ElegantGod commented on issue #11121: ssd inference bug URL: https://github.com/apache/incubator-mxnet/issues/11121#issuecomment-393916464 Thank for you reply. In python vision, I can fix the input size of inference equal to the input size of training by using "self.mod.bind(data_shapes=[('data', (batch_size, 3, data_shape, data_shape))])", which can generate the result close to truth. But in c++ version, I only use "simple_bind" function to generate the executor, the shape of executor.data is the shape of input image. OK, my English is poor, I hope you can understand it, thanks!
---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services
