AIGrammar Python package
Project description
AIGrammar
About
AIGrammar is all in one and easy to use package for model diagnostic and vulnerability checks. It enable with a simple line of code to check model metrics and prediction generalizability, feature contribution, and model vulnerability against adversarial attacks.
Data
- Multicollinearity
- Data drift
Model
- Metric metric comparison
- roc_auc vs average precision
- Optimal threshold vs 50% threshold
Feature importance
- Too high importance
- 0 impact
- Negative influence (FLOFO)
- Causes of overfitting
Adversarial Attack
- Model vulnerability identification based on one feature minimal change for getting opposite outcome.
Usage
Python 3.7+ required.
Installation: pip install AIGrammar
Find an example notebook here.
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
AIGrammar-0.0.4.tar.gz
(2.3 kB
view details)
File details
Details for the file AIGrammar-0.0.4.tar.gz.
File metadata
- Download URL: AIGrammar-0.0.4.tar.gz
- Upload date:
- Size: 2.3 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/3.7.1 importlib_metadata/4.10.1 pkginfo/1.8.2 requests/2.24.0 requests-toolbelt/0.9.1 tqdm/4.47.0 CPython/3.8.3
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 | cc90bf72a1000613d7a094d46cd4d9f1681462de748a91ba62aec2871b80fcf5 |
|
| MD5 | 112d92f5d044d05e466f6d75e5614c65 |
|
| BLAKE2b-256 | f9f23a0117fc843abc5e219839a649076fb5f60313848fff9ed6a846047cefc2 |
