https://github.com/Microsoft/CNTK
History
Tip revision: 5cc66af4ddf6c2ed9574e3e8d7b2ba895cb8da58 authored by Amit Agarwal on 13 December 2016, 06:34:50 UTC
CNTK parallel training: Turned off quantization for gradient matrices where row count is too small to benefit from quantization
Tip revision: 5cc66af
File Mode Size
Configuration Files.md -rw-r--r-- 93.7 KB
External Buffer Behavior.md -rw-r--r-- 6.0 KB
Model Editing Language.md -rw-r--r-- 36.1 KB
Network Description Language.md -rw-r--r-- 28.3 KB
PerformanceChart.png -rw-r--r-- 33.9 KB

back to top