Skip to content
GitLab
Explore
Sign in
Register
Xiph.Org
rnnoise
Repository
rnnoise
training
dump_rnn.py
Find file
Blame
History
Permalink
constraining the weights to +/- 0.5 so they can be quantized with 8 bits
· 4d1e630a
Jean-Marc Valin
authored
Aug 14, 2017
4d1e630a