Given an image, can you find all of the same landmarks in a dataset?

guanghan_nas_2

Google Landmark Retrieval 2021

Louis Official

Rate the effectiveness of argumentative writing elements from students grade 6-12

Kaggle 比赛：<a href="https://www.kaggle.com/competitions/feedback-prize-effectiveness/data">Feedback Prize - Predicting Effective Arguments</a>原版数据集

Feedback Prize - Predicting Effective Arguments

Identify whales and dolphins by unique characteristics

Kaggle 快乐鲸原版数据集，比赛地址：<a href="https://www.kaggle.com/c/happy-whale-and-dolphin">https://www.kaggle.com/c/happy-whale-and-dolphin</a>

Happywhale - Whale and Dolphin Identification

Segment multi-organ functional tissue units

Kaggle 比赛：<a href="https://www.kaggle.com/competitions/hubmap-organ-segmentation/data">https://www.kaggle.com/competitions/hubmap-organ-segmentation/data</a>原版数据集&nbsp;

HuBMAP + HPA - Hacking the Human Body

Predict the relationship between code and comments

Google AI4Code – Understand Code in Python Notebooks

yaan_nas_1

包含：<ul><li>5489 个训练图片</li><li>234 个验证图片</li></ul>标签分布和原版大致相同。

COCO 2017 MINI 版

Identify glomeruli in human kidney tissue images

Kaggle 比赛 <a href="https://www.kaggle.com/c/hubmap-kidney-segmentation">https://www.kaggle.com/c/hubmap-kidney-segmentation</a>原版数据集

HuBMAP - Hacking the Kidney

Predict how DNA, RNA & protein measurements co-vary in single cells


Kaggle 比赛<a href="https://www.kaggle.com/competitions/open-problems-multimodal"> https://www.kaggle.com/competitions/open-problems-multimodal</a>&nbsp;原版数据集

Open Problems - Multimodal Single-Cell Integration

guanghan_nas_1

cifar10

祖安的羊驼

<h2>COCO 数据集（2017部分）</h2>从官方下载，未经任何处理，下载后全部解压，数据集中包含：<ul><li><a href="http://images.cocodataset.org/zips/train2017.zip">2017 Train images [118K/18GB]</a></li><li><a href="http://images.cocodataset.org/zips/val2017.zip">2017 Val images [5K/1GB]</a></li><li><a href="http://images.cocodataset.org/zips/test2017.zip">2017 Test images [41K/6GB]</a></li><li><a href="http://images.cocodataset.org/zips/unlabeled2017.zip">2017 Unlabeled images [123K/19GB]</a></li><li><a href="http://images.cocodataset.org/annotations/annotations_trainval2017.zip">Train/Val annotations [241MB]</a></li><li><a href="http://images.cocodataset.org/annotations/stuff_annotations_trainval2017.zip">Stuff Train/Val annotations [1.1GB]</a></li><li><a href="http://images.cocodataset.org/annotations/panoptic_annotations_trainval2017.zip">Panoptic Train/Val annotations [821MB]</a></li><li><a href="http://images.cocodataset.org/annotations/image_info_test2017.zip">Testing Image info [1MB]</a></li><li><a href="http://images.cocodataset.org/annotations/image_info_unlabeled2017.zip">Unlabeled Image info [4MB]</a></li></ul>官方数据集地址：<a href="https://cocodataset.org/">https://cocodataset.org/</a>&nbsp;

COCO 2017

datasets_9.29.zip

B1randa

img_data.zip

9a22ce0f-9c77-48f0-ba9b-5fa9de6dec69_2ec35b01-65b1-450b-b258-63c5f6aac09a/coronahack-chest-xraydataset.zip

COVID-19-Chest-Xray

Dave

5 000 annotated images with fine annotations

<h2>Cityscapes 数据集</h2>从官方下载（仅解压操作），数据集中包含：<ul><li><a href="https://www.cityscapes-dataset.com/file-handling/?packageID=1">gtFine_trainvaltest.zip (241MB)</a> <a href="https://www.cityscapes-dataset.com/md5-sum/?packageID=1">[md5]</a></li><li><a href="https://www.cityscapes-dataset.com/file-handling/?packageID=3">leftImg8bit_trainvaltest.zip (11GB)</a> <a href="https://www.cityscapes-dataset.com/md5-sum/?packageID=3">[md5]</a></li></ul>官方数据集地址：<a href="https://www.cityscapes-dataset.com/downloads/">https://www.cityscapes-dataset.com/downloads/</a><h2>&nbsp;</h2>&nbsp;&nbsp;

Cityscapes数据集

搬砖人

cifar100

coco2017.zip

奔跑的阿峰

miniImageNet数据集节选自ImageNet数据集。miniImageNet包含100类共60000张彩色图片，其中每类有600个样本.

MiniImageNet

zhen

imagenet train

imagenet val

<h4>THE MNIST DATABASE（手写数字识别）</h4>数据来源：<a href="http://yann.lecun.com/exdb/mnist/">http://yann.lecun.com/exdb/mnist/</a>&nbsp;该数据已将原数据的 4 个压缩包解压，下面是读取数据的代码，添加数据后可以直接使用：&nbsp;<pre><code class="language-python">import numpy as np
import os

def extract(folder): 
 def extract_file(file_name, num, shape):
 lead_bytes = 16 if 'images' in file_name else 1
 with open(os.path.join(folder, file_name), 'rb') as f:
 f.read(lead_bytes)
 buffer = f.read(np.prod((*shape, num)))
 data = np.frombuffer(buffer, dtype=np.uint8)
 data.reshape(num, *shape, 1)
 return data

 return (
 extract_file('train-images-idx3-ubyte', 60000, (28, 28)),
 extract_file('train-labels-idx1-ubyte', 60000, (1,)),
 extract_file('t10k-images-idx3-ubyte', 10000, (28, 28)),
 extract_file('t10k-labels-idx1-ubyte', 10000, (1,))
 )

train_images, train_labels, test_images, test_labels = extract('/home/featurize/data/mnist')</code></pre>&nbsp;

数据集

标签

数据集大小

数据集来源