ONNX EXPORT FOR VIDEO SEGMENTATION

I’m currently working with the SAM2 model and noticed that while ONNX export is supported for image segmentation, there is no straightforward way to export the full video segmentation pipeline (including memory components like memory_encoder, memory_attention, etc.) to ONNX.

It would be very helpful to have guidance or official support for exporting the video segmentation mode of SAM2 to ONNX format. This would enable broader deployment and optimization in real-time or resource-constrained environments.

Could you please provide instructions or add support for ONNX export of video segmentation in SAM2.