Downloads · 30 days
0
nutrivision-om-igsm
aman24012/nutrivision-om-igsm
nutrivision-om-igsm is a machine learning model from aman24012. Use it for the machine learning task on the model card, and read the license before you ship it in a product. It is set up for pytorch. The card lists the license as mit.
Trained weights for NutriVision — single-image food nutrition estimation with ingredient-aware hierarchical frequency fusion. From an RGB photo and an optional CLIP-encoded ingredient list, predicts 5 nutrients (calor…
Access
Public
Updated May 19, 2026
Repo size
829 MB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.pt829 MB · 100%
At a glance
- Library
- pytorch
- License
- mit
- Access
- Public
- Created
- May 19, 2026
- Updated
- May 19, 2026
- SHA
- 78e2479e
Spaces
Model card
From the Hugging Face model README
NutriVision · Om-IGSMNet checkpoint
Trained weights for NutriVision — single-image food nutrition estimation with ingredient-aware hierarchical frequency fusion. From an RGB photo and an optional CLIP-encoded ingredient list, predicts 5 nutrients (calories, mass, fat, carbs, protein).
- Architecture: Om-IGSMNet — dual ConvNeXt-Base + SSRA + ISM + IC-FAFM + IA-MPH
- Trained on: Nutrition5k (Google, CVPR 2021)
- Best result: 13.52 % mean PMAE on the 5:1 dish-level test split
How to load
from huggingface_hub import hf_hub_download
import torch
ckpt = hf_hub_download(repo_id="aman24012/nutrivision-om-igsm", filename="best_model.pt")
state = torch.load(ckpt, map_location="cpu", weights_only=False)
# state["model_state_dict"] is ready for OmIGSMNet.load_state_dict(...)
For a hosted UI: see the matching Space at aman24012/nutrivision.
Per-nutrient PMAE on Nutrition5k
| Cal | Mass | Fat | Carb | Prot | Mean ↓ |
|---|---|---|---|---|---|
| 12.65 | 10.37 | 16.45 | 14.76 | 13.35 | 13.52 |
License
MIT. Trained on Nutrition5k which has its own dataset licence — please respect that licence for derivative works.
- Library
- pytorch
- License
- mit
- Created
- May 19, 2026
- Updated
- May 19, 2026