Made a new PJ model which is FPGA friendly (at least the encoder part). From what I understand, nn.Unflatten() and nn.ConvTranspose2d isn't supported by hls4ml, so if you have any ideas @PRAkTIKal24 how to fix this, please let me know.
This PR only makes additions to the model file and adds an if statement to the encoder/decoder separate saving function in data_processing.py, so everything should work as before.
Made a new PJ model which is FPGA friendly (at least the encoder part). From what I understand,
nn.Unflatten()
andnn.ConvTranspose2d
isn't supported byhls4ml
, so if you have any ideas @PRAkTIKal24 how to fix this, please let me know.This PR only makes additions to the model file and adds an if statement to the encoder/decoder separate saving function in
data_processing.py
, so everything should work as before.