llama.cpp
6b86bcff
- cuda : increase max block size to 1024
Go
Login via GitHub
Home
Pricing
FAQ
Install
Login
via GitHub
Commit
View On
GitHub
Commit
1 year ago
cuda : increase max block size to 1024
References
#4256 - ggml : add ggml_soft_max_ext
Author
ggerganov
Parents
62532c05
Loading