Failure 'origBiasSize == outChans' loading weights to FPGA

[rubende]

Hello,

I have followed the steps indicated in this tutorial (https://github.com/Xilinx/Vitis-AI/blob/master/alveo/notebooks/image_classification_tensorflow.ipynb), and it works. After, I have repeated the tutorial with my own model, I am able to quantize and compile the model, but in the inference I get the following error:

[XDNN] loading xclbin settings from /opt/xilinx/overlaybins/xdnnv3/xdnn_v3_96x16_2pe_8b_9mb_bank03_2.xclbin.json [XDNN] using custom DDR banks 0,3 Path ./freeze_modelV2_partition_01-data.h5 is a file. Loading weights/bias/quant_params to FPGA... python: xdnn.cpp:1357: int XDNNV3FillWeightsBiasQuantBlob(short int*, int, std::__cxx11::string, std::__cxx11::string, const DType*, unsigned int, const DType*, unsigned int, short unsigned int, short unsigned int, unsigned int, unsigned int, int, int, int, int, bool, std::__cxx11::string, int, std::vector&) [with DType = float; std::__cxx11::string = std::__cxx11::basic_string]: Assertion `origBiasSize == outChans' failed.

Any suggestions? There is my inference code (https://pastebin.com/BP4CbUaz). Thank you.