Skip to main content

Ready-to-assemble AI

Project description

Flatpack

Flatpack

NOTE: Flatpack is currently experimental. Please refrain from using it in production environments.

注意: Flatpack 目前仍处于实验阶段。请勿用于生产环境。

Ready-to-assemble AI

At Flatpack, our mission is clear: we are committed to building trust in AI.

Flatpack democratises AI and ML through micro language models and model compression. Our platform enables users to train custom language models with 100M to 10B parameters. We introduce flatpacks (FPKs) to integrate AI and ML into edge computing, electronic components, and robots.

在 Flatpack,我们的使命十分明确: 致力于建立对 AI 的信任。

Flatpack 通过微语言模型和模型压缩普及 AI 和 ML。我们的平台使用户能够训练具有 100M 到 10B 个参数的自定义语言模型。我们引入了 flatpacks (FPK),将 AI 和 ML 集成到边缘计算、电子元件和机器人中。

Flatpack 3.X.X (Aglaonice)

Aglaonice, an ancient Greek astronomer from the 2nd or 1st century BC, was celebrated for her precise lunar eclipse predictions. Her mastery inspired the Greek proverb: "As the moon obeys Aglaonice," signifying unwavering certainty.

3.10.0 (2024-10-21)
Added dark mode support for a better user experience.

3.9.0 (2024-10-04)
Added Monaco Editor and custom hooks support.

3.8.0 (2024-09-24)
Improved package setup and deprecated agent spawning.

3.7.0 (2024-09-04)
Added cron and manual scheduling for builds.

3.6.0 (2024-07-28)
Introduced SQLite database support for each flatpack.

3.5.0 (2024-05-27)
Initial support for model compression in GGUF using llama.cpp.

3.4.0 (2024-04-30)
Added support for spawning agents with micro APIs.

3.3.0 (2024-04-16)
Added a vector database for storing and querying embeddings.

3.2.0 (2024-03-09)
Added support for unboxing local flatpacks using --local.

3.1.0 (2023-12-11)
Introduced a local web interface for a better user experience.

3.0.0 (2023-10-20)
Moved to a predictable and structured release strategy.

Moving to versioning structure: 3.0.0

Our previous releases were a mix of minor tweaks and significant shifts, making it challenging to anticipate the nature of changes. With the introduction of version 3.0.0, we are embracing a predictable and structured release strategy:

  • Major versions (X.0.0): These signify significant changes or updates.
  • Minor versions (3.X.0): Introduce new features without breaking compatibility.
  • Patch versions (3.0.X): Address bug fixes and minor refinements.

This move is not merely about semantic versioning. It is a pledge for clear communication and trust with our users. We invite you to explore our new release strategy and appreciate your patience as we evolve.

Explore the Project on GitHub

License

This project is released under Apache-2.0.

install_requires

DISCLAIMER: This information is only a technical reference, not an endorsement or legal advice. Before using any software for commercial purposes, perform compatibility checks and seek legal advice. You are responsible for ensuring compliance with licensing requirements.

Check out the JLA - Compatibility Checker (European Commission 2024).

Last updated: 2024-10-22

Project details


Release history Release notifications | RSS feed

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

flatpack-3.10.26.tar.gz (56.1 kB view details)

Uploaded Source

Built Distribution

flatpack-3.10.26-py3-none-any.whl (55.4 kB view details)

Uploaded Python 3

File details

Details for the file flatpack-3.10.26.tar.gz.

File metadata

  • Download URL: flatpack-3.10.26.tar.gz
  • Upload date:
  • Size: 56.1 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/5.0.0 CPython/3.11.10

File hashes

Hashes for flatpack-3.10.26.tar.gz
Algorithm Hash digest
SHA256 6cd45f86d364254a92e6d81e89dfa6ced416269b657486b4083e1d873a112f8d
MD5 908b67d634c5060cb7dcdba7d2cf9ced
BLAKE2b-256 b7d304028ad91269720ca4842266a0ce81dfba9e54b3839ff47c0754f5ad84c0

See more details on using hashes here.

File details

Details for the file flatpack-3.10.26-py3-none-any.whl.

File metadata

  • Download URL: flatpack-3.10.26-py3-none-any.whl
  • Upload date:
  • Size: 55.4 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/5.0.0 CPython/3.11.10

File hashes

Hashes for flatpack-3.10.26-py3-none-any.whl
Algorithm Hash digest
SHA256 90c6e0d1bdfafa6466914716d35d993fae9b80ec26661829cdea4a15652d7670
MD5 51d93fcb5a0c0fb3cb39943eacfeba58
BLAKE2b-256 13815eb912fd7b304a8c1cf3ce2e821d5c61a3a283ddcf2bdc43999c181614e8

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page