deepseek-ai/DeepSeek-OCR

Add MDPBench evaluation results

#112 opened 5 days ago by

Delores-Lin

Переведи с сохранением структуры

#110 opened about 1 month ago by

Vl16161616

Add OlmOCRBench evaluation results

#109 opened about 2 months ago by

nielsr

update README to point to new version of this model

1

#108 opened 3 months ago by

googlefan

发票号码连续多个0的时候会识别异常

2

#107 opened 3 months ago by

Alooooo

ocr_test

#106 opened 4 months ago by

fengshengyu

deepseekocr

#105 opened 4 months ago by

sikejack

Is there a V2 in the works?

#104 opened 4 months ago by

athrael-soju

Can I extract the information from the invoice in JSON form using the DeepSeek OCR model?

1

#103 opened 4 months ago by

zabih1

DeepSeek OCR Triton Error [CUDA] an illegal memory access on vLLM 0.11.2

#102 opened 4 months ago by

blazarev

Deepseek-ocr cant read top right corner

👍 4

2

#101 opened 4 months ago by

Bunkid

How to use Large/Gundam mode with vLLM online serving?

1

#100 opened 5 months ago by

prudant

Request: DOI

1

#99 opened 5 months ago by

Amer9i

finetune

#98 opened 5 months ago by

Supperman007

RuntimeError: The expanded size of the tensor (753) must match the existing size (752) at non-singleton dimension 0. Target sizes: [753, 1280]. Tensor sizes: [752, 1]

1

#95 opened 5 months ago by

deathknight0

Update README.md

#94 opened 5 months ago by

mrpathakji

www

#93 opened 5 months ago by

Matolinjo

How do I view the visualization results recognized by my DeepSeekOCR? Why is the image I saved the same as the original one?

#92 opened 5 months ago by

Jiahua0

How can I disable streaming during inference? Also, can Deepseek OCR handle multiple images at the same time?

1

#91 opened 5 months ago by

Tizzzzy

Document support

1

#90 opened 5 months ago by

wolffan

[Issue] Repetitive text with endless loop

➕ 5

13

#89 opened 5 months ago by

abhi22

Create Matthew Koschel The Riemann hypothesis

2

#88 opened 5 months ago by

Mattkos

OCR Request fails the first time, works second time

#87 opened 5 months ago by

naamanm

requirements.txt

#86 opened 5 months ago by

Hifzamudassar

app.py

#85 opened 5 months ago by

Hifzamudassar

【问题反馈】GaoDa-M 模式下动态预处理未正确使用 1024×1024，仍固定为 640×640 问题描述

#84 opened 5 months ago by

liushihowe

Lightweight FastAPI Service for DeepseekOCR + docker-compose

#83 opened 5 months ago by

athrael-soju

🔥🔥🔥微调DeepSeek-OCR视频教程

#82 opened 5 months ago by

leo009

GGUF Available in NexaSDK (only). Here's how to run it locally.

🔥 3

1

#81 opened 5 months ago by

alanzhuly

Error while running this model in colab

3

#80 opened 5 months ago by

dhirajsuvarna

gopu

#78 opened 5 months ago by

Mauricio-100

Request: DOI

#77 opened 5 months ago by

S18069O8C2910251200058

Here is a way to self host this on AWS

#76 opened 5 months ago by

mwalker714

If possible, could you inference include words with strikethrough formatting where applicable?

#74 opened 5 months ago by

willy1212009

elektraspurio98@gmail.com

#73 opened 5 months ago by

josiah91-me-home

CPU Inferencing

1

#72 opened 6 months ago by

chiggidev

Compare images based on scoring

🔥 1

1

#70 opened 6 months ago by

schirrmacher

لخص

1

#67 opened 6 months ago by

janaismail

pull

#66 opened 6 months ago by

LYChuad

Update README.md

1

#65 opened 6 months ago by

hader1890

model.infer output

🔥❤️ 4

#62 opened 6 months ago by

Theoreticallyhugo

This is what ai needed

2

#61 opened 6 months ago by

kingriel

Update README.md

#60 opened 6 months ago by

DDDDARAT

image_generation

#59 opened 6 months ago by

Thinks123

Bboxes

1

#58 opened 6 months ago by

mubashiross

finetune ?

1

#57 opened 6 months ago by

winzzniw555

how to use deep-seek-ocr model

1

#56 opened 6 months ago by

Zhaleh197

How to use it correctly with online serving via vllm openai compatible server?

12

#55 opened 6 months ago by

dhruvil237

Remove hardcoded .cuda() calls to support single forward pass on CPU and ensure DeepSeekOCR model compatibility with transformers==4.52.4

🚀😎 2

4

#54 opened 6 months ago by

kamalrajkannanmcw

🧠 Why does DeepSeek-OCR not use Multi-Head Latent Attention (MLA)?

6

#53 opened 6 months ago by

ZoneTwelve

Add MDPBench evaluation results

Переведи с сохранением структуры

Add OlmOCRBench evaluation results

update README to point to new version of this model

发票号码连续多个0的时候 会识别异常

ocr_test

deepseekocr

Is there a V2 in the works?

Can I extract the information from the invoice in JSON form using the DeepSeek OCR model?

DeepSeek OCR Triton Error [CUDA] an illegal memory access on vLLM 0.11.2

Deepseek-ocr cant read top right corner

How to use Large/Gundam mode with vLLM online serving?

Request: DOI

finetune

RuntimeError: The expanded size of the tensor (753) must match the existing size (752) at non-singleton dimension 0. Target sizes: [753, 1280]. Tensor sizes: [752, 1]

Update README.md

www

How do I view the visualization results recognized by my DeepSeekOCR? Why is the image I saved the same as the original one?

How can I disable streaming during inference? Also, can Deepseek OCR handle multiple images at the same time?

Document support

[Issue] Repetitive text with endless loop

Create Matthew Koschel The Riemann hypothesis

OCR Request fails the first time, works second time

requirements.txt

app.py

【问题反馈】GaoDa-M 模式下动态预处理未正确使用 1024×1024，仍固定为 640×640 问题描述

Lightweight FastAPI Service for DeepseekOCR + docker-compose

🔥🔥🔥微调DeepSeek-OCR视频教程

GGUF Available in NexaSDK (only). Here's how to run it locally.

Error while running this model in colab

gopu

Request: DOI

Here is a way to self host this on AWS

If possible, could you inference include words with strikethrough formatting where applicable?

elektraspurio98@gmail.com

CPU Inferencing

Compare images based on scoring

لخص

pull

Update README.md

model.infer output

This is what ai needed

Update README.md

image_generation

Bboxes

finetune ?

how to use deep-seek-ocr model

How to use it correctly with online serving via vllm openai compatible server?

Remove hardcoded .cuda() calls to support single forward pass on CPU and ensure DeepSeekOCR model compatibility with transformers==4.52.4

🧠 Why does DeepSeek-OCR not use Multi-Head Latent Attention (MLA)?

发票号码连续多个0的时候会识别异常