/cs259-project

Course project of CS259, UCLA, 21 FALL

Primary LanguagePython

1. To run CNN training and model pruning

Install Numpy package on Linux machine

pip3 install torch==1.10.0+cpu torchvision==0.11.1+cpu torchaudio==0.10.0+cpu -f https://download.pytorch.org/whl/cpu/torch_stable.html

Train and prune the network

cd sw
mkdir models # create dir to save model checkpoints in different epochs
python3 model_training.py

2. To synthesis the HLS hardware code

Require vitis_hls installed

cd hw
vitis_hls run_hls.tcl
# to view the report, run `less proj_mulcol/solution1/syn/report/matmul_csynth.rpt`

To run the second version of systolic array, cover mulcol.cpp with mulcol.cpp.bk, then redo it.