Skip to content

Commit

Permalink
Do NOT run pretrained llama (#227)
Browse files Browse the repository at this point in the history
  • Loading branch information
Delaunay authored Jun 19, 2024
1 parent 22d5143 commit 77a3ed7
Showing 1 changed file with 2 additions and 2 deletions.
4 changes: 2 additions & 2 deletions config/base.yaml
Original file line number Diff line number Diff line change
@@ -79,8 +79,8 @@ llama:
# Note: when NOT using pretrained model
# the benchmark becomes much harder as no end token is ever outputted by the model
# which makes inference much slower
argv:
--pretrained: true
# argv:
# --pretrained: true


_hf:

0 comments on commit 77a3ed7

Please sign in to comment.