Clinton ThomasandLucas Bourtoule ee17d0d869 Use server-generated keys for late-interaction query caches (#51445) ...

Co-authored-by: Lucas Bourtoule <35483370+dhalf@users.noreply.github.com>

2026-09-03 07:15:58 +00:00

..

basic

[ROCm][CI] Solidify entrypoint LLM lifecycle (#51280)

2026-08-13 12:24:02 -05:00

classify

[ROCm][CI] Solidify entrypoint LLM lifecycle (#51280)

2026-08-13 12:24:02 -05:00

embed

[Bugfix][Frontend] Let pooling requests set padding (#51157)

2026-08-27 11:06:20 +00:00

reward

[ROCm][CI] Solidify entrypoint LLM lifecycle (#51280)

2026-08-13 12:24:02 -05:00

scoring

Use server-generated keys for late-interaction query caches (#51445)

2026-09-03 07:15:58 +00:00

token_classify

[ROCm][CI] Solidify entrypoint LLM lifecycle (#51280)

2026-08-13 12:24:02 -05:00

token_embed

[ROCm][CI] Solidify entrypoint LLM lifecycle (#51280)

2026-08-13 12:24:02 -05:00

__init__.py

[CI] Split pooling from entrypoints Test (#24632)

2025-09-11 01:53:09 -07:00

test_factories.py

fix(pooling): validate BGE-M3 combined task ownership (#51823)

2026-08-17 10:30:35 +00:00

test_io_processor.py

[Misc] Use VLLMValidationError in pooling input validation (#51931)

2026-08-13 06:13:25 +00:00

test_utils.py

[Perf] Use numpy zero-copy path for embedding float response serialization (#41681)

2026-05-07 19:42:21 -07:00