【DSW Gallery】基于MAE的自监督学习示例

本文涉及的产品
模型训练 PAI-DLC,5000CU*H 3个月
交互式建模 PAI-DSW,5000CU*H 3个月
模型在线服务 PAI-EAS,A10/V100等 500元 1个月
简介: EasyCV是基于Pytorch,以自监督学习和Transformer技术为核心的 all-in-one 视觉算法建模工具,并包含图像分类,度量学习,目标检测,姿态识别等视觉任务的SOTA算法。本文自监督学习-MAE为例,为您介绍如何在PAI-DSW中使用EasyCV。

直接使用

请打开基于MAE的自监督学习示例,并点击右上角 “ 在DSW中打开” 。

image.png

EasyCV自监督训练-MAE

  自监督学习(Self-Supervised Learning)能利用大量无标注的数据进行表征学习,然后在特定下游任务上对参数进行微调。通过这样的方式,能够在较少有标注数据上取得优于有监督学习方法的精度。

  近年来,自监督学习受到了越来越多的关注,如Yann Lecun也在 AAAI 上讲 Self-Supervised Learning 是未来的大势所趋。在CV领域涌现了如SwAV、MOCO、DINO、MoBY等一系列工作。MAE是kaiming继MOCO之后在自监督学习领域的又一力作。

  本文将介绍如何在pai-dsw基于EasyCV快速使用MAE进行度量模型的训练、微调、推理。

运行环境要求

PAI-Pytorch 1.7/1.8镜像, GPU机型 P100 or V100, 内存 32G

安装依赖包

注:在PAI-DSW docker中无需安装相关依赖,可跳过此1,2步骤, 在本地notebook环境中执行1,2 步骤安装环境

1、获取torch和cuda版本,并根据版本号修改mmcv安装命令,安装对应版本的mmcv和nvidia-dali

import torch
import os
os.environ['CUDA']='cu' + torch.version.cuda.replace('.', '')
os.environ['Torch']='torch'+torch.version.__version__.replace('+PAI', '')
!echo $CUDA
!echo $Torch
[2023-02-02 13:22:04,960.960 dsw-16577-76cf6dbbb4-m7pbn:1125 INFO utils.py:30] NOTICE: PAIDEBUGGER is turned off.
cu101
torch1.8.2
# install some python deps
! pip install --upgrade tqdm
! pip install mmcv-full==1.4.4 -f https://download.openmmlab.com/mmcv/dist/cu101/torch1.8.0/index.html
! pip install http://pai-vision-data-hz.oss-cn-zhangjiakou.aliyuncs.com/third_party/nvidia_dali_cuda100-0.25.0-1535750-py3-none-manylinux2014_x86_64.whl

2、安装EasyCV算法包 注:在PAI-DSW docker中预安装了pai-easycv库,可跳过该步骤,若训练测试过程中报错,尝试用下方命令更新easycv版本

#pip install pai-easycv
! echo y | pip uninstall pai-easycv easycv
!pip install pai-easycv

3、简单验证

from easycv.apis import *

正式工作

数据准备

  自监督训练只需要提供无标注图片即可进行, 你可以下载ImageNet数据,或者使用你自己的图片数据。需要提供一个包含若干图片的文件夹路径p,以及一个文件列表,文件列表中是每个图片相对图片目录p的路径.

  图片文件夹结构示例如下, 文件夹路径为./images

images/
├── 0001.jpg
├── 0002.jpg
├── 0003.jpg
|...
└── 9999.jpg
文件列表内容示例如下
0001.jpg
0002.jpg
0003.jpg
...
9999.jpg

为了快速走通流程,我们也提供了一个小的示例数据集,执行如下命令下载解压

! wget http://pai-vision-data-hz.oss-cn-zhangjiakou.aliyuncs.com/data/imagenet_raw_demo/imagenet_raw_demo.tar.gz && tar -zxf imagenet_raw_demo.tar.gz
Will not apply HSTS. The HSTS database must be a regular and non-world-writable file.
ERROR: could not open HSTS store at '/root/.wget-hsts'. HSTS will be disabled.
--2023-02-02 15:01:45--  http://pai-vision-data-hz.oss-cn-zhangjiakou.aliyuncs.com/data/imagenet_raw_demo/imagenet_raw_demo.tar.gz
Resolving pai-vision-data-hz.oss-cn-zhangjiakou.aliyuncs.com (pai-vision-data-hz.oss-cn-zhangjiakou.aliyuncs.com)... 39.98.20.13
Connecting to pai-vision-data-hz.oss-cn-zhangjiakou.aliyuncs.com (pai-vision-data-hz.oss-cn-zhangjiakou.aliyuncs.com)|39.98.20.13|:80... connected.
HTTP request sent, awaiting response... 200 OK
Length: 33109699 (32M) [application/x-gzip]
Saving to: ‘imagenet_raw_demo.tar.gz’
imagenet_raw_demo.t 100%[===================>]  31.58M  15.8MB/s    in 2.0s    
2023-02-02 15:01:47 (15.8 MB/s) - ‘imagenet_raw_demo.tar.gz’ saved [33109699/33109699]
# 重命名文件夹
! mv imagenet_raw_demo  imagenet_raw

训练模型

这个Demo中我们采用MAE自监督算法训练vit-base主干网络, 下载示例配置文件

! rm -rf mae_vit_base_patch16_8xb64_1600e.py
! wget http://pai-vision-data-hz.oss-cn-zhangjiakou.aliyuncs.com/release/doc/easycv/configs/selfsup/mae/mae_vit_base_patch16_8xb64_1600e.py
Will not apply HSTS. The HSTS database must be a regular and non-world-writable file.
ERROR: could not open HSTS store at '/root/.wget-hsts'. HSTS will be disabled.
--2023-02-02 15:02:47--  http://pai-vision-data-hz.oss-cn-zhangjiakou.aliyuncs.com/release/doc/easycv/configs/selfsup/mae/mae_vit_base_patch16_8xb64_1600e.py
Resolving pai-vision-data-hz.oss-cn-zhangjiakou.aliyuncs.com (pai-vision-data-hz.oss-cn-zhangjiakou.aliyuncs.com)... 39.98.20.13
Connecting to pai-vision-data-hz.oss-cn-zhangjiakou.aliyuncs.com (pai-vision-data-hz.oss-cn-zhangjiakou.aliyuncs.com)|39.98.20.13|:80... connected.
HTTP request sent, awaiting response... 200 OK
Length: 2348 (2.3K) [text/x-python]
Saving to: ‘mae_vit_base_patch16_8xb64_1600e.py’
mae_vit_base_patch1 100%[===================>]   2.29K  --.-KB/s    in 0s      
2023-02-02 15:02:48 (228 MB/s) - ‘mae_vit_base_patch16_8xb64_1600e.py’ saved [2348/2348]

为了缩短训练时间,打开配置文件 mae_vit_base_patch16_8xb64_1600e.py,修改total_epoch参数为5, 每隔1次迭代打印一次日志。

# runtime settings
total_epochs = 5
# log config
log_config=dict(interval=1)

正式训练时,建议使用多机8卡或单机8卡配合该配置文件使用,修改update_interval参数进行梯度累积,确保有效batch_size一致

# 查看easycv安装位置
import easycv
print(easycv.__file__)
/home/pai/lib/python3.6/site-packages/easycv/__init__.py
!python -m torch.distributed.launch --nproc_per_node=1 --master_port=29930 /home/pai/lib/python3.6/site-packages/easycv/tools/train.py ./mae_vit_base_patch16_8xb64_1600e.py --work_dir work_dir/selfsup/jpg/mae --launcher pytorch

