https://github.com/norwegianseismicarray/nais

NORSAR AI System (tools for machine learning on waveform data)
https://github.com/norwegianseismicarray/nais

deep-learning keras machine-learning python tensorflow

Last synced: 17 days ago
JSON representation

NORSAR AI System (tools for machine learning on waveform data)

• Host: GitHub
• URL: https://github.com/norwegianseismicarray/nais
• Owner: NorwegianSeismicArray
• License: mit
• Created: 2022-02-22T11:38:16.000Z (almost 3 years ago)
• Default Branch: main
• Last Pushed: 2023-11-03T15:19:54.000Z (about 1 year ago)
• Last Synced: 2024-12-16T07:21:24.940Z (20 days ago)
• Topics: deep-learning, keras, machine-learning, python, tensorflow
• Language: Python
• Homepage:
• Size: 1010 KB
• Stars: 2
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
### [Documentation](https://norwegianseismicarray.github.io/nais/nais)

### NORSAR Artificial Inteligence System (NAIS, yes, the acronym was chosen first)

This repository will eventually contain tools specific to aiding in AI research at NORSAR.

This repository is based on Tensorflow and Keras. 

This repository will contain: 

* Baseline models, for waveforms and spectrograms. Possibly pretrained for certain tasks.

* Standard datasets to test models. Development of new models can be difficult. Standard datasets eliminates errors in the data and lets you focus on developing the model. 

  * Classification

  * Regression

  * Masking (eg. arrival picking)

* Augmentation methods.

## Installation

``pip install nais``

If an error occurs when importing `nais`, likely ``sndfile`` library is not installed (check the error), and needs to be:

``apt-get install libsndfile1-dev``

# Quick example

```python

import numpy as np

from nais.Models import AlexNet1D

X = np.random.normal(size=(16,256,3)) #16 examples of three channel data.

y = np.random.randint(0,1,size=(16,)) #Labels

model = AlexNet1D(num_outputs=1) #binary 

model.compile('adam','binary_crossentropy')

model.fit(X,y)

```