Describe the bug
I assume due to pytorch upgrade. Happens on 2.8, unsure about earlier.
Full message:
[W1106 11:25:56.492120877 AllocatorConfig.cpp:28] Warning: PYTORCH_CUDA_ALLOC_CONF is deprecated, use PYTORCH_ALLOC_CONF instead (function operator())
Comes from this line:
|
os.environ["PYTORCH_CUDA_ALLOC_CONF"] = settings.pytorch.pytorch_cuda_alloc_conf |
Steps/Code to Reproduce
Run python examples/tabpfn_for_multiclass_classification.py
Expected Results
No warning
Actual Results
Warning
Versions