quantize Unet model under openVino NNCF

[timosy]

I wonder if I can quantize a U-net model under openVino NNCF framework.

I prepared and trained a symple model of U-net. Is it possible to quantize teh U-net w. NNCF. Are there any person who tried it? I'd like to know whether a class such as "BatchNorm2d, Conv2D, Down, Up,  DoubleConvBNReLU " are available or not . 

[Peh_Intel · Accepted solution]

Hi timosy,

You can refer to the Semantic segmentation sample for the demonstration Deep Learning model compression capabilities for semantic segmentation problem.

As long as your UNet model is implemented as close as possible to the original papers, then it should be able to be quantized.

Regards,

Peh

[Peh_Intel]

Hi timosy,

This thread will no longer be monitored since this issue has been resolved. If you need any additional information from Intel, please submit a new question. 

Regards,

Peh