gan-go

KO | EN

gitlite — search

#react #llm #go #golang #machine-learning #deep-learning #docker #macos #windows #awesome #pytorch #mcp #ios #openai

(★ 87)

Generative Adversarial Network in Go via Gorgonia

gan-go 최신버젼 다운로드

최종 버전 다운로드 (.zip)

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# Generative adversarial networks
**Recipe for simple GAN in Golang ecosystem via [Gorgonia library](https://gorgonia.org/)**

Generating f(x) = x^2             |  Generating f(x) = sin(x)
:-------------------------:|:-------------------------:
<img src="cmd/examples/parabola/output/out_parabola.gif" width="240">  |  <img src="cmd/examples/sin/output/out_sine.gif" width="240">

## Table of Contents

- [About](#about)
- [Why](#why)
- [Instruments](#instruments)
- [Usage](#usage)
- [ToDo](#todo)
- [Code explanation](#code-explanation)
- [Support and contributing](#support-and-contributing)

## About
This just a repository with simple example how to build GAN in [Gorgonia](https://gorgonia.org/)

What is GAN? 
* Short: such networks are just two neural networks (**Discriminator** and **Generator**) contesting each other. **Generator** must "cheat" on **Discriminator** and the last one should detect lies.
* Long: Wiki-page about GAN's - https://en.wikipedia.org/wiki/Generative_adversarial_network

_Note_: *although there is code with some wrappings/aliasing and helping abstractions and functions in repository, this does not pretend to be high-level machine learning framework*

_Note #2_: *By the way... Code is ugly since I've decided to handle errors instead of using panic(...) calls. Panicing is considered to be in main functions of examples only*

Current examples folder contains limited set of layer types:
- [x] Linear
- [x] Convolutional
- [x] Maxpool
- [x] AvgPool
- [x] Flatten
- [x] Reshape
- [x] Dropout
- [x] Embedding
- [x] LSTM
- [x] RNN

_Note #3_: *some layer types (e.g. LSTM, RNN) and examples (train_cnn, train_embedding, train_lstm, train_rnn) are not related to GAN topic itself. They just show how to build neural networks of other kinds with the same set of abstractions*

## Why
Just want to do that in Golang ecosystem.

## Instruments
Code is written on Golang - https://golang.org/

Used machine learning library - [Gorgonia](https://github.com/gorgonia/gorgonia)

Plotting library - [gonum](https://github.com/gonum/plot#gonum-plot)

## Usage

* Get the repo
    ```shell
    git clone https://github.com/LdDl/gan-go.git
    ```

* Navigate to examples folder
    ```shell
    cd gan-go
    cd cmd/examples
    ```

* Pick one of examples. E.g. parabola:
    ```shell
    cd parabola
    ```

* Run example
    ```
    go run main.go
    ```

* Output

After programm terminates there should be multiple files:
    1. Single image for reference function - reference_function.png
    2. Multiple images for generated data on N-th epoch - gen_reference_fun_%N-th epoch%.png
    3. Single image for generated data on last epoch - gen_reference_func_final.png

Example for parabola:

Actual reference function:

![Reference function](https://raw.githubusercontent.com/LdDl/gan-go/master/cmd/examples/parabola/output/reference_function.png)

Generated data on 0-th epoch:

![Generated data on 0-th epoch](https://raw.githubusercontent.com/LdDl/gan-go/master/cmd/examples/parabola/output/gen_reference_func_0.png)

Generated data on 10-th epoch:

![Generated data on 10-th epoch](https://raw.githubusercontent.com/LdDl/gan-go/master/cmd/examples/parabola/output/gen_reference_func_10.png)

Generated data on 60-th epoch:

![Generated data on 60-th epoch](https://raw.githubusercontent.com/LdDl/gan-go/master/cmd/examples/parabola/output/gen_reference_func_60.png)

Generated data on 150-th epoch:

![Generated data on 150-th epoch](https://raw.githubusercontent.com/LdDl/gan-go/master/cmd/examples/parabola/output/gen_reference_func_150.png)

Generated data on last epoch:

![Generated data on last epoch](https://raw.githubusercontent.com/LdDl/gan-go/master/cmd/examples/parabola/output/gen_reference_func_final.png)

## ToDo
Current stage of TODO list for future releases:
- [x] Reduce duplicating of code for *.Fwd()* method of each neural network type (GAN/Discriminator/Generator)
- [x] Switch **Layer** from *struct* to *interface* or use other technique for building clean code
- [x] Add basic layers: Linear, Convolutional, Maxpool, Flatten
- [ ] Deal with batch process 
- [x] More loss function
    - [x] Cross Entropy
    - [x] Binary Cross Entropy
    - [x] L1
    - [x] Huber (PSEUDO)
- [x] Examples for text data generation
- [x] Simple LSTM
    - [x] Proper layer type
    - [x] Examples
- [x] RNN
- [ ] GRU
- [x] Embedding

## Code explanation
@TODO

## Support and contributing
If you have troubles or questions please [open an issue](https://github.com/LdDl/gan-go/issues/new).

If you want to improve this library / fix some bugs and etc. you can [make PR](https://github.com/LdDl/gan-go/compare)

gan-go

gan-go 최신버젼 다운로드

오늘의 추천 프로젝트

Was this content helpful?

Recent Feedback

gan-go

gan-go 최신버젼 다운로드

오늘의 추천 프로젝트

Was this content helpful?

Recent Feedback

Download README