Added keep tokens to keep so many tokens in a prompt when dropping

This commit is contained in:
Jaret Burkett committed 2024-03-18 13:18:25 -06:00
1 parent 89f4bcad2e
commit 9c1cc9641e
2 files changed
+16 -8

No files matched your search

+1
View File
@@ -460,6 +460,7 @@ class DatasetConfig:
self.token_dropout_rate: float = float(kwargs.get('token_dropout_rate', 0.0))
self.shuffle_tokens: bool = kwargs.get('shuffle_tokens', False)
self.caption_dropout_rate: float = float(kwargs.get('caption_dropout_rate', 0.0))
self.keep_tokens: int = kwargs.get('keep_tokens', 0) # #of first tokens to always keep unless caption dropped
self.flip_x: bool = kwargs.get('flip_x', False)
self.flip_y: bool = kwargs.get('flip_y', False)
self.augments: List[str] = kwargs.get('augments', [])