Giovanni Canali
Giovanni Canali
**Is your feature request related to a problem? Please describe.** Python 3.8 is no longer supported. I suggest to remove tests and support for this version.
Fixes #440: add a type checker to be used for debugging. If the debugging level is set to `DEBUG`, the `enforce_type` decorator will enforce the type hints both for the...
**Describe the bug** PyTorch does not support `torch.compile` on Python 3.14+. This is clearly stated here https://github.com/pytorch/pytorch/blob/main/torch/__init__.py#L2608: ```python if sys.version_info >= (3, 14): raise RuntimeError("torch.compile is not supported on Python...
**Describe the bug** The attributes of any class inheriting from `Equation` are not automatically moved to the correct training device.
**Describe the bug** The classes `HelmholtzProblem` and `Advection` contain unintended self references, which result in unexpected behavior during execution. **Details** - In `HelmholtzProblem`, the method signature includes an extra self...
**Is your feature request related to a problem? Please describe.** Each domain in PINA should have a method which automatically identifies the boundaries of the domain itself.
Currently, there are no tests that guarantee solver convergence. Defining a benchmark and running convergence tests on a rolling basis would help ensure that solver behavior remains consistent with expectations....
**Describe the bug** After upgrading to PyTorch 2.8.0, compilation involving `LabelTensor` fails due to incorrect inference of `runtime_type` in the torch compilation pipeline. This results in a `RuntimeError` during the...
**Is your feature request related to a problem? Please describe.** SOAP is a quasi-second-order optimization algorithm that runs AdamW in the eigenbasis provided by Shampoo. **Additional context** The code is...