Skip to main content

GoodWoolf

GoodWoolf

AI by the People

Foundation Models · Commodity Hardware · Collaborative Training

Website  ·  Docs  ·  Main Repository  ·  Training-Ledger


Building Foundation Models Collaboratively on Commodity Hardware

GoodWoolf builds large foundation models on commodity hardware using the following principles:

  • Open Collaboration. Enable collective model development across independent contributors, coordinated through a shared public training ledger.

  • Virtual Memory Extension: Expanding effective memory beyond physical RAM by leveraging storage-based virtual memory.

  • Data-Independent Scaling: Decoupling total training time from global data volume, bounding system latency to individual run durations rather than aggregate data sizes.

Learn more by following the online documenation:

Project Structure

The project consists of two repositories:

Collaborative Training Workflow

The training process is hardware-agnostic, parallel, and completely asynchronous.

=======================================================
Visual represenation of Collaborative Training Workflow
=======================================================

          +-----------+----------+-----------+
          |           |          |           |
          v           |          |           |
  +----------------+  |          v           |
  | Participant A  |  |  +----------------+  |
  |----------------|  |  | Participant B  |  |
  | Local data     |  |  |----------------|  |
  | Train locally  |  |  | Local data     |  |
  | Publish model  |  |  | Train locally  |  |
  | Pull request   |  |  | Publish model  |  |
  +-------+--------+  |  | Pull request   |  |
          |           v  +--------+-------+  |
          |   +----------------+  |          |
          |   | Participant C  |  |          |
          |   |----------------|  |          |
          |   | Local data     |  |          |
          |   | Train locally  |  |          |
          |   | Model offload  |  |          |
          |   | Pull request   |  |          |
          |   +------+---------+  |          |
          |          |            |          |
          v          v            v          |
       +-----------------------------+       |
       |   PUBLIC TRAINING LEDGER    |-------+
       |-----------------------------|
       | Base model weights          |
       | Training updates / PRs      |<------+
       | Training history            |       |
       +-----------------------------+       |
                     |                       |
                     v                       |
       +-----------------------------+       |
       |   Peers / Independent       |       |
       |   Auditors                  |-------+
       |-----------------------------|
       | Validate and aggregate      |
       | model updates               |
       | Publish updated model       |
       | Pull request                |
       +-----------------------------+

Usage Examples

Prepare training data

goodwoolf prepare <vendor>/<model> <dataset>

Download model weights

goodwoolf pull <vendor>/<model>

Train model

goodwoolf pretrain <vendor>/<model> <dataset>

Generate text

goodwoolf run <vendor>/<model> --prompt "A long time ago..."

Get involved

GoodWoolf is built by an open-source community project.

Whether you're trainig and sharing models, improving the platform, fixing bugs or writing documentation, your contributions are welcome.

License

Metadata

Release files for goodwoolf 0.0.1

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for goodwoolf 0.0.1
File Size Uploaded Provenance Actions
goodwoolf-0.0.1.tar.gz 2.1 MB Without provenance Details for goodwoolf-0.0.1.tar.gz

Built distribution (wheel)

Table of built distributions (wheels) for goodwoolf 0.0.1
File Interpreter ABI Platform Actions
goodwoolf-0.0.1-py3-none-any.whl Python 3 none any Details for goodwoolf-0.0.1-py3-none-any.whl

Total release size: 4.3 MB

Release files / goodwoolf-0.0.1.tar.gz

Download URL goodwoolf-0.0.1.tar.gz
Size 2.1 MB
Tags Source
SHA-256 checksum
How to use checksums
0b55f3274265011dcf7e1201e3eeeef06d623d6d6cfeb0b1064435c147a15807
BLAKE2b-256 checksum
How to use checksums
7c987d347c754ad6979450c9d0188bced9eb0fed4fae8e596a98c8eee010bfbc
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/7.0.0 CPython/3.12.13

Release files / goodwoolf-0.0.1-py3-none-any.whl

Download URL goodwoolf-0.0.1-py3-none-any.whl
Size 2.2 MB
Tags Python 3
SHA-256 checksum
How to use checksums
7688a78e4aa0925d3f56bd4c806ce7c4cea242f9e6b7a2385fa8996e3e38db13
BLAKE2b-256 checksum
How to use checksums
4996a8a87376657adf967ac22d428b0e3f1777f8b93ded0d2a776c8edc89889d
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/7.0.0 CPython/3.12.13

Security for goodwoolf 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.

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.

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page