Wrapper around evotorch to simplify any optimization task
Project description
FunctionStudio
💡 Announcing the Development of a new Optimization Library
As an AI professional, I'm excited to share that I'm developing a new optimization library, planned for release later this year. This library aims to address key challenges in the field:
❓ Why another optimization library?
- Current tools like Scikit-learn, PyTorch, and TensorFlow offer limited support for multi-objective optimization. This library simplifies finding trade-off models for any combination of objectives, aimed to ease the use of multi-objective optimization in model development.
- Users will have the ability to choose whether to train models end-to-end or as separate modules. This feature is particularly useful for optimizing decisions composed of multiple sub-decisions.
- Built on a low-code language (DMN) for both the input and output layers, the library ensures interpretability, interoperability, and interchangeability of model chains.
- The library will support single-objective reinforcement learning, multi-objective reinforcement learning, supervised learning, and multi-objective supervised learning, all through a single, user-friendly interface.
To do
- Implement class called Function allowing to fit any multi-objective function by adding objectives.✅
- Implement supervised example case on how to use function class ✅
- Implement Reinforcement learning example for function ✅
- Implement CompositeFunction class ✅
- Implement reinforcement learning example for CompositeFunction class ✅
- Implement multi-objective reinforcement learning example ✅
- Implement the reporting module to visualize the found pareto front❌
- Implement the dmn input and output layer to convert function from and to dmn❌
- Setup pypi package, requirements file❌
- Make unit tests❌
- document classes in code❌
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
FunctionStudio-0.0.1.tar.gz
(7.7 kB
view hashes)
Built Distribution
Close
Hashes for FunctionStudio-0.0.1-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | bdac5c616890a94e372dbe559f8ad07cd0ce5290b5eb37f83f6bee6d67639ae4 |
|
MD5 | e9acbc4c41f44169e092ccfbc904090c |
|
BLAKE2b-256 | 7549ca7a71c3b8f697ce4c1be729cf5fe24d387efcebb1389195810cab4df1c7 |