How to decrease inference time of LayoutXLM?

[piegu]

Hi,

I'm using Hugging Face libraries in order to run LayoutXLM. How can I decrease inference time? Which code to use?

I've already tested BetterTransformer(Optimum) and ONNXbut none of them accepts LayoutXLMmodel.

• BetterTransformer: NotImplementedError: The model type layoutlmv2 is not yet supported to be used with BetterTransformer.
• ONNX: KeyError: "layoutlmv2 is not supported yet.

Thank you.

Note: I asked this question here, too: https://github.com/microsoft/unilm/issues/1080

[NielsRogge]

Hi,

Feel free to open an issue on the Optimum library!

[piegu]

Issue opened in the Optimum library: https://github.com/huggingface/optimum/issues/1024