Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/rxxuzi/doomsday
https://github.com/rxxuzi/doomsday
Last synced: about 1 month ago
JSON representation
- Host: GitHub
- URL: https://github.com/rxxuzi/doomsday
- Owner: rxxuzi
- License: mit
- Created: 2023-09-03T17:47:15.000Z (over 1 year ago)
- Default Branch: main
- Last Pushed: 2024-01-14T21:17:40.000Z (12 months ago)
- Last Synced: 2024-01-15T01:18:10.875Z (12 months ago)
- Language: Scala
- Size: 168 KB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
[](https://github.com/rxxuzi/doomsday/actions/workflows/scala.yml)
## Introduction
Doomsday is an emerging deep learning framework written in Scala.
Aimed at offering seamless numerical computations with the power of Scala,
it leverages the Breeze library to provide high-speed numerical and linear algebraic operations.
Please note that this library is still under development, and contributions or feedback are highly appreciated.
## Version Information
- **SDK**: 1.8
- **Scala Version**: 3.1.3
- **Breeze**: 2.1.0
## Dependencies
- [Breeze](https://github.com/scalanlp/breeze) - A library for numerical processing and scientific computing in Scala.
## Quick Start
1. **Clone the repository**:
```shell
git clone https://github.com/rxxuzi/doomsday.git
```
2. **Add Doomsday to your sbt project**:
In your `build.sbt` file, add the following dependency:
~~~sbt
libraryDependencies += "com.rxxuzi" %% "doomsday" % "latest.version"
~~~
*(Replace "latest.version" with the current version of Doomsday.)*
3. **Using the library**:
After setting up the dependency, you can start using Doomsday in your project:
~~~scala
import doomsday.core._
import doomsday.function._
import doomsday.models._
import doomsday.dataset._
import doomsday.optimizers._
~~~
Sample code can be found in [example](.docs/examples).
You can now utilize all the functionalities provided by Doomsday.
## Features
- **Flexibility**: Design and train custom neural network architectures tailored to your needs.
- **Optimizers**: Multiple optimization algorithms available like SGD, Adam, RMSprop, etc.
- **Loss Functions**: A variety of loss functions to choose from, including MSE, Cross-Entropy, and more.
- **Regularization Techniques**: Support for techniques like Dropout, Batch Normalization, and L1/L2 regularization.
- **Datasets**: Built-in utilities to handle and preprocess datasets for machine learning tasks.
- **Model Evaluation**: Tools for evaluating model performance using metrics like accuracy, precision, recall, etc.
- **Functional API**: Leverage the power of Scala's functional programming capabilities for more expressive model design.
## Documentation
For detailed usage instructions, troubleshooting, and more, check out our documentation in the **[document](.docs)** folder.
## License
This project is licensed under the MIT License. For more details, see the [LICENSE](LICENSE) file.