issues
search
PygmalionAI
/
aphrodite-engine
PygmalionAI's large-scale inference engine
https://pygmalion.chat
GNU Affero General Public License v3.0
606
stars
78
forks
source link
feat: optimized layernorm kernels
#398
Closed
AlpinDale
closed
1 month ago
AlpinDale
commented
1 month ago
Should reduce latency in the layernorm by a fair bit.
Should reduce latency in the layernorm by a fair bit.