              base_momentum=base_momentum,
+                                              max_momentum=max_momentum,
+                                              div_factor=div_factor,
+                                              final_div_factor=final_div_factor,
+                                              last_epoch=last_epoch),
+            step_on_iteration=True
+        )

pyproject.toml

@@ -70,7 +70,7 @@ select = [
 [tool.ruff.per-file-ignores]
 "__init__.py" = ["F401"]
 
-[tool.pytest]
+[tool.pytest.ini_options]
 minversion = 6.0
 addopts = "--cov=argus"
-testpaths = "tests"
+testpaths = ["tests"]

tests/callbacks/test_lr_schedulers.py

@@ -1,8 +1,6 @@
 import pytest
 from collections import Counter
-from distutils.version import LooseVersion
 
-import torch
 from torch.optim import lr_scheduler
 from torch.optim.optimizer import Optimizer
 
@@ -104,9 +102,7 @@ def test_cosine_annealing_lr(self, test_engine):
         assert cosine_annealing_lr.scheduler.T_max == 10
         assert cosine_annealing_lr.scheduler.eta_min == 0
 
-    @pytest.mark.skipif(LooseVersion(torch.__version__) < LooseVersion("1.4.0"),
-                        reason="Requires torch==1.4.0 or higher")
-    def test_multiplicative_lr(self, test_engine, step_on_iteration, monkeypatch):
+    def test_multiplicative_lr(self, test_engine, step_on_iteration):
         multiplicative_lr = MultiplicativeLR(lambda epoch: 0.95,
                                              step_on_iteration=step_on_iteration)
         multiplicative_lr.attach(test_engine)
@@ -115,26 +111,14 @@ def test_multiplicative_lr(self, test_engine, step_on_iteration, monkeypatch):
         assert multiplicative_lr.scheduler.lr_lambdas[0](1) == 0.95
         assert multiplicative_lr.step_on_iteration == step_on_iteration
 
-        from argus.callbacks.lr_schedulers import torch
-        monkeypatch.setattr(torch, "__version__", '1.3.0')
-        with pytest.raises(ImportError):
-            MultiplicativeLR(lambda epoch: 0.95)
-
-    @pytest.mark.skipif(LooseVersion(torch.__version__) < LooseVersion("1.3.0"),
-                        reason="Requires torch==1.3.0 or higher")
-    def test_one_cycle_lr(self, test_engine, monkeypatch):
+    def test_one_cycle_lr(self, test_engine):
         one_cycle_lr = OneCycleLR(max_lr=0.01, steps_per_epoch=1000, epochs=10)
         one_cycle_lr.attach(test_engine)
         one_cycle_lr.start(test_engine.state)
         assert isinstance(one_cycle_lr.scheduler, lr_scheduler.OneCycleLR)
         assert one_cycle_lr.scheduler.total_steps == 10000
         assert one_cycle_lr.step_on_iteration
 
-        from argus.callbacks.lr_schedulers import torch
-        monkeypatch.setattr(torch, "__version__", '1.1.0')
-        with pytest.raises(ImportError):
-            OneCycleLR(max_lr=0.01, steps_per_epoch=1000, epochs=10)
-
     def test_cosine_annealing_warm_restarts(self, test_engine, step_on_iteration):
         warm_restarts = CosineAnnealingWarmRestarts(T_0=1, T_mult=1, eta_min=0,
                                                     step_on_iteration=step_on_iteration)

tests/test_optimizer.py

@@ -1,6 +1,3 @@
-import pytest
-from distutils.version import LooseVersion
-
 import torch
 
 from argus.optimizer import get_pytorch_optimizers, _is_pytorch_optimizer
@@ -19,8 +16,6 @@ def test_is_pytorch_optimizer():
     assert not _is_pytorch_optimizer(torch.nn.BCELoss)
 
 
-@pytest.mark.skipif(LooseVersion(torch.__version__) < LooseVersion("1.7.0"),
-                    reason="Requires torch==1.7.0 or higher")
 def test_is_multi_tensor_optimizer():
     from torch.optim import _multi_tensor
     assert not _is_pytorch_optimizer(_multi_tensor.SGD)