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
Docs: update small models Torch backend requirements to exclude XPU
myhloli85632fbc689c
diff --git a/docs/en/usage/quick_usage.md b/docs/en/usage/quick_usage.mdindex 267e2410..9601bb07 100644--- a/docs/en/usage/quick_usage.md+++ b/docs/en/usage/quick_usage.md@@ -22,7 +22,7 @@ mineru parse <input_path> --pages all -o <output_path>> [!NOTE]> Runtime acceleration is selected separately for the two model components, based on installed dependencies and detected devices:>-> - Small models use the Torch backend only when `torch`, `torchvision`, `transformers`, `accelerate`, and `safetensors` are **all** installed and a non-CPU device (CUDA/XPU/MPS/...) is detected; otherwise they run ONNX on CPU. Installing Torch alone is not enough.+> - Small models use the Torch backend only when `torch`, `torchvision`, `transformers`, `accelerate`, and `safetensors` are **all** installed and a non-CPU device (CUDA/MPS/...) is detected; otherwise they run ONNX on CPU. Installing Torch alone is not enough.> - The local VLM engine is chosen independently: macOS always uses llama.cpp; on an accelerator device Linux prefers vLLM, then an installed LMDeploy, and Windows uses LMDeploy; otherwise llama.cpp.> - XPU is excluded from automatic LMDeploy selection: Linux uses an installed XPU-compatible vLLM, otherwise llama.cpp; Windows uses llama.cpp.> - Windows users who need CUDA acceleration should first visit the [PyTorch website](https://pytorch.org/get-started/locally/) and install accelerator-enabled `torch` and `torchvision` matching their CUDA version, then install the `mineru[full]` extras.diff --git a/docs/zh/usage/quick_usage.md b/docs/zh/usage/quick_usage.mdindex 65b3c95a..374c8429 100644--- a/docs/zh/usage/quick_usage.md+++ b/docs/zh/usage/quick_usage.md@@ -22,7 +22,7 @@ mineru parse <input_path> --pages all -o <output_path>> [!NOTE]> 运行时加速按两个模型组件分别选择,依据是已安装的依赖和检测到的设备:>-> - 小模型仅在 `torch`、`torchvision`、`transformers`、`accelerate`、`safetensors` **全部安装**且检测到非 CPU 设备(CUDA/XPU/MPS 等)时使用 Torch 后端,否则使用 ONNX(CPU)。只安装 Torch 并不足以启用。+> - 小模型仅在 `torch`、`torchvision`、`transformers`、`accelerate`、`safetensors` **全部安装**且检测到非 CPU 设备(CUDA/MPS 等)时使用 Torch 后端,否则使用 ONNX(CPU)。只安装 Torch 并不足以启用。> - 本地 VLM 引擎独立选择:macOS 固定使用 llama.cpp;加速卡设备上 Linux 优先 vLLM、其次已安装的 LMDeploy,Windows 使用 LMDeploy;否则使用 llama.cpp。> - XPU 不会自动选择 LMDeploy:Linux 上安装了支持 XPU 的 vLLM 时使用 vLLM,否则使用 llama.cpp;Windows 上使用 llama.cpp。> - Windows 用户如需 CUDA 加速,请先前往 [PyTorch 官网](https://pytorch.org/get-started/locally/) 选择与 CUDA 版本匹配的命令安装支持加速的 `torch` 和 `torchvision`,再安装 `mineru[full]`。