How to Upload a YOLO Model to Arkindex
This guide walks you through uploading a YOLO model to Arkindex using the Arkindex CLI, using the Institutional Newspapers Segmenter YOLO26x model from Hugging Face as an example.
1. Install and authenticate the Arkindex CLI
Install the CLI via pip and log in to your Arkindex instance:
pip install arkindex-cli
arkindex login
Follow the prompts to enter your Arkindex instance URL and credentials or API token.
2. Create a model directory
mkdir institutional-newspapers-segmenter
3. Download the YOLO weights
Download the model weights from Hugging Face and save them as model.pt inside your model directory (Arkindex requires the weight file to be named model.pt):
curl -L https://huggingface.co/institutional/institutional-newspapers-segmenter-yolo26x/resolve/main/best.pt > institutional-newspapers-segmenter/model.pt
4. Create the Arkindex configuration file
Create a .arkindex.yml file in the parent directory (the directory containing institutional-newspapers-segmenter, not inside the folder itself):
version: 2
models:
- path: institutional-newspapers-segmenter
name: Institutional Newspapers Segmenter YOLO26x
(You can change the name field to any preferred display name.)
5. Upload the model to Arkindex
Make sure you are in the same directory containing .arkindex.yml, then run:
arkindex models publish
6. Configure the model in the Arkindex UI
Once the model is uploaded, link it to a compatible worker via the web interface:
- In Arkindex, open your user menu in the upper-right corner and select Models.
- Search for the model Institutional Newspapers Segmenter YOLO26x and select it.
- Open the Compatible Workers tab.
- Add Yolo Inference - Image Segmenter as a compatible worker.
The YOLO model is now uploaded and ready for use in Arkindex.