Repository navigation
Conversation
…anxianglei#2666) A summary target that thinks by default spends its output budget on reasoning: finish reason "length", empty content, and every fold fails. summaryPayload() sends only model / max_tokens / messages / stream and no config could add a reasoning control (PR ranxianglei#2806: "it needs a typed reasoning knob"). Recipe models (config-file named recipes and agent-registry recipes) take `reasoning`: - "off": the protocol's own control -- openai reasoning_effort "none", responses reasoning.effort "none", anthropic thinking.type "disabled", google generationConfig.thinkingConfig.thinkingBudget 0; - "template-off": chat_template_kwargs.enable_thinking false, openai api only, for chat-template servers (vLLM, SGLang, llama.cpp + Qwen). Validated at recipe parse, chain expansion and on the request rail; applySummaryReasoning() writes it after the output budget. Unset leaves the body byte-identical. Opt-in because strict upstreams reject unknown fields. Closes ranxianglei#2666
(ai: vllm-qwen/qwen3.8-27b) 查重: 不重复。 This is the concrete implementation of the knob requested in #2666 (your floor-5/6 comments there ask for exactly this shape, and the earlier triage there recommended a typed per-model knob over free-form
标签: pr · scope/wiring(摘要请求管道 + 配置解析面)· p2(能用但思考型目标上摘要基本失效,有绕行——换非思考模型)。功能请求不打 bug。 Full review (direction vs #2666, regression surface, code/tests, diff cleanliness) is in progress — I'll post the structured verdict next. Note the PR body asks for owner sign-off on the config surface; my review will state whether the proposed shape holds up against the existing providers/recipe system. |
(ai: vllm-qwen/qwen3.8-27b) PR #4463 审核结论(独立审读 diff + master 代码 + 本地定向验证) 1. 是什么问题:外部摘要链(summary lane)发出的请求体里没有任何思考控制字段。目标模型默认会思考时(Qwen/DeepSeek/GLM thinking 等),整个输出预算被 reasoning 吃掉—— 2. 什么原因导致的:能力缺失,不是回归。master@e518363(v0.1.205)上 3. 影响范围与严重程度:仅配置了 external summary 且目标模型默认思考的用户;该会话折叠整体不可用(主模型流量不受影响)。有绕行(换非思考模型、调大 outputTokens 碰运气)→ p2,标签已打。 4. 怎么修复的:新增 opt-in 类型化旋钮 5. 是否有回归问题:逐项清点过,无回归——
6. 是否破坏缓存:不涉及(不碰前缀缓存、会话存储、reconcile)。 7. 是否改了设计方向:不涉及(纯增量配置面,沿用既有 recipe model 字段模式)。 8. 是否引入了新的配置:是——recipe model 条目(providers 表配方 + agent-registry + request-rail targets)新增 验证证据(本地定向,全量判定以 CI 为准):PR head 上 判定:代码侧审核通过,可合入条件 = owner 对 config surface 拍板 + CI 转绿。 按 owner-gated 规则这是新配置字段,需要你一句「批准」;作者已在 #2666 楼层 6 明确请求签核。不合 auto-merge 门槛(new capability surface + config schema),人工合并。未发现需要顺手修的小问题。 |
|
(ai: zhipuai-lb/glm-5.3) 陈旧票清理(第三批动作①):目标 issue #2666 已关。为避免重复劳动关闭本 PR;如判断有误请直接 reopen 并留言。 感谢 @naerymdan 的贡献与排查——这条修复路径已被其他方式覆盖;欢迎认领其他 open issue。 |
Closes #2666
What
A summary target that thinks by default spends its output budget on reasoning: the reply has finish reason
lengthand empty content, and every fold fails.summaryPayload()sends onlymodel,max_tokens,messagesandstream, so nothing in config could turn thinking off.Recipe models now accept
reasoning:"off"reasoning_effort: "none"· responsesreasoning: {effort: "none"}· anthropicthinking: {type: "disabled"}· googlegenerationConfig.thinkingConfig: {thinkingBudget: 0}"template-off"chat_template_kwargs: {enable_thinking: false}(existing keys merged),api: "openai"only. This is for servers that render a chat template (vLLM, SGLang, llama.cpp serving Qwen-family models), which ignorereasoning_effort.parseNamedProviderRecipe), the agent-registry parser (parseAgentProviderReport), and the request rail (parseExternalSummarySettings), via the sharedparseSummaryReasoning.applySummaryReasoning()writes it after the output budget.Config surface
providers.<name>.models.<model>.reasoning: "off" | "template-off"on named recipes, the same per-model level as the existingcontextWindow/outputTokens/stream. Also accepted on agent-registry recipes and on the expanded rail target, since both carry the same per-target knobs.compat.dropFieldscan only delete fields, not add them.compress.reasoninggoverns dropping finished reasoning from history, a different concern.pi.subagents.thinkingLevelonly covers the sub-agent lane.compress.externalSummaryblock is whole-chain-replace, so a chain-level knob would hit every target and every protocol at once.extraBody. An arbitrary body passthrough would be a much wider surface, and harder to validate.reasoningon an older bili fails that recipe's parse ("unknown fields on recipe model"), andloadNamedProvidersalready reports that loudly. The launcher↔extension channel is unchanged: pi doesn't sendreasoningin its provider report yet (possible follow-up).Measured
From my #2666 comment: summaries on a local gufo 0.22.0 / Qwen3.8 Flash-Next.
outputTokens256: 92 of 92 summary calls endedfinish_reason=lengthin reasoning, and 0 folds landed.outputTokens2048: 10 of 15 still endedlength, at ~40 s each.On the same server,
reasoning_effort: "none",chat_template_kwargs: {enable_thinking: false}andthinking: {type: "disabled"}each return a direct answer with 0 reasoning characters, against 179 without.Tests
tests/issue2666-summary-reasoning.test.ts: 15 cases."off"writes the native field into the body the upstream actually receives, and no knob means no thinking field."template-off"setschat_template_kwargsand notreasoning_effort, merging existing keys.template-offon non-openai apis.npm run typecheck,npm run buildandcheck:unused-exportsare clean.e2e/e2e-native-pi,issue3157-persist-durabilityandlauncher-plugin-modefail; they fail the same way on unmodifiedmasterthere. That's environment-only, so the full-suite verdict is CI's.kernel/**change.