rai-check-dl
Deep learning audits for image classification, medical imaging, and scientific AI.
Features
- Image classification accuracy and per-class recall checks
- Robustness checks for recorded or callback-generated transformation predictions
- Built-in brightness, contrast, horizontal-flip, and Gaussian-noise transformations
- Grad-CAM heatmaps with PyTorch hooks or TensorFlow
GradientTape - Medical imaging checks for patient leakage across splits and site-level accuracy bias
- Segmentation Dice, IoU, mask-quality, and subgroup checks
- Object-detection mAP, IoU-threshold, class-coverage, and subgroup checks
- DICOM metadata, scanner/protocol bias, near-duplicate leakage, calibration, and patient aggregation
- Grad-CAM transformation stability and optional localization-overlap checks
Python API
from rai_audit.dl import ImageClassificationAudit
report = ImageClassificationAudit(
y_true=y_true,
y_pred=y_pred,
transformed_predictions={"sensor_noise": noisy_predictions},
persist=False,
).run()
To evaluate transformations directly, pass images and a predictor callback.
CLI
Audit recorded predictions from CSV:
rai-check dl run --data predictions.csv --task image --format html
Medical imaging CSV audits can add --patient-id, --split, and --site.
Examples
Metadata
Release files for rai-check-dl 0.0.1
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | Actions |
|---|---|---|---|---|
| rai_check_dl-0.0.1.tar.gz | 18.9 kB | With provenance | Details for rai_check_dl-0.0.1.tar.gz |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| rai_check_dl-0.0.1-py3-none-any.whl | Python 3 | none | any | Details for rai_check_dl-0.0.1-py3-none-any.whl |
Total release size: 41.7 kB
Release files / rai_check_dl-0.0.1.tar.gz
| Download URL | rai_check_dl-0.0.1.tar.gz |
|---|---|
| Size | 18.9 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
da81a189cfe16903a36e3da938793148e89c9f33ecfdb0ce2c80871ef2fdd1df
|
|
BLAKE2b-256 checksum How to use checksums |
95202058999ecb08c604a847e880f691fa71ce7a45e4a493670837b53e683bf3
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.14
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Sep 21, 2026.
Transparency logRelease files / rai_check_dl-0.0.1-py3-none-any.whl
| Download URL | rai_check_dl-0.0.1-py3-none-any.whl |
|---|---|
| Size | 22.8 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
cfc9555117b9bbf288f5719208ab7b522ee7a477b46edaaeeaa6432df15bca9d
|
|
BLAKE2b-256 checksum How to use checksums |
082acc336fed90a382ce2418bdbc56313759295bfe94261398f91db65d332c55
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.14
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Sep 21, 2026.
Transparency logSecurity for rai-check-dl 0.0.1
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
Full provenance
All 2 files in this release have been uploaded with provenance metadata.
Attestation for rai_check_dl-0.0.1.tar.gz
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Sep 21, 2026.
Transparency logA note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.