Open addisonklinke opened 3 months ago
Hey, I'm assuming you're referencing the torch.onnx.export() api. If you can try torch.onnx.export(...,dynamo=True) you might have a better outcome!
Actually my use-case is in PySpark and not torch. From my understanding, DEFAULT_OPSET_NUMBER
is internal to onnxconverter-common
so any 3rd party library which uses it for conversion would be limited to that opset
@MaanavD any direction for how we could bump the opset version?
@addisonklinke going to add @gramalingam to this thread, I think he'd know best!
@xadupre knows more about these converters.
DEFAULT_OPSET_NUMBER
is currently 15 and was last updated in Nov 2021. This corresponds to a max ofonnx==1.10.2
from the official versioning table which is 6 minor versions behind the latest 1.16.0. Additionally, 1.10.2 only has wheels for Python <=3.9 which is EOL Oct 2025 and makes using a more modern env difficultWhat are the limitations in upgrading this / why is it lagging so far behind the
onnx
releases?