6 Replies Latest reply on Jun 26, 2018 2:09 AM by Intel Corporation

    Failure in faster-rcnn training

    Norton_Wang

      I try to train my model with py-faster-rcnn provide in the path /glob/deep-learning/py-faster-rcnn.

       

      trainBUG1.PNG

      But The following error occurred:

      0420 07:05:22.753042  3197 smooth_L1_loss_layer.cpp:54] Not Implemented Yet

      trainBUG2.PNG

       

      I found out that the source code of  py-faster-rcnn hasn't implemented the roi_pooling_layer and smooth_L1_loss_layer for CPU.

      Then I found the modified layer file on github GitHub - neuleaf/faster-rcnn-cpu .

      What should I do next?