Skip to content

Commit 9062cfe

Browse files
committed
refactoring: remove redundant document marker token id
1 parent c0fedd7 commit 9062cfe

File tree

1 file changed

+2
-2
lines changed

1 file changed

+2
-2
lines changed

fastembed/late_interaction_multimodal/colpali.py

Lines changed: 2 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -34,15 +34,15 @@
3434

3535

3636
class ColPali(LateInteractionMultimodalEmbeddingBase, OnnxMultimodalModel[np.ndarray]):
37-
DOCUMENT_MARKER_TOKEN_ID = 2
3837
QUERY_PREFIX = "Query: "
3938
BOS_TOKEN = "<s>"
4039
PAD_TOKEN = "<pad>"
4140
QUERY_MARKER_TOKEN_ID = [2, 5098]
4241
IMAGE_PLACEHOLDER_SIZE = (3, 448, 448)
4342
EMPTY_TEXT_PLACEHOLDER = np.array(
4443
[257152] * 1024 + [2, 50721, 573, 2416, 235265, 108]
45-
) # This is a tokenization of '<image>' * 1024 + '<bos>Describe the image.\n' line which is used as placeholder while processing just image
44+
) # This is a tokenization of '<image>' * 1024 + '<bos>Describe the image.\n' line which is used as placeholder
45+
# while processing an image
4646
EVEN_ATTENTION_MASK = np.array([1] * 1030)
4747

4848
def __init__(

0 commit comments

Comments
 (0)