a
PSic @ sd d dl mZ d dlZd dlmZmZmZmZ d dlZ ddl
m Z m Z ddl
mZ G dd deZdS ) ) annotationsN)AnyCallableOptionalTuple ) download_and_extract_archiveverify_str_arg)
VisionDatasetc sp e Zd ZdZdZddddddd d
d fd d
ZddddZddddZd
dddZd dddZ Z
S ) FGVCAircrafta `FGVC Aircraft `_ Dataset.
The dataset contains 10,000 images of aircraft, with 100 images for each of 100
different aircraft model variants, most of which are airplanes.
Aircraft models are organized in a three-levels hierarchy. The three levels, from
finer to coarser, are:
- ``variant``, e.g. Boeing 737-700. A variant collapses all the models that are visually
indistinguishable into one class. The dataset comprises 100 different variants.
- ``family``, e.g. Boeing 737. The dataset comprises 70 different families.
- ``manufacturer``, e.g. Boeing. The dataset comprises 30 different manufacturers.
Args:
root (string): Root directory of the FGVC Aircraft dataset.
split (string, optional): The dataset split, supports ``train``, ``val``,
``trainval`` and ``test``.
annotation_level (str, optional): The annotation level, supports ``variant``,
``family`` and ``manufacturer``.
transform (callable, optional): A function/transform that takes in an PIL image
and returns a transformed version. E.g, ``transforms.RandomCrop``
target_transform (callable, optional): A function/transform that takes in the
target and transforms it.
download (bool, optional): If True, downloads the dataset from the internet and
puts it in root directory. If dataset is already downloaded, it is not
downloaded again.
zWhttps://www.robots.ox.ac.uk/~vgg/data/fgvc-aircraft/archives/fgvc-aircraft-2013b.tar.gztrainvalvariantNFstrzOptional[Callable]boolNone)rootsplitannotation_level transformtarget_transformdownloadreturnc s t j|||d t|dd| _t|dd| _tj| jd| _ |rL|
| s\t dtj| j dd d
d d| j }t
|d }d
d |D | _W d n1 s0 Y tt| jtt| j| _tj| j dd} tj| j dd| j d| j d}
g | _g | _t
|
d ^}|D ]H} | dd\} }
| jtj| | d | j| j|
q$W d n1 s0 Y d S )N)r r r )trainvalr testr )r
familyZ manufacturerzfgvc-aircraft-2013bz;Dataset not found. You can use download=True to download itdataz variants.txtz families.txtzmanufacturers.txtrc S s g | ] }| qS )strip).0liner r ^/var/www/html/django/DPS/env/lib/python3.9/site-packages/torchvision/datasets/fgvc_aircraft.py
J z)FGVCAircraft.__init__..imagesZimages__z.txt r z.jpg)super__init__r _splitZ_annotation_levelospathjoinr
_data_path _download
_check_exists RuntimeErroropenclassesdictziprangelen class_to_idx _image_files_labelsr r append)selfr r r r r r Zannotation_filefZimage_data_folderZ labels_filer! Z
image_name
label_name __class__r r" r) * s@ . $ zFGVCAircraft.__init__int)r c C s
t | jS N)r7 r9 r< r r r" __len__Z s zFGVCAircraft.__len__zTuple[Any, Any]c C sP | j | | j| }}tj|d}| jr8| |}| jrH| |}||fS )NRGB)r9 r: PILImager2 convertr r )r< idx
image_filelabelimager r r" __getitem__] s
zFGVCAircraft.__getitem__c C s | r dS t| j| j dS )zW
Download the FGVC Aircraft dataset archive and extract it under root.
N)r0 r _URLr rC r r r" r/ i s zFGVCAircraft._downloadc C s t j| jot j| jS rB )r+ r, existsr. isdirrC r r r" r0 q s zFGVCAircraft._check_exists)r r
NNF) __name__
__module__ __qualname____doc__rN r) rD rM r/ r0
__classcell__r r r? r" r s 0 r )
__future__r r+ typingr r r r PIL.ImagerF utilsr r visionr
r r r r r" s