daisybell
A scanner that will scan your AI models for problems. Currently it focuses on bias testing. It is currently alpha.
How to Use
First install it:
pip install daisybell
Run it in this manner (currently supports models from HuggingFace’s repository):
daisybell roberta-base
The scan can output files for further analysis:
daisybell roberta-base --output results/roberta-base
We will infer the task(s) of model by default but to provide specific tasks to test explicitly use the –task switch:
daisybell cross-encoder/nli-distilroberta-base --task zero-shot-classification
Daisybell can even scan chatbot models like StableLM for AI safety and alignment issues:
daisybell StabilityAI/stablelm-tuned-alpha-3b -d cuda:0
That’s it for now. More will come.
Future Work
More bias tests. More metrics for bias testing based on the research in the field.
Integration with other types of testing (eg. adversarial robustness)
More kinds of models besides HuggingFace models. We are especially interested in MLFlow integration.
Documentation.
Please contribute if you can. Help is always helpful.
License
Apache
Credit
A project of IQT Labs.
Metadata
Release files for daisybell 0.4.4
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 |
|---|---|---|---|---|
| daisybell-0.4.4.tar.gz | 15.2 kB | Without provenance | Details for daisybell-0.4.4.tar.gz |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| daisybell-0.4.4-py3-none-any.whl | Python 3 | none | any | Details for daisybell-0.4.4-py3-none-any.whl |
Total release size: 35.3 kB
Release files / daisybell-0.4.4.tar.gz
| Download URL | daisybell-0.4.4.tar.gz |
|---|---|
| Size | 15.2 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
df370407ed1b0b6488a8e96a47d42a9b53d42e5475a7689034b024de6cd50cb8
|
|
BLAKE2b-256 checksum How to use checksums |
ecd2f5115692550c9181656fba187441b70b63e770be58dd583f3db5fc2045c6
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
poetry/1.8.3 CPython/3.10.12 Linux/6.5.0-1022-azure
|
Release files / daisybell-0.4.4-py3-none-any.whl
| Download URL | daisybell-0.4.4-py3-none-any.whl |
|---|---|
| Size | 20.1 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
c81727543e84c2e2cc77cde22af343cea30f3ed89166d83567abcb63da13bb7a
|
|
BLAKE2b-256 checksum How to use checksums |
6c69338a31fc44dcd06817a99da4f57ac7150f160db677e7e820a54102b27535
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
poetry/1.8.3 CPython/3.10.12 Linux/6.5.0-1022-azure
|
Security for daisybell 0.4.4
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.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A 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.