A dynamically recon gurable accelerator design using a sparse-winograd decomposition algorithm for cnns

HIGHLIGHTS

• who: Tech Science Press and colleagues from the National University of Defence Technology, Changsha, China have published the Article: A Dynamically Recon gurable Accelerator Design Using a Sparse-Winograd Decomposition Algorithm for CNNs, in the Journal: (JOURNAL)
• what: In light of this, this paper proposes a block parallel computing method based on the Sparse-Winograd algorithm and a reconfigurable decomposition method for lightweight CNNs, which solves the problems associated with accelerating different convolution types. To solve the above problems, the authors propose a method of FDWA based on the Sparse-Winograd algorithm, which operated by . . .
  

   

  

  If you want to have access to all the content you need to log in!

  Thanks :)

  Username or E-mail:

  If you don't have an account, you can create one here.