Parameter counts are different for some models when you import the same model from prototxt (Caffe) and json (Keras). The count for prototxt is wrong.
Steps to reproduce:
Export one of the following models to Keras and Caffe from the model zoo. Note the parameter count:
Inception V3
Inception V4
Yolo
ResNet 101
Import the models back to Fabrik via Keras and Caffe import. Note that the parameter count is different for the Keras method and the Caffe method. The parameter count for the Caffe method would not match the Keras method.
Parameter Counts:
Name
Model Zoo
Keras Import
Caffe Import
Inception V3
23,851,784
23,851,784
30,013,000
Inception V4
42,712,937
42,712,937
47,357,129
Yolo
271,703,550
217,177,598
3,354,516,990
ResNet 101
44,654,504
44,654,504
23,490,158,504
Notes:
I have not tested for all models, so errors may exist in some others as well.
Some models like VGG16 have no errors. As in, parameter counts are equal for Model zoo, Caffe and Keras.
Parameter count in Keras mostly matches the model zoo values, except in Yolo (for the ones I have tested).
Some models have very high error in parameter count in Caffe.
Parameter counts are different for some models when you import the same model from
prototxt
(Caffe) andjson
(Keras). The count for prototxt is wrong.Steps to reproduce:
Export one of the following models to Keras and Caffe from the model zoo. Note the parameter count:
Import the models back to Fabrik via Keras and Caffe import. Note that the parameter count is different for the Keras method and the Caffe method. The parameter count for the Caffe method would not match the Keras method.
Parameter Counts:
Notes: