FlyAI训练加速模型

TensorFlow公开模型

原地址: https://github.com/tensorflow/models/tree/master/research/slim

inception_v1_2016_08_28.tar.gz
加速地址: https://www.flyai.com/m/inception_v1_2016_08_28.tar.gz
inception_v2_2016_08_28.tar.gz
加速地址: https://www.flyai.com/m/inception_v2_2016_08_28.tar.gz
inception_v3_2016_08_28.tar.gz
加速地址: https://www.flyai.com/m/inception_v3_2016_08_28.tar.gz
inception_v4_2016_09_09.tar.gz
加速地址: https://www.flyai.com/m/inception_v4_2016_09_09.tar.gz
inception_resnet_v2_2016_08_30.tar.gz
加速地址: https://www.flyai.com/m/inception_resnet_v2_2016_08_30.tar.gz
resnet_v1_50_2016_08_28.tar.gz
加速地址: https://www.flyai.com/m/resnet_v1_50_2016_08_28.tar.gz
resnet_v1_101_2016_08_28.tar.gz
加速地址: https://www.flyai.com/m/resnet_v1_101_2016_08_28.tar.gz
resnet_v1_152_2016_08_28.tar.gz
加速地址: https://www.flyai.com/m/resnet_v1_152_2016_08_28.tar.gz
resnet_v2_50_2017_04_14.tar.gz
加速地址: https://www.flyai.com/m/resnet_v2_50_2017_04_14.tar.gz
resnet_v2_101_2017_04_14.tar.gz
加速地址: https://www.flyai.com/m/resnet_v2_101_2017_04_14.tar.gz
resnet_v2_152_2017_04_14.tar.gz
加速地址: https://www.flyai.com/m/resnet_v2_152_2017_04_14.tar.gz
vgg_16_2016_08_28.tar.gz
加速地址: https://www.flyai.com/m/vgg_16_2016_08_28.tar.gz
vgg_19_2016_08_28.tar.gz
加速地址: https://www.flyai.com/m/vgg_19_2016_08_28.tar.gz
mobilenet_v1_1.0_224.tgz
加速地址: https://www.flyai.com/m/mobilenet_v1_1.0_224.tgz
mobilenet_v1_0.50_160.tgz
加速地址: https://www.flyai.com/m/mobilenet_v1_0.50_160.tgz
mobilenet_v1_0.25_128.tgz
加速地址: https://www.flyai.com/m/mobilenet_v1_0.25_128.tgz
mobilenet_v2_1.4_224.tgz
加速地址: https://www.flyai.com/m/mobilenet_v2_1.4_224.tgz
mobilenet_v2_1.0_224.tgz
加速地址: https://www.flyai.com/m/mobilenet_v2_1.0_224.tgz
nasnet-a_mobile_04_10_2017.tar.gz
加速地址: https://www.flyai.com/m/nasnet-a_mobile_04_10_2017.tar.gz
nasnet-a_large_04_10_2017.tar.gz
加速地址: https://www.flyai.com/m/nasnet-a_large_04_10_2017.tar.gz
pnasnet-5_large_2017_12_13.tar.gz
加速地址: https://www.flyai.com/m/pnasnet-5_large_2017_12_13.tar.gz
pnasnet-5_mobile_2017_12_13.tar.gz
加速地址: https://www.flyai.com/m/pnasnet-5_mobile_2017_12_13.tar.gz

BERT预训练模型

原地址: https://github.com/google-research/bert

BERT-Base, Uncased: 12-layer, 768-hidden, 12-heads, 110M parameters
加速地址: https://www.flyai.com/m/uncased_L-12_H-768_A-12.zip
BERT-Large, Uncased: 24-layer, 1024-hidden, 16-heads, 340M parameters
加速地址: https://www.flyai.com/m/uncased_L-24_H-1024_A-16.zip
BERT-Base, Cased: 12-layer, 768-hidden, 12-heads , 110M parameters
加速地址: https://www.flyai.com/m/cased_L-12_H-768_A-12.zip
BERT-Large, Cased: 24-layer, 1024-hidden, 16-heads, 340M parameters
加速地址: https://www.flyai.com/m/cased_L-24_H-1024_A-16.zip
BERT-Base, Multilingual Cased (New, recommended): 104 languages, 12-layer, 768-hidden, 12-heads, 110M parameters
加速地址: https://www.flyai.com/m/multi_cased_L-12_H-768_A-12.zip
BERT-Base, Chinese: Chinese Simplified and Traditional, 12-layer, 768-hidden, 12-heads, 110M parameters
加速地址: https://www.flyai.com/m/chinese_L-12_H-768_A-12.zip