导出模型

模型训练完成,使用export命令导出模型进行推理,导出的模型包含推理时所需的预处理信息、后处理信息

# 查看训练产生的pt文件
! ls  work_dir/selfsup/jpg/mae/*
work_dir/selfsup/jpg/mae/20230202_150414.log
work_dir/selfsup/jpg/mae/20230202_150414.log.json
work_dir/selfsup/jpg/mae/epoch_5.pth

对模型的字段进行修改,以便用于fintune任务

import torch 
weight_path = 'work_dir/selfsup/jpg/mae/epoch_5.pth'
state_dict = torch.load(weight_path)['state_dict']
state_dict_out = {}
for key in state_dict:
    state_dict_out['model.' + key.replace('encoder.','')] = state_dict[key]
torch.save(state_dict_out,weight_path)
[2023-02-02 15:08:11,405.405 dsw-16577-76cf6dbbb4-m7pbn:1794 INFO utils.py:30] NOTICE: PAIDEBUGGER is turned off.

使用自监督模型进行图像分类fintune

下载分类任务示例配置文件

! rm -rf mae_vit_base_patch16_8xb64_100e_lrdecay065_fintune.py
! wget http://pai-vision-data-hz.oss-cn-zhangjiakou.aliyuncs.com/release/doc/easycv/configs/selfsup/mae/mae_vit_base_patch16_8xb64_100e_lrdecay065_fintune.py
Will not apply HSTS. The HSTS database must be a regular and non-world-writable file.
ERROR: could not open HSTS store at '/root/.wget-hsts'. HSTS will be disabled.
--2023-02-02 15:11:23--  http://pai-vision-data-hz.oss-cn-zhangjiakou.aliyuncs.com/release/doc/easycv/configs/selfsup/mae/mae_vit_base_patch16_8xb64_100e_lrdecay065_fintune.py
Resolving pai-vision-data-hz.oss-cn-zhangjiakou.aliyuncs.com (pai-vision-data-hz.oss-cn-zhangjiakou.aliyuncs.com)... 39.98.20.13
Connecting to pai-vision-data-hz.oss-cn-zhangjiakou.aliyuncs.com (pai-vision-data-hz.oss-cn-zhangjiakou.aliyuncs.com)|39.98.20.13|:80... connected.
HTTP request sent, awaiting response... 200 OK
Length: 4040 (3.9K) [text/x-python]
Saving to: ‘mae_vit_base_patch16_8xb64_100e_lrdecay065_fintune.py’
mae_vit_base_patch1 100%[===================>]   3.95K  --.-KB/s    in 0s      
2023-02-02 15:11:23 (189 MB/s) - ‘mae_vit_base_patch16_8xb64_100e_lrdecay065_fintune.py’ saved [4040/4040]

修改配置文件 mae_vit_base_patch16_8xb64_100e_lrdecay065_fintune.py,修改model.pretrained参数为work_dir/selfsup/jpg/mae/epoch_5.pth,为了缩短训练时间,修改total_epoch参数为5,每隔1次迭代打印一次日志。

# runtime settings
total_epochs = 5
# log config
log_config=dict(interval=1)
# 
pretrained='work_dir/selfsup/jpg/mae/epoch_5.pth'

正式训练时,建议使用多机8卡或单机8卡配合该配置文件使用,修改update_interval参数进行梯度累积,确保有效batch_size一致

分类模型训练

这里提供了单卡进行训练和验证集评估的命令

!python -m torch.distributed.launch --nproc_per_node=1 --master_port=29930 \
/home/pai/lib/python3.6/site-packages/easycv/tools/train.py mae_vit_base_patch16_8xb64_100e_lrdecay065_fintune.py --work_dir work_dir/selfsup/jpg/mae_fintune --launcher pytorch

预测

对训练好的模型导出并预测

! python -m easycv.tools.export mae_vit_base_patch16_8xb64_100e_lrdecay065_fintune.py  work_dir/selfsup/jpg/mae_fintune/ClsEvaluator_neck_top1_best.pth  work_dir/selfsup/jpg/mae_fintune/best_export.pth
[2023-02-02 15:20:34,506.506 dsw-16577-76cf6dbbb4-m7pbn:5559 INFO utils.py:30] NOTICE: PAIDEBUGGER is turned off.
mae_vit_base_patch16_8xb64_100e_lrdecay065_fintune.py
load model from init weights
load model from init weights
load checkpoint from local path: work_dir/selfsup/jpg/mae_fintune/ClsEvaluator_neck_top1_best.pth

下载测试图片和标签文件

! wget http://pai-vision-data-hz.oss-cn-zhangjiakou.aliyuncs.com/data/cifar10/qince_data/predict/aeroplane_s_000004.png
! wget http://pai-vision-data-hz.oss-cn-zhangjiakou.aliyuncs.com/release/doc/easycv/configs/selfsup/mae/label_map.txt
Will not apply HSTS. The HSTS database must be a regular and non-world-writable file.
ERROR: could not open HSTS store at '/root/.wget-hsts'. HSTS will be disabled.
--2023-02-02 15:21:19--  http://pai-vision-data-hz.oss-cn-zhangjiakou.aliyuncs.com/data/cifar10/qince_data/predict/aeroplane_s_000004.png
Resolving pai-vision-data-hz.oss-cn-zhangjiakou.aliyuncs.com (pai-vision-data-hz.oss-cn-zhangjiakou.aliyuncs.com)... 39.98.20.13
Connecting to pai-vision-data-hz.oss-cn-zhangjiakou.aliyuncs.com (pai-vision-data-hz.oss-cn-zhangjiakou.aliyuncs.com)|39.98.20.13|:80... connected.
HTTP request sent, awaiting response... 200 OK
Length: 2391 (2.3K) [image/png]
Saving to: ‘aeroplane_s_000004.png’
aeroplane_s_000004. 100%[===================>]   2.33K  --.-KB/s    in 0s      
2023-02-02 15:21:19 (186 MB/s) - ‘aeroplane_s_000004.png’ saved [2391/2391]
Will not apply HSTS. The HSTS database must be a regular and non-world-writable file.
ERROR: could not open HSTS store at '/root/.wget-hsts'. HSTS will be disabled.
--2023-02-02 15:21:20--  http://pai-vision-data-hz.oss-cn-zhangjiakou.aliyuncs.com/release/doc/easycv/configs/selfsup/mae/label_map.txt
Resolving pai-vision-data-hz.oss-cn-zhangjiakou.aliyuncs.com (pai-vision-data-hz.oss-cn-zhangjiakou.aliyuncs.com)... 39.98.20.13
Connecting to pai-vision-data-hz.oss-cn-zhangjiakou.aliyuncs.com (pai-vision-data-hz.oss-cn-zhangjiakou.aliyuncs.com)|39.98.20.13|:80... connected.
HTTP request sent, awaiting response... 200 OK
Length: 10473 (10K) [text/plain]
Saving to: ‘label_map.txt’
label_map.txt       100%[===================>]  10.23K  --.-KB/s    in 0s      
2023-02-02 15:21:20 (278 MB/s) - ‘label_map.txt’ saved [10473/10473]

导入模型权重,并预测测试图片的分类结果

import cv2
from easycv.predictors.classifier import TorchClassifier
output_ckpt = 'work_dir/selfsup/jpg/mae_fintune/best_export.pth'
tcls = TorchClassifier(output_ckpt, topk=1, label_map_path='label_map.txt')
img = cv2.imread('aeroplane_s_000004.png')
# input image should be RGB order
img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
output = tcls.predict([img])
print(output)
/home/pai/lib/python3.6/site-packages/ipykernel_launcher.py:5: DeprecationWarning: Call to deprecated class TorchClassifier (Please use ClassificationPredictor.).
  """
load model from init weights
load model from init weights
load checkpoint from local path: work_dir/selfsup/jpg/mae_fintune/best_export.pth
/home/pai/lib/python3.6/site-packages/torchvision/transforms/transforms.py:258: UserWarning: Argument interpolation should be of type InterpolationMode instead of int. Please, use InterpolationMode enum.
  "Argument interpolation should be of type InterpolationMode instead of int. "
[{'class': [701], 'class_name': ['prison'], 'class_probs': {'kit_fox': 0.0006673375, 'English_setter': 0.0006035781, 'Siberian_husky': 0.0008663863, 'Australian_terrier': 0.0019608254, 'English_springer': 0.0005221807, 'grey_whale': 0.00063806656, 'lesser_panda': 0.0003269123, 'Egyptian_cat': 0.0010239016, 'ibex': 0.0013876773, 'Persian_cat': 0.00087532273, 'cougar': 0.0005133071, 'gazelle': 0.0006251378, 'porcupine': 0.0011010804, 'sea_lion': 0.0009487947, 'malamute': 0.0008857793, 'badger': 0.00095850596, 'Great_Dane': 0.0005924807, 'Walker_hound': 0.0004092168, 'Welsh_springer_spaniel': 0.0003430144, 'whippet': 0.0012620587, 'Scottish_deerhound': 0.0009736246, 'killer_whale': 0.00077678234, 'mink': 0.0005921396, 'African_elephant': 0.0016753499, 'Weimaraner': 0.0004948448, 'soft-coated_wheaten_terrier': 0.0010146626, 'Dandie_Dinmont': 0.0007387661, 'red_wolf': 0.0007098433, 'Old_English_sheepdog': 0.0005576845, 'jaguar': 0.00085798727, 'otterhound': 0.0005155093, 'bloodhound': 0.000760222, 'Airedale': 0.0008089066, 'hyena': 0.00036170988, 'meerkat': 0.0009608063, 'giant_schnauzer': 0.0014625018, 'titi': 0.0022610286, 'three-toed_sloth': 0.00049388607, 'sorrel': 0.0009254522, 'black-footed_ferret': 0.0011746412, 'dalmatian': 0.0012251538, 'black-and-tan_coonhound': 0.00057177624, 'papillon': 0.0012972491, 'skunk': 0.0011283859, 'Staffordshire_bullterrier': 0.001174961, 'Mexican_hairless': 0.00096787966, 'Bouvier_des_Flandres': 0.0018662175, 'weasel': 0.0010575509, 'miniature_poodle': 0.0012047039, 'Cardigan': 0.00051354064, 'malinois': 0.00043676127, 'bighorn': 0.0011948944, 'fox_squirrel': 0.0005488761, 'colobus': 0.0004639502, 'tiger_cat': 0.00042578278, 'Lhasa': 0.00073431176, 'impala': 0.0008406444, 'coyote': 0.0009409984, 'Yorkshire_terrier': 0.00050984224, 'Newfoundland': 0.0030436, 'brown_bear': 0.0009942558, 'red_fox': 0.00023898148, 'Norwegian_elkhound': 0.0013350636, 'Rottweiler': 0.0010885244, 'hartebeest': 0.0016698914, 'Saluki': 0.0009344835, 'grey_fox': 0.0008787476, 'schipperke': 0.00034686414, 'Pekinese': 0.00076353835, 'Brabancon_griffon': 0.0006259315, 'West_Highland_white_terrier': 0.0007767992, 'Sealyham_terrier': 0.0003513658, 'guenon': 0.0007001311, 'mongoose': 0.0003822867, 'indri': 0.003609633, 'tiger': 0.0007440456, 'Irish_wolfhound': 0.0017718839, 'wild_boar': 0.002473396, 'EntleBucher': 0.0008174394, 'zebra': 0.0012578121, 'ram': 0.00082847854, 'French_bulldog': 0.0008900284, 'orangutan': 0.00032684745, 'basenji': 0.0009302878, 'leopard': 0.00027074857, 'Bernese_mountain_dog': 0.0012590293, 'Maltese_dog': 0.0006665097, 'Norfolk_terrier': 0.0008061021, 'toy_terrier': 0.00036666094, 'vizsla': 0.0014387047, 'cairn': 0.00055520685, 'squirrel_monkey': 0.000687506, 'groenendael': 0.000625328, 'clumber': 0.0029570605, 'Siamese_cat': 0.0005550988, 'chimpanzee': 0.00093030784, 'komondor': 0.0019317855, 'Afghan_hound': 0.00055170926, 'Japanese_spaniel': 0.00085775316, 'proboscis_monkey': 0.0006491716, 'guinea_pig': 0.0031287922, 'white_wolf': 0.00071867363, 'ice_bear': 0.000760691, 'gorilla': 0.00037043117, 'borzoi': 0.0007290872, 'toy_poodle': 0.0019134342, 'Kerry_blue_terrier': 0.00062811025, 'ox': 0.000470462, 'Scotch_terrier': 0.0006669656, 'Tibetan_mastiff': 0.00055876496, 'spider_monkey': 0.0010898418, 'Doberman': 0.0011695517, 'Boston_bull': 0.0012972043, 'Greater_Swiss_Mountain_dog': 0.00068416016, 'Appenzeller': 0.00067680137, 'Shih-Tzu': 0.00074696424, 'Irish_water_spaniel': 0.00037025844, 'Pomeranian': 0.00043323633, 'Bedlington_terrier': 0.000662382, 'warthog': 0.0002796068, 'Arabian_camel': 0.001094503, 'siamang': 0.00062231475, 'miniature_schnauzer': 0.00091091415, 'collie': 0.00041920063, 'golden_retriever': 0.00086782576, 'Irish_terrier': 0.0008741582, 'affenpinscher': 0.0010470334, 'Border_collie': 0.001344511, 'hare': 0.00068683823, 'boxer': 0.00092890306, 'silky_terrier': 0.00037482253, 'beagle': 0.00075919175, 'Leonberg': 0.0028384586, 'German_short-haired_pointer': 0.00039628547, 'patas': 0.0006132576, 'dhole': 0.0004852794, 'baboon': 0.0015588455, 'macaque': 0.00077748805, 'Chesapeake_Bay_retriever': 0.00026608192, 'bull_mastiff': 0.0005500803, 'kuvasz': 0.0004735009, 'capuchin': 0.00045627708, 'pug': 0.0004272182, 'curly-coated_retriever': 0.00088318466, 'Norwich_terrier': 0.00073626626, 'flat-coated_retriever': 0.0012534767, 'hog': 0.00049689575, 'keeshond': 0.0029105514, 'Eskimo_dog': 0.0008534481, 'Brittany_spaniel': 0.0008315074, 'standard_poodle': 0.0007357677, 'Lakeland_terrier': 0.001161858, 'snow_leopard': 0.0007181353, 'Gordon_setter': 0.00063368015, 'dingo': 0.00085809594, 'standard_schnauzer': 0.0010800055, 'hamster': 0.0010458771, 'Tibetan_terrier': 0.0008203154, 'Arctic_fox': 0.0014584167, 'wire-haired_fox_terrier': 0.0010811656, 'basset': 0.0007325018, 'water_buffalo': 0.0022239466, 'American_black_bear': 0.00095368, 'Angora': 0.0005397964, 'bison': 0.0014236068, 'howler_monkey': 0.00083370507, 'hippopotamus': 0.0030128977, 'chow': 0.0010758714, 'giant_panda': 0.00084381434, 'American_Staffordshire_terrier': 0.00056128606, 'Shetland_sheepdog': 0.00031425315, 'Great_Pyrenees': 0.0009117502, 'Chihuahua': 0.00039096142, 'tabby': 0.0012651962, 'marmoset': 0.0011296106, 'Labrador_retriever': 0.000810957, 'Saint_Bernard': 0.0005878524, 'armadillo': 0.0009452825, 'Samoyed': 0.0007942208, 'bluetick': 0.0004587803, 'redbone': 0.00033727402, 'polecat': 0.0005682515, 'marmot': 0.00049401016, 'kelpie': 0.00050273834, 'gibbon': 0.0012528409, 'llama': 0.00028378292, 'miniature_pinscher': 0.00028555433, 'wood_rabbit': 0.00045620077, 'Italian_greyhound': 0.0017857451, 'lion': 0.0017187167, 'cocker_spaniel': 0.0011800345, 'Irish_setter': 0.00037181537, 'dugong': 0.0004985526, 'Indian_elephant': 0.0038731324, 'beaver': 0.00081938715, 'Sussex_spaniel': 0.0010430053, 'Pembroke': 0.00035126496, 'Blenheim_spaniel': 0.001339322, 'Madagascar_cat': 0.0006958304, 'Rhodesian_ridgeback': 0.00044000885, 'lynx': 0.0007535317, 'African_hunting_dog': 0.00041206204, 'langur': 0.0014016775, 'Ibizan_hound': 0.0012293942, 'timber_wolf': 0.0012037078, 'cheetah': 0.0013392578, 'English_foxhound': 0.00049555744, 'briard': 0.0005603474, 'sloth_bear': 0.00067047833, 'Border_terrier': 0.0011209487, 'German_shepherd': 0.0003471842, 'otter': 0.00074742624, 'koala': 0.0012796782, 'tusker': 0.0012231773, 'echidna': 0.0006569209, 'wallaby': 0.0017348453, 'platypus': 0.00084433006, 'wombat': 0.0012885607, 'revolver': 0.00032871883, 'umbrella': 0.000562244, 'schooner': 0.000296088, 'soccer_ball': 0.0003618831, 'accordion': 0.0009141253, 'ant': 0.0011580512, 'starfish': 0.0003766985, 'chambered_nautilus': 0.0011685746, 'grand_piano': 0.0005815985, 'laptop': 0.0010458522, 'strawberry': 0.0011316275, 'airliner': 0.0015322901, 'warplane': 0.0012332336, 'airship': 0.00031386214, 'balloon': 0.00082097185, 'space_shuttle': 0.00053222646, 'fireboat': 0.0007481256, 'gondola': 0.0012073473, 'speedboat': 0.0018494261, 'lifeboat': 0.00048196854, 'canoe': 0.0004692828, 'yawl': 0.0025077173, 'catamaran': 0.0009199604, 'trimaran': 0.00047925056, 'container_ship': 0.0008788729, 'liner': 0.00045518728, 'pirate': 0.0008025049, 'aircraft_carrier': 0.0014064736, 'submarine': 0.00094974844, 'wreck': 0.0009211716, 'half_track': 0.0025719875, 'tank': 0.0010164235, 'missile': 0.00096704747, 'bobsled': 0.0009352393, 'dogsled': 0.00089393975, 'bicycle-built-for-two': 0.0006481305, 'mountain_bike': 0.00044295142, 'freight_car': 0.00072417833, 'passenger_car': 0.0010054957, 'barrow': 0.0011820616, 'shopping_cart': 0.0006315327, 'motor_scooter': 0.0012318507, 'forklift': 0.0008912859, 'electric_locomotive': 0.00043462266, 'steam_locomotive': 0.0014380314, 'amphibian': 0.0013264907, 'ambulance': 0.0008022707, 'beach_wagon': 0.0006814724, 'cab': 0.0011039875, 'convertible': 0.00026016115, 'jeep': 0.00195107, 'limousine': 0.0006085943, 'minivan': 0.001311419, 'Model_T': 0.002430188, 'racer': 0.0003428685, 'sports_car': 0.00050586375, 'go-kart': 0.0015384373, 'golfcart': 0.0004850899, 'moped': 0.0019142939, 'snowplow': 0.0006055381, 'fire_engine': 0.0013242919, 'garbage_truck': 0.0007222062, 'pickup': 0.001595455, 'tow_truck': 0.00079983054, 'trailer_truck': 0.0008890617, 'moving_van': 0.00082359574, 'police_van': 0.00041793904, 'recreational_vehicle': 0.0009667033, 'streetcar': 0.00074825884, 'snowmobile': 0.0039802683, 'tractor': 0.00093732105, 'mobile_home': 0.0007239634, 'tricycle': 0.0012059201, 'unicycle': 0.0009181362, 'horse_cart': 0.00055886596, 'jinrikisha': 0.0016387817, 'oxcart': 0.00031353126, 'bassinet': 0.0007486063, 'cradle': 0.0008687476, 'crib': 0.0018985302, 'four-poster': 0.00026404974, 'bookcase': 0.0013073584, 'china_cabinet': 0.00107492, 'medicine_chest': 0.000447954, 'chiffonier': 0.0007046773, 'table_lamp': 0.000543701, 'file': 0.0006581826, 'park_bench': 0.0013155353, 'barber_chair': 0.00076055666, 'throne': 0.0005820926, 'folding_chair': 0.0035606618, 'rocking_chair': 0.00017051968, 'studio_couch': 0.0009297978, 'toilet_seat': 0.00060623296, 'desk': 0.00061987026, 'pool_table': 0.0008919749, 'dining_table': 0.0009493097, 'entertainment_center': 0.0005797904, 'wardrobe': 0.0020887179, 'Granny_Smith': 0.000841478, 'orange': 0.0001226445, 'lemon': 0.0016463246, 'fig': 0.0005141846, 'pineapple': 0.0013097357, 'banana': 0.00058463984, 'jackfruit': 0.0008032825, 'custard_apple': 0.0008301924, 'pomegranate': 0.0010332471, 'acorn': 0.0007280805, 'hip': 0.0016770704, 'ear': 0.00047219478, 'rapeseed': 0.0020728048, 'corn': 0.0011806451, 'buckeye': 0.00035773474, 'organ': 0.0011563181, 'upright': 0.00040048748, 'chime': 0.0006850233, 'drum': 0.0003219981, 'gong': 0.00056942843, 'maraca': 0.00038401593, 'marimba': 0.0016179933, 'steel_drum': 0.00038199915, 'banjo': 0.0014720429, 'cello': 0.00046661607, 'violin': 0.0008638141, 'harp': 0.00078270026, 'acoustic_guitar': 0.0011032243, 'electric_guitar': 0.0003665835, 'cornet': 0.000872279, 'French_horn': 0.0017834828, 'trombone': 0.0004510387, 'harmonica': 0.00048546432, 'ocarina': 0.00041573335, 'panpipe': 0.0005880968, 'bassoon': 0.0017976358, 'oboe': 0.00079598906, 'sax': 0.0014975878, 'flute': 0.0008456801, 'daisy': 0.0007133399, "yellow_lady's_slipper": 0.0005282468, 'cliff': 0.00091643014, 'valley': 0.0012302834, 'alp': 0.00025667544, 'volcano': 0.0014490306, 'promontory': 0.0008656905, 'sandbar': 0.0014829241, 'coral_reef': 0.0010663585, 'lakeside': 0.0008574207, 'seashore': 0.0018335097, 'geyser': 0.0019565504, 'hatchet': 0.0005222706, 'cleaver': 0.00068574655, 'letter_opener': 0.0003563477, 'plane': 0.0009281646, 'power_drill': 0.00061272416, 'lawn_mower': 0.0007980305, 'hammer': 0.0011221775, 'corkscrew': 0.0017910097, 'can_opener': 0.0018746562, 'plunger': 0.0005097243, 'screwdriver': 0.0012582233, 'shovel': 0.0004525653, 'plow': 0.0006649758, 'chain_saw': 0.0006513588, 'cock': 0.0013244139, 'hen': 0.0014041583, 'ostrich': 0.0009083533, 'brambling': 0.00051652134, 'goldfinch': 0.0006296056, 'house_finch': 0.00047804936, 'junco': 0.0014165796, 'indigo_bunting': 0.0028740333, 'robin': 0.0007866315, 'bulbul': 0.001917282, 'jay': 0.00076105696, 'magpie': 0.0013161178, 'chickadee': 0.00085238036, 'water_ouzel': 0.0002879172, 'kite': 0.00062900217, 'bald_eagle': 0.0010329705, 'vulture': 0.00042114107, 'great_grey_owl': 0.0026120266, 'black_grouse': 0.0007101799, 'ptarmigan': 0.00026333507, 'ruffed_grouse': 0.000757097, 'prairie_chicken': 0.00075126265, 'peacock': 0.00057938276, 'quail': 0.0020931046, 'partridge': 0.0023236645, 'African_grey': 0.0006997198, 'macaw': 0.0003764595, 'sulphur-crested_cockatoo': 0.00051658193, 'lorikeet': 0.001089204, 'coucal': 0.0006429889, 'bee_eater': 0.0005939811, 'hornbill': 0.00042015742, 'hummingbird': 0.0012699981, 'jacamar': 0.00038460435, 'toucan': 0.0012428985, 'drake': 0.0004888178, 'red-breasted_merganser': 0.001494211, 'goose': 0.00026711606, 'black_swan': 0.00068730576, 'white_stork': 0.0010806646, 'black_stork': 0.00054795877, 'spoonbill': 0.0007178469, 'flamingo': 0.0007972173, 'American_egret': 0.0019316667, 'little_blue_heron': 0.001196161, 'bittern': 0.0017733854, 'crane': 0.0010716481, 'limpkin': 0.00054351514, 'American_coot': 0.0005197537, 'bustard': 0.0008223057, 'ruddy_turnstone': 0.00079439674, 'red-backed_sandpiper': 0.0007670702, 'redshank': 0.00090661953, 'dowitcher': 0.00044888392, 'oystercatcher': 0.0027146111, 'European_gallinule': 0.00105227, 'pelican': 0.00041989624, 'king_penguin': 0.0013082908, 'albatross': 0.000823284, 'great_white_shark': 0.00095469254, 'tiger_shark': 0.0007659887, 'hammerhead': 0.00034065126, 'electric_ray': 0.00046376462, 'stingray': 0.003160964, 'barracouta': 0.0005162672, 'coho': 0.00077205466, 'tench': 0.00026017704, 'goldfish': 0.0006990679, 'eel': 0.0005111648, 'rock_beauty': 0.00080139533, 'anemone_fish': 0.0005497227, 'lionfish': 0.0005297166, 'puffer': 0.0003801149, 'sturgeon': 0.0008049187, 'gar': 0.0004115417, 'loggerhead': 0.0009446465, 'leatherback_turtle': 0.0016841586, 'mud_turtle': 0.0018437328, 'terrapin': 0.0021972153, 'box_turtle': 0.00040449982, 'banded_gecko': 0.0018808753, 'common_iguana': 0.0007888826, 'American_chameleon': 0.00079707993, 'whiptail': 0.00044323158, 'agama': 0.0006787146, 'frilled_lizard': 0.00070589053, 'alligator_lizard': 0.0007762472, 'Gila_monster': 0.00093608676, 'green_lizard': 0.0021551407, 'African_chameleon': 0.0002796997, 'Komodo_dragon': 0.00046416905, 'triceratops': 0.0020266527, 'African_crocodile': 0.0013681892, 'American_alligator': 0.0003778704, 'thunder_snake': 0.0013599838, 'ringneck_snake': 0.0007473412, 'hognose_snake': 0.00066585944, 'green_snake': 0.00066184276, 'king_snake': 0.0002789968, 'garter_snake': 0.0004152252, 'water_snake': 0.0014960772, 'vine_snake': 0.001124733, 'night_snake': 0.0006035525, 'boa_constrictor': 0.00072426005, 'rock_python': 0.0010061796, 'Indian_cobra': 0.000972281, 'green_mamba': 0.0010819746, 'sea_snake': 0.0006987462, 'horned_viper': 0.0012374495, 'diamondback': 0.0011098131, 'sidewinder': 0.00096575415, 'European_fire_salamander': 0.0008947692, 'common_newt': 0.00032571264, 'eft': 0.0008172559, 'spotted_salamander': 0.0018453375, 'axolotl': 0.0011403434, 'bullfrog': 0.0008467938, 'tree_frog': 0.0015964701, 'tailed_frog': 0.0011335404, 'whistle': 0.0016750292, 'wing': 0.0002675159, 'paintbrush': 0.0007092226, 'hand_blower': 0.0009814496, 'oxygen_mask': 0.0005475067, 'snorkel': 0.0032566949, 'loudspeaker': 0.0009572188, 'microphone': 0.00074006227, 'screen': 0.0023267374, 'mouse': 0.0009983559, 'electric_fan': 0.002048344, 'oil_filter': 0.0013005986, 'strainer': 0.00050701346, 'space_heater': 0.0007621984, 'stove': 0.00046690213, 'guillotine': 0.0012463164, 'barometer': 0.0006752621, 'rule': 0.003495674, 'odometer': 0.0009359415, 'scale': 0.00031048447, 'analog_clock': 0.00077728427, 'digital_clock': 0.00041153363, 'wall_clock': 0.0014889947, 'hourglass': 0.0008962763, 'sundial': 0.0010768155, 'parking_meter': 0.0008074726, 'stopwatch': 0.000351166, 'digital_watch': 0.00034004668, 'stethoscope': 0.00074013066, 'syringe': 0.0011007161, 'magnetic_compass': 0.0015685349, 'binoculars': 0.0005992295, 'projector': 0.001144601, 'sunglasses': 0.0011693134, 'loupe': 0.000992853, 'radio_telescope': 0.0009821221, 'bow': 0.0003284918, 'cannon': 0.0018880258, 'assault_rifle': 0.0011750954, 'rifle': 0.0006499053, 'projectile': 0.00074486097, 'computer_keyboard': 0.00048503975, 'typewriter_keyboard': 0.0012519436, 'lighter': 0.00097056787, 'abacus': 0.00044205255, 'cash_machine': 0.0005575715, 'slide_rule': 0.0008857869, 'desktop_computer': 0.0005085318, 'hand-held_computer': 0.000819174, 'notebook': 0.00052175234, 'web_site': 0.00091285346, 'harvester': 0.0017603951, 'thresher': 0.0014266062, 'printer': 0.0017153149, 'slot': 0.0007183797, 'vending_machine': 0.00067375985, 'sewing_machine': 0.0010078662, 'joystick': 0.00035324582, 'switch': 0.00062357594, 'hook': 0.00021052865, 'car_wheel': 0.0029452974, 'paddlewheel': 0.0017677146, 'pinwheel': 0.00078193774, "potter's_wheel": 0.00057588087, 'gas_pump': 0.0012868207, 'carousel': 0.00087172474, 'swing': 0.0005312612, 'reel': 0.00079802464, 'radiator': 0.0006469106, 'puck': 0.000825484, 'hard_disc': 0.00039842707, 'sunglass': 0.00037853874, 'pick': 0.0011971801, 'car_mirror': 