Model Database's logo
Tasks

Feature Extraction

Feature extraction refers to the process of transforming raw data into numerical features that can be processed while preserving the information in the original dataset.

Inputs
Input

India, officially the Republic of India, is a country in South Asia.

Feature Extraction Model
Output
Dimension 1 Dimension 2 Dimension 3
2.583383083343506 2.757075071334839 0.9023529887199402
8.29393482208252 1.1071064472198486 2.03399395942688
-0.7754912972450256 -1.647324562072754 -0.6113331913948059
0.07087723910808563 1.5942802429199219 1.4610432386398315

About Feature Extraction

About the Task

Feature extraction is the task of building features intended to be informative from a given dataset, facilitating the subsequent learning and generalization steps in various domains of machine learning.

Use Cases

Feature extraction can be used to do transfer learning in natural language processing, computer vision and audio models.

Inference

Feature Extraction 

from transformers import pipeline
checkpoint = "facebook/bart-base"
feature_extractor = pipeline("feature-extraction",framework="pt",model=checkpoint)
text = "Transformers is an awesome library!"

#Reducing along the first dimension to get a 768 dimensional array
feature_extractor(text,return_tensors = "pt")[0].numpy().mean(axis=0) 

'''tensor([[[ 2.5834,  2.7571,  0.9024,  ...,  1.5036, -0.0435, -0.8603],
         [-1.2850, -1.0094, -2.0826,  ...,  1.5993, -0.9017,  0.6426],
         [ 0.9082,  0.3896, -0.6843,  ...,  0.7061,  0.6517,  1.0550],
         ...,
         [ 0.6919, -1.1946,  0.2438,  ...,  1.3646, -1.8661, -0.1642],
         [-0.1701, -2.0019, -0.4223,  ...,  0.3680, -1.9704, -0.0068],
         [ 0.2520, -0.6869, -1.0582,  ...,  0.5198, -2.2106,  0.4547]]])'''

Useful resources

Compatible libraries

Sentence Transformers Transformers Transformers.js
Feature Extraction demo
using facebook/bart-base
Feature Extraction
This model can be loaded on the Inference API on-demand.
Models for Feature Extraction
Browse Models (5,347)

Note A powerful feature extraction model for natural language processing tasks.

Datasets for Feature Extraction
Browse Datasets (134)

Note Wikipedia dataset containing cleaned articles of all languages. Can be used to train `feature-extraction` models.

Spaces using Feature Extraction

No example Space is defined for this task.

Note Contribute by proposing a Space for this task !

Metrics for Feature Extraction