BERT PyTorch版

BERT-Base, Uncased: 12-layer, 768-hidden, 12-heads, 388M parameters
加速地址: https://www.flyai.com/m/bert-base-uncased.tar.gz
BERT-Large, Uncased: 24-layer, 1024-hidden, 16-heads, 1.16GB parameters
加速地址: https://www.flyai.com/m/bert-large-uncased.tar.gz
BERT-Base, Cased: 12-layer, 768-hidden, 12-heads , 385M parameters
加速地址: https://www.flyai.com/m/bert-base-cased.tar.gz
BERT-Large, Cased: 24-layer, 1024-hidden, 16-heads, 1.16GB parameters
加速地址: https://www.flyai.com/m/bert-large-cased.tar.gz
BERT-Base, Multilingual Cased (New, recommended): 104 languages, 12-layer, 768-hidden, 12-heads, 632M parameters
加速地址: https://www.flyai.com/m/bert-base-multilingual-cased.tar.gz
BERT-Base, Chinese: Chinese Simplified and Traditional, 12-layer, 768-hidden, 12-heads, 364M parameters
加速地址: https://www.flyai.com/m/bert-base-chinese.tar.gz

Keras公开模型

原地址: https://github.com/fchollet/deep-learning-models/releases/

V0.8

densenet121_weights_tf_dim_ordering_tf_kernels.h5(32.5 MB)
加速地址: https://www.flyai.com/m/v0.8|densenet121_weights_tf_dim_ordering_tf_kernels.h5
densenet121_weights_tf_dim_ordering_tf_kernels_notop.h5(28.6 MB)
加速地址: https://www.flyai.com/m/v0.8|densenet121_weights_tf_dim_ordering_tf_kernels_notop.h5
densenet169_weights_tf_dim_ordering_tf_kernels.h5(57.1 MB)
加速地址: https://www.flyai.com/m/v0.8|densenet169_weights_tf_dim_ordering_tf_kernels.h5
densenet169_weights_tf_dim_ordering_tf_kernels_notop.h5(50.7 MB)
加速地址: https://www.flyai.com/m/v0.8|densenet169_weights_tf_dim_ordering_tf_kernels_notop.h5
densenet201_weights_tf_dim_ordering_tf_kernels.h5(80.2 MB)
加速地址: https://www.flyai.com/m/v0.8|densenet201_weights_tf_dim_ordering_tf_kernels.h5
densenet201_weights_tf_dim_ordering_tf_kernels_notop.h5(72.9 MB)
加速地址: https://www.flyai.com/m/v0.8|densenet201_weights_tf_dim_ordering_tf_kernels_notop.h5
NASNet-large-no-top.h5(328 MB)
加速地址: https://www.flyai.com/m/v0.8|NASNet-large-no-top.h5
NASNet-large.h5(343 MB)
加速地址: https://www.flyai.com/m/v0.8|NASNet-large.h5
NASNet-mobile-no-top.h5(19.1 MB)
加速地址: https://www.flyai.com/m/v0.8|NASNet-mobile-no-top.h5
NASNet-mobile.h5(23.1 MB)
加速地址: https://www.flyai.com/m/v0.8|NASNet-mobile.h5

V0.7

inception_resnet_v2_weights_tf_dim_ordering_tf_kernels.h5(215 MB)
加速地址: https://www.flyai.com/m/v0.7|inception_resnet_v2_weights_tf_dim_ordering_tf_kernels.h5
inception_resnet_v2_weights_tf_dim_ordering_tf_kernels_notop.h5(209 MB)
加速地址: https://www.flyai.com/m/v0.7|inception_resnet_v2_weights_tf_dim_ordering_tf_kernels_notop.h5

V0.5

inception_v3_weights_tf_dim_ordering_tf_kernels.h5(91.7 MB)
加速地址: https://www.flyai.com/m/v0.5|inception_v3_weights_tf_dim_ordering_tf_kernels.h5
inception_v3_weights_tf_dim_ordering_tf_kernels_notop.h5(83.8 MB)
加速地址: https://www.flyai.com/m/v0.5|inception_v3_weights_tf_dim_ordering_tf_kernels_notop.h5

