[Weekly Review] 2020/08/17-23

Published: by Creative Commons Licence (Last updated: )

2020/08/17-23

This week, I still worked on the Loosely Timed TLM.

I chose to use the for-loop to run the computational units and then clock the cycle instead of using the quantum keeper to sync the local clocks.

However, it still cost me more than 40 minutes to run a full test. So I think I really need to do some refactors to improve the efficiency.

This is one slide of Prof. Xavier Bresson. And it's quite straight forward to show me what dose the weight look like.

However, if it's not a full connection layer, then I think the weight will be much more abstract.

FC Layer

Back to Top
Share Post: