github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-07 08:30:05 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #11653] Support dots.ocr (rednote-hilab) #69761

New Issue
Open
opened 2026-05-04 19:07:11 -05:00 by GiteaMirror · 4 comments
GiteaMirror commented 2026-05-04 19:07:11 -05:00
Owner
Copy Link

Originally created by @Leroy-X on GitHub (Aug 4, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/11653

dots.ocr is a powerful, multilingual document parser that unifies layout detection and content recognition within a single vision-language model while maintaining good reading order. Despite its compact 1.7B-parameter LLM foundation, it achieves state-of-the-art(SOTA) performance.

https://huggingface.co/rednote-hilab/dots.ocr
https://github.com/rednote-hilab/dots.ocr

Originally created by @Leroy-X on GitHub (Aug 4, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/11653 dots.ocr is a powerful, multilingual document parser that unifies layout detection and content recognition within a single vision-language model while maintaining good reading order. Despite its compact 1.7B-parameter LLM foundation, it achieves state-of-the-art(SOTA) performance. https://huggingface.co/rednote-hilab/dots.ocr https://github.com/rednote-hilab/dots.ocr
GiteaMirror added the model label 2026-05-04 19:07:11 -05:00
GiteaMirror commented 2026-05-04 19:07:23 -05:00
Author
Owner
Copy Link

@berserker1 commented on GitHub (Aug 6, 2025):

Currently the model runs only through vllm right?

<!-- gh-comment-id:3157431406 --> @berserker1 commented on GitHub (Aug 6, 2025): Currently the model runs only through vllm right?
GiteaMirror commented 2026-05-04 19:07:25 -05:00
Author
Owner
Copy Link

@LuSrackhall commented on GitHub (Aug 16, 2025):

Currently the model runs only through vllm right?

However, vllm requires too much VRAM, and many people don't have hardware capable of supporting it.

However, if not to use ollama; you can also use dots.ocr through Hugging Face inference.

After setting up the environment, you can use the following command to experience the parsing effect locally:

python3 demo/demo_hf.py

You can replace this path to parse the image you want to test.

Image

Finally, I also hope to be able to experience dots.ocr through Ollama.

<!-- gh-comment-id:3193833491 --> @LuSrackhall commented on GitHub (Aug 16, 2025): > Currently the model runs only through vllm right? However, vllm requires too much VRAM, and many people don't have hardware capable of supporting it. However, if not to use ollama; you can also use dots.ocr through Hugging Face inference. After setting up the environment, you can use the following command to experience the parsing effect locally: ```bash python3 demo/demo_hf.py ``` You can replace this path to parse the image you want to test. <img width="1288" height="469" alt="Image" src="https://github.com/user-attachments/assets/f2c49a26-d44f-46b4-81dd-c8af1422f2d2" /> Finally, I also hope to be able to experience dots.ocr through Ollama.
GiteaMirror commented 2026-05-04 19:07:25 -05:00
Author
Owner
Copy Link

@mkvn-1 commented on GitHub (Aug 17, 2025):

Currently the model runs only through vllm right?

However, vllm requires too much VRAM, and many people don't have hardware capable of supporting it.

However, if not to use ollama; you can also use dots.ocr through Hugging Face inference.

After setting up the environment, you can use the following command to experience the parsing effect locally:

python3 demo/demo_hf.py
You can replace this path to parse the image you want to test.

Image Finally, I also hope to be able to experience dots.ocr through Ollama.

Can you pls share a Jupyter notebook showing how to load this model with Transformers? That would be very useful for everyone.

<!-- gh-comment-id:3194402292 --> @mkvn-1 commented on GitHub (Aug 17, 2025): > > Currently the model runs only through vllm right? > > However, vllm requires too much VRAM, and many people don't have hardware capable of supporting it. > > However, if not to use ollama; you can also use dots.ocr through Hugging Face inference. > > After setting up the environment, you can use the following command to experience the parsing effect locally: > > python3 demo/demo_hf.py > You can replace this path to parse the image you want to test. > > <img alt="Image" width="1288" height="469" src="https://private-user-images.githubusercontent.com/142690689/478706750-f2c49a26-d44f-46b4-81dd-c8af1422f2d2.png?jwt=eyJ0eXAiOiJKV1QiLCJhbGciOiJIUzI1NiJ9.eyJpc3MiOiJnaXRodWIuY29tIiwiYXVkIjoicmF3LmdpdGh1YnVzZXJjb250ZW50LmNvbSIsImtleSI6ImtleTUiLCJleHAiOjE3NTU0MzkwNDEsIm5iZiI6MTc1NTQzODc0MSwicGF0aCI6Ii8xNDI2OTA2ODkvNDc4NzA2NzUwLWYyYzQ5YTI2LWQ0NGYtNDZiNC04MWRkLWM4YWYxNDIyZjJkMi5wbmc_WC1BbXotQWxnb3JpdGhtPUFXUzQtSE1BQy1TSEEyNTYmWC1BbXotQ3JlZGVudGlhbD1BS0lBVkNPRFlMU0E1M1BRSzRaQSUyRjIwMjUwODE3JTJGdXMtZWFzdC0xJTJGczMlMkZhd3M0X3JlcXVlc3QmWC1BbXotRGF0ZT0yMDI1MDgxN1QxMzUyMjFaJlgtQW16LUV4cGlyZXM9MzAwJlgtQW16LVNpZ25hdHVyZT0xYWMxNmMxODViMmZjZWExNmYzNzNkODhmODlmNzVlZGJjNWFmMzliNGE1OTJjZGMxNmVjYjliN2FhYTdhNTQyJlgtQW16LVNpZ25lZEhlYWRlcnM9aG9zdCJ9.mY44aMMyPHt2XxB_tmaRjjQfP388-Tzk602gPtNscCQ"> > Finally, I also hope to be able to experience dots.ocr through Ollama. Can you pls share a Jupyter notebook showing how to load this model with Transformers? That would be very useful for everyone.
GiteaMirror commented 2026-05-04 19:07:26 -05:00
Author
Owner
Copy Link

@LuSrackhall commented on GitHub (Aug 18, 2025):

Can you pls share a Jupyter notebook showing how to load this model with Transformers? That would be very useful for everyone.

@mkvn-1 Follow the steps in the official documentation to deploy successfully. My deployment process is relatively niche, but it also follows the official steps.

If you're interested—my deployment process uses Pixi to manage dependencies.

You can refer to my related repository DotsOCR for details. It's very useful when leveraging different cloud-native platforms for free. After cloning the relevant code, use the pixi i command to install dependencies. Then, follow the documentation tutorial to download model weights, and you can run it. I’ve updated the general process in the README of this repository.

Below is a screenshot of a successful deployment on the Tencent CNB platform:
Image

Image

I hope my deployment process can help you.

<!-- gh-comment-id:3195280141 --> @LuSrackhall commented on GitHub (Aug 18, 2025): > Can you pls share a Jupyter notebook showing how to load this model with Transformers? That would be very useful for everyone. @mkvn-1 Follow the steps in the official documentation to deploy successfully. My deployment process is relatively niche, but it also follows the official steps. If you're interested—my deployment process uses Pixi to manage dependencies. You can refer to my related repository [DotsOCR](https://github.com/LuSrackhall/DotsOcr) for details. It's very useful when leveraging different cloud-native platforms for free. After cloning the relevant code, use the `pixi i` command to install dependencies. Then, follow the documentation tutorial to download model weights, and you can run it. I’ve updated the general process in the README of this repository. Below is a screenshot of a successful deployment on the Tencent CNB platform: <img width="1907" height="1080" alt="Image" src="https://github.com/user-attachments/assets/0944f117-1264-4a92-9ff8-1a7d50eb637f" /> <img width="1038" height="859" alt="Image" src="https://github.com/user-attachments/assets/e6e1c2aa-262d-4601-95a6-399ef6245169" /> I hope my deployment process can help you.
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
amd
api
app
bug
build
cli
cloud
compatibility
context-length
create
docker
documentation
embeddings
feature request
feedback wanted
good first issue
gpt-oss
gpu
harmony
help wanted
image
install
intel
js
launch
linux
macos
memory
mlx
model
needs more info
networking
nvidia
ollama.com
performance
pull-request
Mirrored from GitHub Pull Request
 python
question
registry
rendering
thinking
tools
top
vulkan
windows
wsl
No Label model
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
GiteaMirror ninjasurge
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/ollama#69761
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?