V0.4

xception_weights_tf_dim_ordering_tf_kernels.h5(87.6 MB)
加速地址: https://www.flyai.com/m/v0.4|xception_weights_tf_dim_ordering_tf_kernels.h5
xception_weights_tf_dim_ordering_tf_kernels_notop.h5(79.8 MB)
加速地址: https://www.flyai.com/m/v0.4|xception_weights_tf_dim_ordering_tf_kernels_notop.h5

V0.3

music_tagger_crnn_weights_tf_kernels_tf_dim_ordering.h5(1.57 MB)
加速地址: https://www.flyai.com/m/v0.3|music_tagger_crnn_weights_tf_kernels_tf_dim_ordering.h5
music_tagger_crnn_weights_tf_kernels_th_dim_ordering.h5(1.57 MB)
加速地址: https://www.flyai.com/m/v0.3|music_tagger_crnn_weights_tf_kernels_th_dim_ordering.h5

V0.2

inception_v3_weights_tf_dim_ordering_tf_kernels.h5(90.7 MB)
加速地址: https://www.flyai.com/m/v0.2|inception_v3_weights_tf_dim_ordering_tf_kernels.h5
inception_v3_weights_tf_dim_ordering_tf_kernels_notop.h5(82.9 MB)
加速地址: https://www.flyai.com/m/v0.2|inception_v3_weights_tf_dim_ordering_tf_kernels_notop.h5
inception_v3_weights_th_dim_ordering_th_kernels.h5(90.7 MB)
加速地址: https://www.flyai.com/m/v0.2|inception_v3_weights_th_dim_ordering_th_kernels.h5
inception_v3_weights_th_dim_ordering_th_kernels_notop.h5(82.9 MB)
加速地址: https://www.flyai.com/m/v0.2|inception_v3_weights_th_dim_ordering_th_kernels_notop.h5
resnet50_weights_tf_dim_ordering_tf_kernels.h5(98.1 MB)
加速地址: https://www.flyai.com/m/v0.2|resnet50_weights_tf_dim_ordering_tf_kernels.h5
resnet50_weights_tf_dim_ordering_tf_kernels_notop.h5(90.3 MB)
加速地址: https://www.flyai.com/m/v0.2|resnet50_weights_tf_dim_ordering_tf_kernels_notop.h5
resnet50_weights_th_dim_ordering_th_kernels.h5(98.1 MB)
加速地址: https://www.flyai.com/m/v0.2|resnet50_weights_th_dim_ordering_th_kernels.h5
resnet50_weights_th_dim_ordering_th_kernels_notop.h5(90.3 MB)
加速地址: https://www.flyai.com/m/v0.2|resnet50_weights_th_dim_ordering_th_kernels_notop.h5

V0.1

resnet50_weights_tf_dim_ordering_tf_kernels.h5(98.1 MB)
加速地址: https://www.flyai.com/m/v0.1|resnet50_weights_tf_dim_ordering_tf_kernels.h5
resnet50_weights_tf_dim_ordering_tf_kernels_notop.h5(90.3 MB)
加速地址: https://www.flyai.com/m/v0.1|resnet50_weights_tf_dim_ordering_tf_kernels_notop.h5
resnet50_weights_th_dim_ordering_th_kernels.h5(98.1 MB)
加速地址: https://www.flyai.com/m/v0.1|resnet50_weights_th_dim_ordering_th_kernels.h5
resnet50_weights_th_dim_ordering_th_kernels_notop.h5(90.3 MB)
加速地址: https://www.flyai.com/m/v0.1|resnet50_weights_th_dim_ordering_th_kernels_notop.h5
vgg16_weights_tf_dim_ordering_tf_kernels.h5(528 MB)
加速地址: https://www.flyai.com/m/v0.1|vgg16_weights_tf_dim_ordering_tf_kernels.h5
vgg16_weights_tf_dim_ordering_tf_kernels_notop.h5(56.2 MB)
加速地址: https://www.flyai.com/m/v0.1|vgg16_weights_tf_dim_ordering_tf_kernels_notop.h5
vgg16_weights_th_dim_ordering_th_kernels.h5(528 MB)
加速地址: https://www.flyai.com/m/v0.1|vgg16_weights_th_dim_ordering_th_kernels.h5
vgg16_weights_th_dim_ordering_th_kernels_notop.h5(56.2 MB)
加速地址: https://www.flyai.com/m/v0.1|vgg16_weights_th_dim_ordering_th_kernels_notop.h5
vgg19_weights_tf_dim_ordering_tf_kernels.h5(548 MB)
加速地址: https://www.flyai.com/m/v0.1|vgg19_weights_tf_dim_ordering_tf_kernels.h5
vgg19_weights_tf_dim_ordering_tf_kernels_notop.h5(76.4 MB)
加速地址: https://www.flyai.com/m/v0.1|vgg19_weights_tf_dim_ordering_tf_kernels_notop.h5
vgg19_weights_th_dim_ordering_th_kernels.h5(548 MB)
加速地址: https://www.flyai.com/m/v0.1|vgg19_weights_th_dim_ordering_th_kernels.h5
vgg19_weights_th_dim_ordering_th_kernels_notop.h5(76.4 MB)
加速地址: https://www.flyai.com/m/v0.1|vgg19_weights_th_dim_ordering_th_kernels_notop.h5

