jonathanjordan21/donut_fine_tuning_food_composition_id
Individual publisher
Nutrition panels on packaging have to be retyped before anyone can use the numbers. These are downloadable model weights, with code, that read food composition details straight from document images and answer questions about them.
QUICK INFO
SOLUTION TYPE
Models, data & tools
TECHNOLOGY CATEGORY
Models & fine-tunes
AREA OF WORK
Food safety, quality & nutrition
ACCESS
Free · Open source
DEPLOYMENT
Self-hosted / on-premise · Downloadable
SOURCE REVIEWED
August 17, 2026
What it helps with
- Read a nutrition panel from a photo
- Stop retyping numbers off food packaging
- Pull food composition data from label images
- Ask questions about a scanned nutrition table
- Digitise nutrition panels in bulk
- Build a label reading tool
Scope & audience
WHO IT'S FOR
Developer / AI builder
AREA OF WORK
Food safety, quality & nutrition
COMMODITY / SCOPE
Cross-commodity
Access & technical details
ACCESS
Anyone can get it
INTERFACE
Model download / weights · Code repository
DEPLOYMENT
You run it yourself
LANGUAGES
Not stated by provider
Featured media
Source information
This listing was compiled from publicly available source material. Verify current details with the provider.
PROVIDER DOCUMENTATION
Reviewed August 2026
Explore more like this
Llama 3.3 70B Instruct
Meta
Models, data & tools
Teams building farm advice tools often start with an openly available general model they can host. These are downloadable model weights, also served through programming interfaces, for a general purpose language model tested on agricultural advisory questions.
Free
Cloud-hosted · Downloadable
Models & fine-tunes
Llama 3.3 70B Instruct
advise-educate
extension-advisory-and-education
buildable-or-downloadable-component
models-and-fine-tunes
Free
Cloud-hosted · Downloadable
August 17, 2026
- Open model to host for farm advice
- Start a farm advice tool on open weights
- General model tested on agricultural questions
- Run a farm assistant on our own hardware
- Call an open general model from an app
- Compare open models on farm advice
SmolVLM2
Hugging Face
Models, data & tools
A small open vision-language model family for image, multi-image and video understanding in 256M, 500M and 2.2B sizes. Hugging Face reported an internal local iPhone demonstration with the 500M model. The downloadable model is current, but this listing does not represent a publicly obtainable iPhone application.
Free · Open source
Self-hosted / on-premise · Downloadable
Models & fine-tunes
SmolVLM2
build-integrate-ai
agrifood-data-and-ai-infrastructure
buildable-or-downloadable-component
models-and-fine-tunes
Free · Open source
Self-hosted / on-premise · Downloadable
August 26, 2026
- Analyze short field video on a local device
- Describe crop or equipment images offline
- Build camera-based visual question answering
- Fine-tune a small VLM for scouting imagery
- Search video clips by natural language
- Prototype edge video understanding on Apple hardware
Is this your Resource?
Contact us to claim this listing and add details, screenshots or a demo.