The fastai library allows you to train a
Model on a certain
DataBunch very easily by binding them together inside a
Learner object. This module regroups the tools the library provides to help you preprocess and group your data in this format.
This submodule handles the collaborative filtering problems.
This sub-package deals with tabular (or structured) data.
This sub-package contains everything you need for Natural Language Processing.
This sub-package contains the classes that deal with Computer Vision.
In each case (except for
collab), the module is organized this way:
This sub-module deals with the pre-processing (data augmentation for images, cleaning for tabular data, tokenizing and numericalizing for text).
This sub-module defines the dataset class(es) to deal with this kind of data.
This sub-module defines the specific models used for this kind of data.
When it exists, this sub-module contains functions that will directly bind this data with a suitable model and add the necessary callbacks.
The general structure is:
from fastai.[APPLICATION] import *
For example, to use collab:
from fastai.collab import *