Fixing intel Supports windowing. (#2637)

0c478846 · Nicolas Patry · GitHub · 3dbdf63e · 0c478846
Unverified Commit 0c478846 authored Oct 11, 2024 by Nicolas Patry Committed by GitHub Oct 11, 2024
Hide whitespace changes
Inline Side-by-side

Showing with 2 additions and 2 deletions

server/text_generation_server/layers/attention/common.py server/text_generation_server/layers/attention/common.py +2 -2

No files found.
--- a/server/text_generation_server/layers/attention/common.py
+++ b/server/text_generation_server/layers/attention/common.py
@@ -68,5 +68,5 @@ else:
        def clamp(self, max):
            if SYSTEM == "rocm":
                return self
-            raise NotImplementedError("Not implemented seqlen for paged")
+            self.input_lengths = torch.clamp(self.input_lengths, max=max)
-            return Seqlen(torch.clamp(self.input_lengths, max=max))
+            return self