[Model] Streamline EmbeddingGemma 2 config resolution and harmonize Triton prefill attention (#60289)

Signed-off-by: Luciano Martins <lucianommartins@users.noreply.github.com>
Co-authored-by: Luciano Martins <lucianommartins@users.noreply.github.com>
Co-authored-by: Matthew Bonanni <mbonanni@redhat.com>
This commit is contained in:
Luciano Martins
2026-10-07 14:56:40 -07:00
committed by GitHub
co-authored by Luciano Martins Matthew Bonanni
parent 947dd62c07
commit b803edbc9f
3 changed files with 57 additions and 164 deletions
+1 -1
View File
@@ -1139,7 +1139,7 @@ def get_sentence_transformer_tokenizer_config(
logger.info("Found sentence-transformers tokenize configuration.")
if all(k in encoder_dict for k in ("max_seq_length", "do_lower_case")):
if "max_seq_length" in encoder_dict:
return encoder_dict
return None