vllm/tests/entrypoints at fd3e0cac12e99ced4922ee13562afde46c5cdf2a - vllm

Files

T

mistral0105 fd3e0cac12 Address review on Anthropic cache usage reporting

- api_router: stop silently overriding --enable-prompt-tokens-details for
  AnthropicServingMessages; pass through the user's CLI setting like the
  other serving objects.
- _compute_cache_usage: rewrite docstring to document where
  prompt_tokens_details is attached in vLLM's OpenAI streaming path
  (terminal include_usage chunk only), why message_start cannot populate
  cache fields today, and why cache_creation_input_tokens defaults to 0
  rather than None when cache info is present.
- AnthropicUsage construction: omit cache fields entirely when the
  underlying cache info is unknown (cache_read is None), rather than
  emitting null. Applied uniformly to non-streaming responses,
  message_start, and message_delta so "unknown" is signaled by key
  absence rather than null, distinguishing it from a real zero.
- Tests: add TestStreamingCacheUsageSemantics covering the three usage
  states (cache hit, cache miss with details, no details at all) for
  both message_start and message_delta.

Signed-off-by: mistral0105 <zhangshuoming17@mails.ucas.ac.cn>

2026-06-19 12:40:01 +00:00

anthropic

Address review on Anthropic cache usage reporting

2026-06-19 12:40:01 +00:00

generate

[Frontend] Resettle generative scoring entrypoint. (#44153 )

2026-06-01 07:54:59 +00:00

llm

[CI] Consolidate multimodal entrypoint tests. (#44819 )

2026-06-08 04:48:08 -07:00

multimodal

[CI] Consolidate multimodal entrypoint tests. (#44819 )