text-generation-inference
0a609731
- Fixing the batching tokenization in flash causal lm.
Go
Login via GitHub
Home
Pricing
FAQ
Install
Login
via GitHub
Commit
View On
GitHub
Commit
1 year ago
Fixing the batching tokenization in flash causal lm.
References
#2449 - Lots of improvements (Still 2 allocators)
Author
Narsil
Parents
c6f1a612
Loading