spooNN
spooNN copied to clipboard
fpgasystems
FPGA-based neural network inference project with an end-to-end approach (from training to implementation to deployment)
Why when I change bitw to int 4 , the training correctness is downing to 0.1 or 1. I am not sure where the problem is. Can anybody help me?
Dear Sir, I am trying your notebook "halfsqueezenet-objdetect.ipynb" but dma.recvchannel.wait() stuck and keep waiting after calling the fire function. Strange that function "weightsfactors_transfer" went fine. I tried your "mnist-cnn.ipynb" and...
For linuix code, you write command"$ export PYTHONPATH=/path/to/spooNN/hls-nn-lib/training:$PYTHONPATH", but export command isnt work for windows. What should I do?
Helllo, I wanted to thank you for such interesting work. I wanted to ask what the difference between the IPs for conv layer is? For eg: How are `CONV2D_ACT_KP` ,...
I am not able to understand what these factorA and factorB params are in the trained network. Can someone provide a hint ..
Hello, I'd worked on your repo, finding it's too slow when detecting. It turns on the for circulations slow the code. I fixed it.
Hello, How can I test mnist.t as a testbench for the verilog files created using hls? Any help is greatly appreciated. Regards,
Hello, I'am an engineering student and I tried to use your code with dog-cat available database. I finished the training but when I want to compile the C++ code using...
following your guide, I open an empyt hls_project when I /path/to/spooNN/halfsqueezenet$ ./scripts/make_IP.sh /path/to/spooNN/halfsqueezenet. how should i do
Metadata
Owner
Metadata
FPGA-based neural network inference project with an end-to-end approach (from training to implementation to deployment)