Implement 3-dimensional (3-D) multiply-accumulate (MAC) by using a HLS-friendly 1-D convolution structure which flattened the 3-D convolution

doonny / PipeCNN

An OpenCL-based FPGA Accelerator for Convolutional Neural Networks

Apache License 2.0

1.22k stars 370 forks source link

Implement 3-dimensional (3-D) multiply-accumulate (MAC) by using a HLS-friendly 1-D convolution structure which flattened the 3-D convolution #73

Closed ArjunaDeSoysa closed 6 years ago

ArjunaDeSoysa commented 6 years ago

Dear @aazz44ss, @doonny I have a problem in the topic as i said, how can convert in 3-D MAC using 1D and the what is the procedure in multiplier adder tree. Can you please tell me the whats the procedure of these. If you are busy can you please send me the some links or resource regarding above things.

Thank You

ArjunaDeSoysa commented 6 years ago

Dear @thinkoco , @doonny , @aazz44ss , @hammadj, @davenso can i know the what is these parameters (f^' , W_l , D_i) are? Please help me to find outthis one. Thank you. 1234

aazz44ss commented 6 years ago

maybe you need to read some document about convolution .

ArjunaDeSoysa commented 6 years ago

Okay i got it. Thank you.