EfficientNet backbone input resolution

alexander-pv / maskrcnn_tf2

Mask R-CNN for object detection and instance segmentation with Keras and TensorFlow V2 and ONNX and TensorRT optimization support.

Other

40 stars 11 forks source link

EfficientNet backbone input resolution #5

Open VeeranjaneyuluToka opened 2 years ago

VeeranjaneyuluToka commented 2 years ago

Hi, I have one more question regarding efficientnet input resolution. Looks like you are passing 512x512 irrespective of EfficientNet-B version, but they have some recommendation of input resolution in original paper. Have you tested them and they work fine even if input resolution is uniform.

Thanks, Veeru.

alexander-pv commented 2 years ago

Hi, @VeeranjaneyuluToka ,

I made tests with 256, 512 image sizes. So there is a lot of research to make. You can change input size in src/common/config.py as well as all the others MaskRCNN parameters.