refactoring: remove redundant document marker token id

qdrant · Feb 6, 2025 · 9062cfe · 9062cfe
1 parent c0fedd7
commit 9062cfe
Showing 1 changed file with 2 additions and 2 deletions.
diff --git a/fastembed/late_interaction_multimodal/colpali.py b/fastembed/late_interaction_multimodal/colpali.py
@@ -34,15 +34,15 @@
 
 
 class ColPali(LateInteractionMultimodalEmbeddingBase, OnnxMultimodalModel[np.ndarray]):
-    DOCUMENT_MARKER_TOKEN_ID = 2
     QUERY_PREFIX = "Query: "
     BOS_TOKEN = "<s>"
     PAD_TOKEN = "<pad>"
     QUERY_MARKER_TOKEN_ID = [2, 5098]
     IMAGE_PLACEHOLDER_SIZE = (3, 448, 448)
     EMPTY_TEXT_PLACEHOLDER = np.array(
         [257152] * 1024 + [2, 50721, 573, 2416, 235265, 108]
-    )  # This is a tokenization of '<image>' * 1024 + '<bos>Describe the image.\n' line which is used as placeholder while processing just image
+    )  # This is a tokenization of '<image>' * 1024 + '<bos>Describe the image.\n' line which is used as placeholder
+    # while processing an image
     EVEN_ATTENTION_MASK = np.array([1] * 1030)
 
     def __init__(