This release has been yanked by its maintainers, and will be ignored by installers, except when explicitly specified.
Consider using release 0.2.1 instead.
Reason given by maintainers: Missing bundled pretrained model resources; use 0.2.1.
OmniMIRA
An anatomically indexed foundation model across structural, molecular and metabolic neuroimaging
OmniMIRA learns brain-region representations from T1-weighted MRI, amyloid PET, FDG PET and CT. It combines local image features within atlas-defined regions, allowing the same embeddings to be used for downstream prediction and regional analysis.
Release status
Version 0.2 provides the OmniMIRA v9 inference architecture, its multi-atlas resources and an epoch-1000 checkpoint. It produces anatomically indexed features and is intended for research use, not clinical diagnosis.
Installation
Requires Python 3.9 or later and PyTorch 2.0 or later.
pip install omnimira
Usage
The wheel contains the checkpoint, model-space templates and atlas definitions. Input NIfTI images must already be spatially normalized to MNI space. The package resamples them to the committed model grid and applies modality-specific intensity normalization.
from omnimira import from_pretrained
model = from_pretrained(device="cpu")
features = model.extract("scan_mni.nii.gz", modality="t1")
print(features["aal3"].shape) # (166, 128)
The command-line interface writes an NPZ with ROI identifiers, names and features:
omnimira scan_mni.nii.gz features.npz --modality t1
The output contains 128-dimensional embeddings for each region: 166 AAL3 regions, 69 Harvard–Oxford regions and 7 Yeo networks.
The modality options are t1 (T1-weighted MRI), av45 (amyloid PET), fdg (FDG PET) and ct (CT).
Documentation
Tests
Model and loss tests use synthetic inputs:
pip install pytest
python -m pytest -q tests/test_model_forward.py tests/test_losses.py
The full test suite also requires the atlas and template files.
Questions
For questions about the code, please open an issue.
License
Source code is available under the Apache 2.0 license. The bundled pretrained model weights are licensed separately under CC BY-NC 4.0, which does not permit commercial use. Third-party resources have separate terms.
For research use only; not for clinical diagnosis.
1 maintainer
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Filter files by name, interpreter, ABI, and platform.
If you're not sure about the file name format, learn more about wheel file names.
Copy a direct link to the current filters
File details
Details for the file omnimira-0.2.0.tar.gz.
File metadata
- Download URL: omnimira-0.2.0.tar.gz
- Upload date:
- Size: 51.3 kB
- Tags: Source
- Uploaded using Trusted Publishing? Yes
- Uploaded via:
twine/7.0.0 CPython/3.13.14
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
204187ee52f1ca5581d5bb39ef7d39451a26299bd854617dc2d8767b0096e829
|
|
| MD5 |
74d47f09e72ce31e89b937630bb9c508
|
|
| BLAKE2b-256 |
b65b5521f7e3d89f9802b6ca3e191ffbc7738d6d84be03e179f4788673ca35a5
|
Provenance
The following attestation bundles were made for omnimira-0.2.0.tar.gz:
Publisher:
publish-pypi.yml on YixinChen-AI/MIRA-Brain-FM
-
Statement:
-
Statement type:
https://in-toto.io/Statement/v1 -
Predicate type:
https://docs.pypi.org/attestations/publish/v1 -
Subject name:
omnimira-0.2.0.tar.gz -
Subject digest:
204187ee52f1ca5581d5bb39ef7d39451a26299bd854617dc2d8767b0096e829 - Sigstore transparency entry: 2802801578
- Sigstore integration time:
-
Permalink:
YixinChen-AI/MIRA-Brain-FM@2c09b5248d836e17e959caa93756adefb1c0c70b -
Branch / Tag:
refs/tags/v0.2.0 - Owner: https://github.com/YixinChen-AI
-
Access:
public
-
Token Issuer:
https://token.actions.githubusercontent.com -
Runner Environment:
github-hosted -
Publication workflow:
publish-pypi.yml@2c09b5248d836e17e959caa93756adefb1c0c70b -
Trigger Event:
push
-
Statement type:
File details
Details for the file omnimira-0.2.0-py3-none-any.whl.
File metadata
- Download URL: omnimira-0.2.0-py3-none-any.whl
- Upload date:
- Size: 42.2 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? Yes
- Uploaded via:
twine/7.0.0 CPython/3.13.14
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
42ebd5660c9f51871049e69e37a667b2a526e1231e414bca9f3fc9e641e9dc3f
|
|
| MD5 |
4f5f42e7f79cdba52110abcfe5536052
|
|
| BLAKE2b-256 |
ddc99202d6ac70bace6c68727134f6913a8f382118ab879a4f48f0f5937d7dae
|
Provenance
The following attestation bundles were made for omnimira-0.2.0-py3-none-any.whl:
Publisher:
publish-pypi.yml on YixinChen-AI/MIRA-Brain-FM
-
Statement:
-
Statement type:
https://in-toto.io/Statement/v1 -
Predicate type:
https://docs.pypi.org/attestations/publish/v1 -
Subject name:
omnimira-0.2.0-py3-none-any.whl -
Subject digest:
42ebd5660c9f51871049e69e37a667b2a526e1231e414bca9f3fc9e641e9dc3f - Sigstore transparency entry: 2802801631
- Sigstore integration time:
-
Permalink:
YixinChen-AI/MIRA-Brain-FM@2c09b5248d836e17e959caa93756adefb1c0c70b -
Branch / Tag:
refs/tags/v0.2.0 - Owner: https://github.com/YixinChen-AI
-
Access:
public
-
Token Issuer:
https://token.actions.githubusercontent.com -
Runner Environment:
github-hosted -
Publication workflow:
publish-pypi.yml@2c09b5248d836e17e959caa93756adefb1c0c70b -
Trigger Event:
push
-
Statement type:
