Low precision decentralized distributed training over IID and non-IID data
2022
drop in accuracy. The proposed low precision decentralized training decreases computational complexity, memory usage, and communication cost by and compute energy by a factor of , while trading off less than a 1% accuracy for both IID and non-IID data. In particular, for higher skew values, we observe an increase in accuracy (by
- Correction
- Source
- Cite
- Save
- Machine Reading By IdeaReader
0
References
0
Citations
NaN
KQI