From 0b7dfdcb902b40a3a0f6048ccce2a7de3c38a425 Mon Sep 17 00:00:00 2001 From: majiayu000 <1835304752@qq.com> Date: Fri, 2 Jan 2026 06:56:08 +0800 Subject: [PATCH] fix: add thinking field to model_config.yaml examples Add thinking switch configuration for Docker and Helm deployed services: - thinking: boolean to enable thinking mode - thinking_budget_tokens: max tokens for thinking output Signed-off-by: majiayu000 <1835304752@qq.com> --- .../deployment/docker-compose/conf/model_config.yaml | 12 ++++++++++++ .../helm-chart/umbrella/conf/model_config.yaml | 12 ++++++++++++ 2 files changed, 24 insertions(+) diff --git a/release/deployment/docker-compose/conf/model_config.yaml b/release/deployment/docker-compose/conf/model_config.yaml index 2aab6ca79..018424dec 100644 --- a/release/deployment/docker-compose/conf/model_config.yaml +++ b/release/deployment/docker-compose/conf/model_config.yaml @@ -29,3 +29,15 @@ models: min: "0.001" max: "1.0" default_val: "0.7" + - name: "thinking" + label: "thinking" + desc: "Enable thinking mode for reasoning models that support extended thinking capabilities." + type: "boolean" + default_val: "false" + - name: "thinking_budget_tokens" + label: "thinking_budget_tokens" + desc: "Maximum number of tokens for thinking content output when thinking mode is enabled." + type: "int" + min: "1024" + max: "65536" + default_val: "4096" diff --git a/release/deployment/helm-chart/umbrella/conf/model_config.yaml b/release/deployment/helm-chart/umbrella/conf/model_config.yaml index 2aab6ca79..018424dec 100644 --- a/release/deployment/helm-chart/umbrella/conf/model_config.yaml +++ b/release/deployment/helm-chart/umbrella/conf/model_config.yaml @@ -29,3 +29,15 @@ models: min: "0.001" max: "1.0" default_val: "0.7" + - name: "thinking" + label: "thinking" + desc: "Enable thinking mode for reasoning models that support extended thinking capabilities." + type: "boolean" + default_val: "false" + - name: "thinking_budget_tokens" + label: "thinking_budget_tokens" + desc: "Maximum number of tokens for thinking content output when thinking mode is enabled." + type: "int" + min: "1024" + max: "65536" + default_val: "4096"