🤖 image-text-to-text

DeepSeek-OCR

deepseek-ai/DeepSeek-OCR

Get AI Model →
2.1M
Downloads
❤️
3.2K
Likes
🏷️
14
Tags
📦
transformers
Library
Model Details
Full Model IDdeepseek-ai/DeepSeek-OCR
Pipeline / Taskimage-text-to-text
Librarytransformers
Downloads (all-time)2.1M
Likes3.2K
Last Modified11/4/2025
Author / Orgdeepseek-ai
PrivateNo — public
⚡ Quick Usage (Python)

Using the 🤗 Transformers library. Install with pip install transformers

from transformers import pipeline

# Load the model
pipe = pipeline("image-text-to-text", model="deepseek-ai/DeepSeek-OCR")

# Run inference
result = pipe("Your input here")
print(result)
🏷️ Tags
transformerssafetensorsdeepseek_vl_v2feature-extractiondeepseekvision-languageocrcustom_codeimage-text-to-textmultilingualarxiv:2510.18234license:miteval-resultsregion:us
More image-text-to-text Models
See all →
Qwen3-VL-2B-Instruct

Qwen/Qwen3-VL-2B-Instruct

64.9M❤️ 368
Get AI Model →
Qwen2.5-VL-7B-Instruct

Qwen/Qwen2.5-VL-7B-Instruct

8.8M❤️ 1.5K
Get AI Model →
GLM-OCR

zai-org/GLM-OCR

7.3M❤️ 1.6K
Get AI Model →
🚀 Use This Model

Access model files, inference API, and full documentation on Hugging Face.

Open on Hugging Face →Browse Model Files ↗← Browse All Models
🤖 Task: image-text-to-text

This model is designed for the image-text-to-text task. Explore more models for this use case.

All image-text-to-text Models →
📊 Popularity
Downloads2.1M
❤️ Community Likes3.2K
🛠️ Requirements
  • Install: pip install transformers
  • Python 3.8+ recommended for Transformers.
  • GPU (CUDA) speeds up inference significantly.
  • Use model.half() for fp16 on limited VRAM.
👋 Need help with code?