Merge branch 'main' into increase-coverage-for-tensor-strategy

.github/dependabot.yml

@@ -8,7 +8,7 @@ updates:
   - package-ecosystem: "pip"
     directory: "/" # Location of package manifests
     schedule:
-      interval: "daily"
+      interval: "weekly"
     target-branch: "main"
     reviewers:
       - "qthequartermasterman"
@@ -19,10 +19,14 @@ updates:
         exclude-patterns:
           - "hypothesis"
           - "torch*"
+          - "transformers"
       direct-dependencies:
         patterns:
           - "hypothesis"
           - "torch*"
+      optional-dependencies:
+        patterns:
+          - "transformers"
   - package-ecosystem: github-actions
     directory: /
     schedule:

.github/workflows/build-reusable.yml

@@ -48,7 +48,7 @@ jobs:
     - name: Install Specific PyTorch version
       if: ${{ inputs.torch-version != '' }}
       run: |
-        uv pip install torch==${{ inputs.torch-version }}
+        uv pip install 'torch==${{ inputs.torch-version }}+cpu' --extra-index-url https://download.pytorch.org/whl/cpu
     - name: Install dependencies
       run: |
         uv pip install ".[dev,huggingface]"

.github/workflows/build.yml

@@ -19,16 +19,15 @@ permissions:
 jobs:
   build:
     strategy:
+      fail-fast: false
       matrix:
         python-version: ['3.9', '3.10', '3.11', '3.12']
+        torch-version: ['2.1.2', '2.2.2', '2.3.1', '2.4.0', '']
+        exclude:
+          - python-version: '3.12'
+            torch-version: '2.1.2'
     uses: ./.github/workflows/build-reusable.yml
     with:
       python-version: ${{ matrix.python-version }}
+      torch-version: ${{ matrix.torch-version }}
       lint: ${{ matrix.python-version == '3.9' }}
-
-  build-pytorch-1-13:
-    uses: ./.github/workflows/build-reusable.yml
-    with:
-      python-version: '3.9'
-      lint: false
-      torch-version: '1.13.1'

hypothesis_torch/__init__.py

@@ -4,7 +4,7 @@
 lacks built-in support for Pytorch tensors and modules, so this library provides strategies for generating them.
 """
 
-__version__ = "0.7.11"
+__version__ = "0.7.13"
 import importlib.util
 
 from hypothesis_torch.device import (

hypothesis_torch/optim.py

@@ -13,12 +13,15 @@
 
 from hypothesis_torch import inspection_util
 
-OptimizerConstructorWithOnlyParameters: TypeAlias = Callable[[Iterable[torch.nn.Parameter]], torch.optim.Optimizer]
+# We do an alias here to avoid some type checking issues with torch 2.4.0
+Optimizer: TypeAlias = torch.optim.Optimizer  # pyright: ignore[reportPrivateImportUsage]
 
-OPTIMIZERS: Final[tuple[type[torch.optim.Optimizer], ...]] = tuple(
+OptimizerConstructorWithOnlyParameters: TypeAlias = Callable[[Iterable[torch.nn.Parameter]], Optimizer]
+
+OPTIMIZERS: Final[tuple[type[Optimizer], ...]] = tuple(
     optimizer
-    for optimizer in inspection_util.get_all_subclasses(torch.optim.Optimizer)
-    if optimizer is not torch.optim.Optimizer and "NewCls" not in optimizer.__name__
+    for optimizer in inspection_util.get_all_subclasses(Optimizer)
+    if optimizer is not Optimizer and "NewCls" not in optimizer.__name__
 )
 
 _ZERO_TO_ONE_FLOATS: Final[st.SearchStrategy[float]] = st.floats(
@@ -63,8 +66,8 @@ def betas(draw: st.DrawFn) -> tuple[float, float]:
 
 
 def optimizer_type_strategy(
-    allowed_optimizer_types: Sequence[type[torch.optim.Optimizer]] | None = None,
-) -> st.SearchStrategy[type[torch.optim.Optimizer]]:
+    allowed_optimizer_types: Sequence[type[Optimizer]] | None = None,
+) -> st.SearchStrategy[type[Optimizer]]:
     """Strategy for generating torch optimizers.
 
     Args:
@@ -81,7 +84,7 @@ def optimizer_type_strategy(
 @st.composite
 def optimizer_strategy(
     draw: st.DrawFn,
-    optimizer_type: type[torch.optim.Optimizer] | st.SearchStrategy[type[torch.optim.Optimizer]] | None = None,
+    optimizer_type: type[Optimizer] | st.SearchStrategy[type[Optimizer]] | None = None,
     **kwargs: Any,  # noqa: ANN401
 ) -> OptimizerConstructorWithOnlyParameters:
     """Strategy for generating torch optimizers.
@@ -121,7 +124,7 @@ def optimizer_strategy(
     hypothesis.note(f"Chosen optimizer type: {optimizer_type}")
     hypothesis.note(f"Chosen optimizer hyperparameters: {kwargs}")
 
-    def optimizer_factory(params: Iterable[torch.nn.Parameter]) -> torch.optim.Optimizer:
+    def optimizer_factory(params: Iterable[torch.nn.Parameter]) -> Optimizer:
         return optimizer_type(params, **kwargs)
 
     return optimizer_factory

hypothesis_torch/tensor.py

@@ -184,7 +184,8 @@ def downcast(x: float) -> float:
     hypothesis.assume(memory_format != torch.channels_last or len(tensor.shape) == 4)
     # channel_last_3d memory format is only supported for 5D tensors
     hypothesis.assume(memory_format != torch.channels_last_3d or len(tensor.shape) == 5)
-    tensor = tensor.to(memory_format=memory_format)
+    # Pyright/mypy falsely reports an error here on py3.9 torch 2.1.2.
+    tensor = tensor.to(memory_format=memory_format)  # type: ignore
 
     return tensor
 

pyproject.toml

@@ -9,7 +9,7 @@ description = "Hypothesis strategies for various Pytorch structures, including t
 dynamic = ["version"]
 dependencies = [
     "hypothesis>=6.0.0",
-    "torch>=1.13.0",
+    "torch>=2.1.0",
 ]
 requires-python = ">=3.9"
 authors=[{name="Andrew P. Sansom", email="[email protected]"}]