search

fangwei123456 / spikingjelly

SpikingJelly is an open-source deep learning framework for Spiking Neural Network (SNN) based on PyTorch.
https://spikingjelly.readthedocs.io
Other
1.24k stars 235 forks source link

Deploy Spikformer on FPGA. Save pytorch model parameters in ANN-SNN way, can be used to build FPGA #500

Open Xi-W opened 4 months ago

Xi-W commented 4 months ago

Deploy Spikformer on FPGA. Save pytorch model parameters in ANN-SNN way, can be used to build FPGA？

fangwei123456 commented 4 months ago

Hi, I do not know how to use FPGAs. We can wait for someone who knows it to answer this question.

Xi-W commented 4 months ago

Hi, I do not know how to use FPGAs. We can wait for someone who knows it to answer this question.

In the FPGA, the data is calculated according to the time order, such as the classic Lenet network, the network weight and bias are trained on the PC side and placed in the FPGA, and the input data is entered into the convolution operation according to the time order, and the final result is obtained. Is there a strict order between the neurons of your SNN+Transformer network?