0.0015085902, 'solar_dish': 0.0008647065, 'remote_control': 0.0010237267, 'disk_brake': 0.0010853297, 'buckle': 0.0003961923, 'hair_slide': 0.0011131054, 'knot': 0.0002489634, 'combination_lock': 0.00045069132, 'padlock': 0.0006865642, 'nail': 0.0009666945, 'safety_pin': 0.00082895433, 'screw': 0.000519691, 'muzzle': 0.0016549203, 'seat_belt': 0.00044220942, 'ski': 0.0004584521, 'candle': 0.0009548372, "jack-o'-lantern": 0.0009540223, 'spotlight': 0.0015289737, 'torch': 0.0009064998, 'neck_brace': 0.0010379179, 'pier': 0.00093379506, 'tripod': 0.0004079615, 'maypole': 0.0008402531, 'mousetrap': 0.00092358957, 'spider_web': 0.00036225514, 'trilobite': 0.0030412993, 'harvestman': 0.00038505578, 'scorpion': 0.0008469358, 'black_and_gold_garden_spider': 0.0014735492, 'barn_spider': 0.0003356832, 'garden_spider': 0.00054690865, 'black_widow': 0.0012433381, 'tarantula': 0.00065229845, 'wolf_spider': 0.0004580926, 'tick': 0.00067108404, 'centipede': 0.000753112, 'isopod': 0.0008716396, 'Dungeness_crab': 0.0006183142, 'rock_crab': 0.00035523614, 'fiddler_crab': 0.0006676319, 'king_crab': 0.00036002512, 'American_lobster': 0.00039007002, 'spiny_lobster': 0.0013909374, 'crayfish': 0.0017018714, 'hermit_crab': 0.000498715, 'tiger_beetle': 0.000639993, 'ladybug': 0.00085118436, 'ground_beetle': 0.00031612744, 'long-horned_beetle': 0.00030696206, 'leaf_beetle': 0.00047393283, 'dung_beetle': 0.0007297292, 'rhinoceros_beetle': 0.00061509, 'weevil': 0.00046350397, 'fly': 0.0014952285, 'bee': 0.0012114372, 'grasshopper': 0.0011853548, 'cricket': 0.0005844737, 'walking_stick': 0.00040300237, 'cockroach': 0.00073412363, 'mantis': 0.00063629827, 'cicada': 0.00022824801, 'leafhopper': 0.00027522002, 'lacewing': 0.0013366669, 'dragonfly': 0.0024538569, 'damselfly': 0.00022150888, 'admiral': 0.0020160535, 'ringlet': 0.0018166861, 'monarch': 0.00076583313, 'cabbage_butterfly': 0.00077342405, 'sulphur_butterfly': 0.0011847485, 'lycaenid': 0.0006704226, 'jellyfish': 0.00081835577, 'sea_anemone': 0.0011766598, 'brain_coral': 0.00046982957, 'flatworm': 0.0012634732, 'nematode': 0.0009636003, 'conch': 0.0006142067, 'snail': 0.00017130001, 'slug': 0.00075874606, 'sea_slug': 0.0014216286, 'chiton': 0.0006182942, 'sea_urchin': 0.00017237256, 'sea_cucumber': 0.0010081974, 'iron': 0.00018330582, 'espresso_maker': 0.0037977756, 'microwave': 0.0005525739, 'Dutch_oven': 0.0006398203, 'rotisserie': 0.0007416046, 'toaster': 0.0030331607, 'waffle_iron': 0.0006374371, 'vacuum': 0.001281261, 'dishwasher': 0.0011918882, 'refrigerator': 0.00090237235, 'washer': 0.0014799209, 'Crock_Pot': 0.001018635, 'frying_pan': 0.0004062346, 'wok': 0.00057704386, 'caldron': 0.00059988076, 'coffeepot': 0.00048700083, 'teapot': 0.0010013032, 'spatula': 0.0004927666, 'altar': 0.0022658801, 'triumphal_arch': 0.0008646381, 'patio': 0.001307513, 'steel_arch_bridge': 0.0009046664, 'suspension_bridge': 0.0022883005, 'viaduct': 0.0013950545, 'barn': 0.0011256731, 'greenhouse': 0.0006194669, 'palace': 0.0013485613, 'monastery': 0.00064643164, 'library': 0.0012675138, 'apiary': 0.0012780542, 'boathouse': 0.00061652093, 'church': 0.0004499872, 'mosque': 0.001469035, 'stupa': 0.001939639, 'planetarium': 0.0028417956, 'restaurant': 0.0005366011, 'cinema': 0.0008090114, 'home_theater': 0.00051942613, 'lumbermill': 0.0010888351, 'coil': 0.00062355393, 'obelisk': 0.00052302075, 'totem_pole': 0.0006346021, 'castle': 0.0010238918, 'prison': 0.037591904, 'grocery_store': 0.00050340284, 'bakery': 0.00178884, 'barbershop': 0.00057094253, 'bookshop': 0.0008253245, 'butcher_shop': 0.0006958083, 'confectionery': 0.0010081554, 'shoe_shop': 0.0017710367, 'tobacco_shop': 0.0005890935, 'toyshop': 0.00059027725, 'fountain': 0.00065871666, 'cliff_dwelling': 0.00040663886, 'yurt': 0.0013837194, 'dock': 0.0012450936, 'brass': 0.0027852838, 'megalith': 0.00071522524, 'bannister': 0.0013596058, 'breakwater': 0.00011946146, 'dam': 0.0016865965, 'chainlink_fence': 0.0014193305, 'picket_fence': 0.0011133159, 'worm_fence': 0.00082927453, 'stone_wall': 0.0011926274, 'grille': 0.00087253674, 'sliding_door': 0.0014886255, 'turnstile': 0.0012448393, 'mountain_tent': 0.00038131158, 'scoreboard': 0.00053523394, 'honeycomb': 0.0072708465, 'plate_rack': 0.0011068652, 'pedestal': 0.00028404867, 'beacon': 0.003178685, 'mashed_potato': 0.0018442032, 'bell_pepper': 0.00056632253, 'head_cabbage': 0.00058140384, 'broccoli': 0.00029479896, 'cauliflower': 0.0010468279, 'zucchini': 0.0003000654, 'spaghetti_squash': 0.0010554418, 'acorn_squash': 0.0004984984, 'butternut_squash': 0.0020677387, 'cucumber': 0.0014997541, 'artichoke': 0.0003227595, 'cardoon': 0.0022013949, 'mushroom': 0.0008607164, 'shower_curtain': 0.0004795841, 'jean': 0.0005667199, 'carton': 0.0030431573, 'handkerchief': 0.0015435346, 'sandal': 0.00075286604, 'ashcan': 0.000556034, 'safe': 0.00038855596, 'plate': 0.002078893, 'necklace': 0.0026141165, 'croquet_ball': 0.0010774418, 'fur_coat': 0.00043184432, 'thimble': 0.0005171653, 'pajama': 0.0006903484, 'running_shoe': 0.0010900572, 'cocktail_shaker': 0.00047785995, 'chest': 0.00052981183, 'manhole_cover': 0.00081432355, 'modem': 0.00051041314, 'tub': 0.0016273508, 'tray': 0.001271315, 'balance_beam': 0.00054115406, 'bagel': 0.00081605726, 'prayer_rug': 0.0009082688, 'kimono': 0.0018243879, 'hot_pot': 0.000492199, 'whiskey_jug': 0.0010275064, 'knee_pad': 0.00037368093, 'book_jacket': 0.0007844238, 'spindle': 0.0015724301, 'ski_mask': 0.0010130922, 'beer_bottle': 0.0008288061, 'crash_helmet': 0.0009783998, 'bottlecap': 0.00051618373, 'tile_roof': 0.00029629504, 'mask': 0.0009694052, 'maillot': 0.0011045595, 'Petri_dish': 0.0010503475, 'football_helmet': 0.00054752803, 'bathing_cap': 0.00091218116, 'teddy': 0.0016758133, 'holster': 