This repository contains the Python client library of Epoch AI. At the moment, only one feature is supported: reading from our database of ML models and benchmark results.
Installation
pip install epochai
Usage
Reading from our Airtable database of ML models and benchmark results
A few preparatory steps are required:
- Open our Airtable base
- Airtable doesn't allow public API access, so you'll have to make a copy of the base (unless you are an Epoch AI team member).
- Define the
AIRTABLE_BASE_IDenvironment variable with the ID of the base you just copied. (The ID is in the URL and starts withapp.) - Create an Airtable API key with access to the base, and the following scopes:
data.records:read,schema.bases:read. Define theAIRTABLE_API_KEYenvironment variable with the key.
You're now ready to use the library. The database models are defined in epochai.airtable.models.
You can get started with our example script examples/airtable.py, or try the snippets below.
from epochai.airtable.models import MLModel, Task, Score, Organization, BenchmarkRun
# Get everything at the start to minimize API calls
scores = Score.all(memoize=True)
runs = BenchmarkRun.all(memoize=True)
models = MLModel.all(memoize=True)
tasks = Task.all(memoize=True)
organizations = Organization.all(memoize=True)
Print information about a model:
print_model_info("claude-3-5-sonnet-20240620")
Print the highest scores for a benchmark and scorer:
print_high_scores(
task_path="bench.task.hendrycks_math.hendrycks_math_lvl_5",
scorer="model_graded_equiv",
scores=scores
)
Track the best-performing model to date over time:
print_performance_timeline(
task_path="bench.task.gpqa.gpqa_diamond",
scorer="choice",
scores=scores
)
Metadata
Release files for epochai 0.1.3
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 |
|---|---|---|---|---|
| epochai-0.1.3.tar.gz | 6.3 kB | Without provenance | Details for epochai-0.1.3.tar.gz |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| epochai-0.1.3-py3-none-any.whl | Python 3 | none | any | Details for epochai-0.1.3-py3-none-any.whl |
Total release size: 10.8 kB
Release files / epochai-0.1.3.tar.gz
| Download URL | epochai-0.1.3.tar.gz |
|---|---|
| Size | 6.3 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
b72e433ee3e77d873bfc670aac49bace519cc8e6b0ba1bdaf27c1b7edff4e934
|
|
BLAKE2b-256 checksum How to use checksums |
1e1717bbb09803106770c1d0ee893283a13d7153ab9479b72d2cd140aa0bd5ec
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
poetry/1.8.2 CPython/3.12.6 Darwin/24.3.0
|
Release files / epochai-0.1.3-py3-none-any.whl
| Download URL | epochai-0.1.3-py3-none-any.whl |
|---|---|
| Size | 4.6 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
9953ea4d85a4ad0f6050a87bfa189ceff1f4a04d762bd77a1a776a3bd7f73410
|
|
BLAKE2b-256 checksum How to use checksums |
ba9d8e5701b2154e667cfa716b7a928e95e5879d45198437e21771d67679c335
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
poetry/1.8.2 CPython/3.12.6 Darwin/24.3.0
|
Security for epochai 0.1.3
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.