Transforms complex documents like PDFs and Office docs into LLM-ready markdown/JSON for your Agentic workflows.
Python ai4sciencedocument-analysisdocxextract-datalayout-analysisocrparserpdfpdf-converterpdf-extractor-llmpdf-extractor-pretrainpdf-extractor-ragpdf-parserpptxpythonxlsx
feat: update engine options to include llama-cpp and adjust documentation
myhlolibab3dc92aae3
diff --git a/docs/next/cli/mineru-kit-vlm-server.md b/docs/next/cli/mineru-kit-vlm-server.mdindex d3d69935..140dd990 100644--- a/docs/next/cli/mineru-kit-vlm-server.md+++ b/docs/next/cli/mineru-kit-vlm-server.md@@ -6,8 +6,7 @@非目标: 完整 Parse API;通用聊天能力说明;底层 engine 私有参数表来源: 由根目录旧 CLI 底稿迁移整理而来-自动服务引擎仅选择 vLLM/LMDeploy;MLX 必须显式使用 `--engine mlx`。-本命令不提供 llama HTTP server;ARM Mac 的默认文档解析在进程内使用 llama。+服务启动入口保留 vLLM、LMDeploy、llama.cpp 和 MLX;MLX 必须显式选择。## 1. 定位@@ -76,10 +75,11 @@ API 服务通过 `--vlm-server-url` 或全局 `model.vlm.server_url` 连接它mineru-kit vlm-server --engine automineru-kit vlm-server --engine vllmmineru-kit vlm-server --engine lmdeploy+mineru-kit vlm-server --engine llama-cppmineru-kit vlm-server --engine mlx```-`--engine` 决定使用哪类底层 serving engine。当前合法值为 `auto`、`vllm`、`lmdeploy`、`mlx`;`auto` 按 vLLM、LMDeploy 的顺序选择已安装且可用的 engine;MLX 仅接受显式选择。+`--engine` 决定使用哪类底层 serving engine。当前合法值为 `auto`、`vllm`、`lmdeploy`、`llama-cpp`、`mlx`;`auto` 按 vLLM、LMDeploy、llama.cpp 的顺序选择已安装且可用的 engine;MLX 仅接受显式选择。除统一参数外,其余参数原样透传到底层 engine server。diff --git a/pyproject.toml b/pyproject.tomlindex 8c58a449..bdb7dbd3 100644--- a/pyproject.toml+++ b/pyproject.toml@@ -41,7 +41,7 @@ dependencies = ["docvortex>=0.6.0,<0.7","gradio>=6.21,<7","mineru[torch] ; sys_platform == 'darwin' and platform_machine == 'arm64'",- "mineru-vl-utils>=2.0.7,<3",+ "mineru-vl-utils>=2.0.8,<3","mineru-llama-cpp>=0.1.3,<1","click>=8.1.7","loguru>=0.7.2",