0.00057679816, 'pop_bottle': 0.00042593875, 'photocopier': 0.00086837035, 'vestment': 0.00033034515, 'crossword_puzzle': 0.0010363212, 'golf_ball': 0.0006904559, 'trifle': 0.0010752911, 'suit': 0.00048054964, 'water_tower': 0.0005003764, 'feather_boa': 0.0007605053, 'cloak': 0.00024744385, 'red_wine': 0.00067424827, 'drumstick': 0.00065143243, 'shield': 0.00052620575, 'Christmas_stocking': 0.0008204674, 'hoopskirt': 0.0008784633, 'menu': 0.0007362759, 'stage': 0.0006689148, 'bonnet': 0.0009185699, 'meat_loaf': 0.0007909464, 'baseball': 0.004425101, 'face_powder': 0.0005708383, 'scabbard': 0.0015747502, 'sunscreen': 0.0009909504, 'beer_glass': 0.002912529, 'hen-of-the-woods': 0.0008806184, 'guacamole': 0.0005704701, 'lampshade': 0.000777545, 'wool': 0.0012973505, 'hay': 0.0008455041, 'bow_tie': 0.00035972652, 'mailbag': 0.00040281023, 'water_jug': 0.0005977341, 'bucket': 0.0010246082, 'dishrag': 0.00057096954, 'soup_bowl': 0.0012191479, 'eggnog': 0.0010109239, 'mortar': 0.0004677529, 'trench_coat': 0.0011158193, 'paddle': 0.0011707654, 'chain': 0.0005027149, 'swab': 0.0008694847, 'mixing_bowl': 0.00073785475, 'potpie': 0.00094167195, 'wine_bottle': 0.00041367978, 'shoji': 0.00072146737, 'bulletproof_vest': 0.001549902, 'drilling_platform': 0.00095722085, 'binder': 0.0006494586, 'cardigan': 0.0003860092, 'sweatshirt': 0.00061620853, 'pot': 0.0010708579, 'birdhouse': 0.0006192731, 'hamper': 0.000680508, 'ping-pong_ball': 0.00036742087, 'pencil_box': 0.0008063758, 'pay-phone': 0.00020054205, 'consomme': 0.00063226453, 'apron': 0.0013311043, 'punching_bag': 0.0012086808, 'backpack': 0.0010986532, 'groom': 0.0010530879, 'bearskin': 0.0023114928, 'pencil_sharpener': 0.00031419293, 'broom': 0.000524321, 'mosquito_net': 0.0009460951, 'abaya': 0.00058811624, 'mortarboard': 0.0005296005, 'poncho': 0.00033354198, 'crutch': 0.0018589825, 'Polaroid_camera': 0.0073109074, 'space_bar': 0.00091180863, 'cup': 0.0004914166, 'racket': 0.0006037721, 'traffic_light': 0.00032293762, 'quill': 0.0005353401, 'radio': 0.0008333485, 'dough': 0.0017043498, 'cuirass': 0.0005229454, 'military_uniform': 0.0009487664, 'lipstick': 0.00076575374, 'shower_cap': 0.0013812194, 'monitor': 0.0011247786, 'oscilloscope': 0.000900707, 'mitten': 0.0005356782, 'brassiere': 0.00084054185, 'French_loaf': 0.0004829366, 'vase': 0.00054381037, 'milk_can': 0.0012251097, 'rugby_ball': 0.0007054033, 'paper_towel': 0.0014432722, 'earthstar': 0.001632474, 'envelope': 0.00088640035, 'miniskirt': 0.00090875855, 'cowboy_hat': 0.00039142516, 'trolleybus': 0.0003915289, 'perfume': 0.00051921565, 'bathtub': 0.00015694281, 'hotdog': 0.0006433339, 'coral_fungus': 0.0005394452, 'bullet_train': 0.0022107903, 'pillow': 0.00051732216, 'toilet_tissue': 0.0014297587, 'cassette': 0.0012342895, "carpenter's_kit": 0.00074536796, 'ladle': 0.0008255416, 'stinkhorn': 0.000737084, 'lotion': 0.0012245589, 'hair_spray': 0.00036913148, 'academic_gown': 0.0015320709, 'dome': 0.00074378395, 'crate': 0.00083024806, 'wig': 0.0013755367, 'burrito': 0.0011630281, 'pill_bottle': 0.0017772405, 'chain_mail': 0.002070553, 'theater_curtain': 0.0011759119, 'window_shade': 0.0009147061, 'barrel': 0.00052829145, 'washbasin': 0.0017775083, 'ballpoint': 0.00044607572, 'basketball': 0.0013331033, 'bath_towel': 0.0007362566, 'cowboy_boot': 0.00046114664, 'gown': 0.00095460087, 'window_screen': 0.00084049394, 'agaric': 0.0011892238, 'cellular_telephone': 0.00034753102, 'nipple': 0.0016348582, 'barbell': 0.0041211806, 'mailbox': 0.000823139, 'lab_coat': 0.0010116922, 'fire_screen': 0.0009556892, 'minibus': 0.00033872112, 'packet': 0.00038947322, 'maze': 0.0006725029, 'pole': 0.0014283361, 'horizontal_bar': 0.0008366088, 'sombrero': 0.00040451752, 'pickelhaube': 0.0014557278, 'rain_barrel': 0.0014428993, 'wallet': 0.00064433966, 'cassette_player': 0.00065501226, 'comic_book': 0.0010568693, 'piggy_bank': 0.00032782956, 'street_sign': 0.0021055064, 'bell_cote': 0.0009688309, 'fountain_pen': 0.00056377694, 'Windsor_tie': 0.00040692632, 'volleyball': 0.0010836242, 'overskirt': 0.00085448404, 'sarong': 0.0002908299, 'purse': 0.0005682637, 'bolo_tie': 0.0005429458, 'bib': 0.0019812153, 'parachute': 0.00029132568, 'sleeping_bag': 0.0008590762, 'television': 0.0003643546, 'swimming_trunks': 0.0024170724, 'measuring_cup': 0.00039348195, 'espresso': 0.00026041383, 'pizza': 0.0011925868, 'breastplate': 0.000664609, 'shopping_basket': 0.0005954881, 'wooden_spoon': 0.00054428336, 'saltshaker': 0.001431945, 'chocolate_sauce': 0.00071529514, 'ballplayer': 0.0006756751, 'goblet': 0.0007100134, 'gyromitra': 0.0009879427, 'stretcher': 0.000595041, 'water_bottle': 0.0008173642, 'dial_telephone': 0.00023926221, 'soap_dispenser': 0.00128625, 'jersey': 0.00069372286, 'school_bus': 0.00029718337, 'jigsaw_puzzle': 0.00071380194, 'plastic_bag': 0.0016863031, 'reflex_camera': 0.0016982144, 'diaper': 0.0012433642, 'Band_Aid': 0.0010888961, 'ice_lolly': 0.0010593379, 'velvet': 0.0004295196, 'tennis_ball': 0.0009705796, 'gasmask': 0.0011191316, 'doormat': 0.0015863173, 'Loafer': 0.0010363159, 'ice_cream': 0.00028664025, 'pretzel': 0.00067002926, 'quilt': 0.002805399, 'tape_player': 0.0010363003, 'clog': 0.00045157672, 'iPod': 0.001187576, 'bolete': 0.0016062711, 'scuba_diver': 0.00035697714, 'pitcher': 0.0008663289, 'matchstick': 0.0006115164, 'bikini': 0.0009137165, 'sock': 0.0010276816, 'CD_player': 0.0011872674, 'lens_cap': 0.0027006543, 'thatch': 0.00030401116, 'vault': 0.0009730892, 'beaker': 0.00028762178, 'bubble': 0.00094937376, 'cheeseburger': 0.0004585568, 'parallel_bars': 0.0005167571, 'flagpole': 0.0017233518, 'coffee_mug': 0.00042655034, 'rubber_eraser': 0.00073954894, 'stole': 0.0011692847, 'carbonara': 0.00083475123, 'dumbbell': 0.00044254755}}]


