Skip to content

Commit

Permalink
webis-de#6 tokens als liste
Browse files Browse the repository at this point in the history
  • Loading branch information
lisalehna committed Dec 5, 2024
1 parent 6c56367 commit b911165
Showing 1 changed file with 3 additions and 3 deletions.
6 changes: 3 additions & 3 deletions lightning_ir/models/mvr/tokenizer.py
Original file line number Diff line number Diff line change
Expand Up @@ -7,11 +7,11 @@ def __init__(
):
super().__init__()
if num_viewer_tokens is not None:
viewer_tokens = " ".join(f"[VIE{idx}]" for idx in range(num_viewer_tokens))
viewer_tokens = [f"[VIE{idx}]" for idx in range(num_viewer_tokens)]
self.add_tokens(viewer_tokens, special_tokens=True)
viewer_token_ids = [(f"[VIE{viewer_token_id}]", self.viewer_token_id(f"[VIE{viewer_token_id}]")) for viewer_token_id in range(num_viewer_tokens)]
viewer_token_ids = [(viewer_tokens[viewer_token_id], self.viewer_token_id(viewer_token_id)) for viewer_token_id in range(num_viewer_tokens)]
self.doc_post_processor = TemplateProcessing(
single=f"{viewer_tokens} $0 [SEP]",
single=f"{" ".join(viewer_tokens)} $0 [SEP]",
pair=f"[CLS] {self.QUERY_TOKEN} $A [SEP] $B:1 [SEP]:1",
special_tokens=[
("[CLS]", self.cls_token_id),
Expand Down

0 comments on commit b911165

Please sign in to comment.