deepseek-ai/DeepSeek-OCR Installation Guide

⚡

Automated Install (Recommended)

Quick installation instructions for deepseek-ai/DeepSeek-OCR. This is the fastest way to complete project installation and setup.

Install via curl

curl -fsSL https://hexmos.com/ipm-install | bash && 
ipm i deepseek-ai/DeepSeek-OCR

Install via PowerShell

iwr https://hexmos.com/ipm-install-ps -UseBasicParsing | iex; 
ipm i deepseek-ai/DeepSeek-OCR

Install via npx

npx @hexmos/ipm i deepseek-ai/DeepSeek-OCR

Prerequisites

Conda

package_manager

Required for environment management and python versioning.

CUDA

system_tool

Version: 11.8

NVIDIA GPU with CUDA 11.8 support is required for optimal performance.

Git

system_tool

Used to clone the repository.

Manual Installation Methods

Conda Environment Setup (vLLM & Transformers)

git clone https://github.com/deepseek-ai/DeepSeek-OCR.git

cd DeepSeek-OCR

conda create -n deepseek-ocr python=3.12.9 -y && source activate deepseek-ocr

pip install torch==2.6.0 torchvision==0.21.0 torchaudio==2.6.0 --index-url https://download.pytorch.org/whl/cu118

pip install https://github.com/vllm-project/vllm/releases/download/v0.8.5/vllm-0.8.5+cu118-cp312-cp312-manylinux1_x86_64.whl

pip install -r requirements.txt

pip install flash-attn==2.7.3 --no-build-isolation

Post Installation Steps

To run vLLM image inference: Navigate to `DeepSeek-OCR/DeepSeek-OCR-vllm` and run `python run_dpsk_ocr_image.py`.
To run Transformers inference: Navigate to `DeepSeek-OCR/DeepSeek-OCR-hf` and run `python run_dpsk_ocr.py`.
Ensure you have edited `config.py` in the vllm directory to set your input/output paths.
Download model weights from Hugging Face if you are not using the automated downloader in the scripts.

← Back to tool All Installation Guides