相关实践学习
使用PAI-EAS一键部署ChatGLM及LangChain应用
本场景中主要介绍如何使用模型在线服务(PAI-EAS)部署ChatGLM的AI-Web应用以及启动WebUI进行模型推理,并通过LangChain集成自己的业务数据。
机器学习概览及常见算法
机器学习(Machine Learning, ML)是人工智能的核心,专门研究计算机怎样模拟或实现人类的学习行为,以获取新的知识或技能,重新组织已有的知识结构使之不断改善自身的性能,它是使计算机具有智能的根本途径,其应用遍及人工智能的各个领域。 本课程将带你入门机器学习,掌握机器学习的概念和常用的算法。
相关文章
|
消息中间件 运维 算法
【DSW Gallery】IsolationForest算法解决异常检测问题
IsolationForest 是一种无监督的异常检测算法, 用于对无 label 的数据进行异常检测,并且支持将 IsolationForest 模型部署成一个流服务,用来对实时数据进行异常检测。该 Demo 将介绍如何在 DSW 中使用 IsolationForest 算法解决异常检测问题。
【DSW Gallery】IsolationForest算法解决异常检测问题
|
机器学习/深度学习 算法
【DSW Gallery】如何使用EasyRec训练DeepFM模型
本文基于EasyRec 0.4.7 展示了如何使用EasyRec快速的训练一个DeepFM模型
【DSW Gallery】如何使用EasyRec训练DeepFM模型
|
人工智能 并行计算 算法
【DSW Gallery】基于MOCOV2的自监督学习示例
EasyCV是基于Pytorch,以自监督学习和Transformer技术为核心的 all-in-one 视觉算法建模工具,并包含图像分类,度量学习,目标检测,姿态识别等视觉任务的SOTA算法。本文以自监督学习-MOCO为例,为您介绍如何在PAI-DSW中使用EasyCV。
【DSW Gallery】基于MOCOV2的自监督学习示例
|
算法 PyTorch 算法框架/工具
【DSW Gallery】基于EasyCV的视频分类示例
EasyCV是基于Pytorch,以自监督学习和Transformer技术为核心的 all-in-one 视觉算法建模工具,并包含图像分类,度量学习,目标检测,姿态识别等视觉任务的SOTA算法。本文以视频分类为例,为您介绍如何在PAI-DSW中使用EasyCV。
【DSW Gallery】基于EasyCV的视频分类示例
|
机器学习/深度学习 并行计算 数据可视化
【DSW Gallery】EasyCV-基于关键点的视频分类示例
EasyCV是基于Pytorch,以自监督学习和Transformer技术为核心的 all-in-one 视觉算法建模工具,并包含图像分类,度量学习,目标检测,姿态识别等视觉任务的SOTA算法。本文以基于关键点的视频分类为例,为您介绍如何在PAI-DSW中使用EasyCV。
【DSW Gallery】EasyCV-基于关键点的视频分类示例
|
消息中间件 运维 算法
【DSW Gallery】OneClassSVM 算法解决异常检测问题
OneClassSVM 是一种无监督的异常检测算法, 用于对无 label 的数据进行异常检测,并且支持将 OneClassSVM 模型部署成一个流服务,用来对实时数据进行异常检测。该D emo 将介绍如何在 DSW 中使用 OneClassSVM 算法解决异常检测问题。
【DSW Gallery】OneClassSVM 算法解决异常检测问题
|
并行计算 算法 自动驾驶
【DSW Gallery】基于EasyCV的BEVFormer 3D检测示例
EasyCV是基于Pytorch,以自监督学习和Transformer技术为核心的 all-in-one 视觉算法建模工具,并包含图像分类,度量学习,目标检测,姿态识别等视觉任务的SOTA算法。本文将以BEVFormer 3D检测为例,为您介绍如何在PAI-DSW中使用EasyCV。
【DSW Gallery】基于EasyCV的BEVFormer 3D检测示例
|
算法 PyTorch 算法框架/工具
【DSW Gallery】基于YOLOX模型和iTAG标注数据的图像检测示例
EasyCV是基于Pytorch,以自监督学习和Transformer技术为核心的 all-in-one 视觉算法建模工具,并包含图像分类,度量学习,目标检测,姿态识别等视觉任务的SOTA算法。本文将为您介绍如何在PAI-DSW中使用EasyCV和PAI-iTAG标注的检测数据训练YOLOX模型。
【DSW Gallery】基于YOLOX模型和iTAG标注数据的图像检测示例
|
机器学习/深度学习 并行计算 算法
【DSW Gallery】基于残差网络的度量学习示例
EasyCV是基于Pytorch,以自监督学习和Transformer技术为核心的 all-in-one 视觉算法建模工具,并包含图像分类,度量学习,目标检测,姿态识别等视觉任务的SOTA算法。本文以度量学习为例,为您介绍如何在PAI-DSW中使用EasyCV。
【DSW Gallery】基于残差网络的度量学习示例
|
存储 机器学习/深度学习 分布式计算
【DSW Gallery】特征平台
特征平台是专门用来存储,共享,管理机器学习模型特征的存储库。特征平台可以方便的向多人、多团队共享特征,提供安全,高效且统一的存储,保证离线在线的一致性。
【DSW Gallery】特征平台