Medical-SAM-Adapter icon indicating copy to clipboard operation
Medical-SAM-Adapter copied to clipboard

export onnx

Open Awesome0324 opened this issue 11 months ago • 1 comments

Hello, I encountered some issues while exporting MSA to onnx.

I used MSA to train the REFUGE dataset and set multiask_output to 2 as required, successfully saving checkpoint best. pth.

But when I wanted to export checkpoint_best.pth as onnx, it prompted me that the forward parameter in torch's onnx was missing the multiask_output parameter. I would like to inquire about how to solve this problem.

Awesome0324 avatar Mar 19 '24 01:03 Awesome0324