PyTorch公开模型

原地址: https://pytorch.org/docs/stable/torchvision/models.html

alexnet-owt-4df8aa71.pth (234MB)
加速地址: https://www.flyai.com/m/alexnet-owt-4df8aa71.pth
densenet121-a639ec97.pth (31MB)
加速地址: https://www.flyai.com/m/densenet121-a639ec97.pth
densenet161-8d451a50.pth (111MB)
加速地址: https://www.flyai.com/m/densenet161-8d451a50.pth
densenet169-b2777c0a.pth (55MB)
加速地址: https://www.flyai.com/m/densenet169-b2777c0a.pth
densenet201-c1103571.pth (78MB)
加速地址: https://www.flyai.com/m/densenet201-c1103571.pth
inception_v3_google-1a9a5a14.pth (104MB)
加速地址: https://www.flyai.com/m/inception_v3_google-1a9a5a14.pth
resnet18-5c106cde.pth (45MB)
加速地址: https://www.flyai.com/m/resnet18-5c106cde.pth
resnet34-333f7ec4.pth (84MB)
加速地址: https://www.flyai.com/m/resnet34-333f7ec4.pth
resnet50-19c8e357.pth (97.75MB)
加速地址: https://www.flyai.com/m/resnet50-19c8e357.pth
resnet152-b121ed2d.pth (231MB)
加速地址: https://www.flyai.com/m/resnet152-b121ed2d.pth
resnet101-5d3b4d8f.pth (171MB)
加速地址: https://www.flyai.com/m/resnet101-5d3b4d8f.pth
squeezenet1_0-a815701f.pth (4.8MB)
加速地址: https://www.flyai.com/m/squeezenet1_0-a815701f.pth
squeezenet1_1-f364aa15.pth (4.8MB)
加速地址: https://www.flyai.com/m/squeezenet1_0-a815701f.pth
vgg11-bbd30ac9.pth (507MB)
加速地址: https://www.flyai.com/m/vgg11-bbd30ac9.pth
vgg11_bn-6002323d.pth (507MB)
加速地址: https://www.flyai.com/m/vgg11_bn-6002323d.pth
vgg13_bn-abd245e5.pth (508MB)
加速地址: https://www.flyai.com/m/vgg13_bn-abd245e5.pth
vgg13-c768596a.pth (508MB)
加速地址: https://www.flyai.com/m/vgg13-c768596a.pth
vgg16-397923af.pth (528MB)
加速地址: https://www.flyai.com/m/vgg16-397923af.pth
vgg16_bn-6c64b313.pth (528MB)
加速地址: https://www.flyai.com/m/vgg16_bn-6c64b313.pth
vgg19_bn-c79401a0.pth (549MB)
加速地址: https://www.flyai.com/m/vgg19_bn-c79401a0.pth
vgg19-dcbb9e9d.pth (549MB)
加速地址: https://www.flyai.com/m/vgg19-dcbb9e9d.pth

公开词向量

GloVe

原地址: https://github.com/stanfordnlp/GloVe

Common Crawl (42B tokens, 1.9M vocab, uncased, 300d vectors, 1.75 GB download)
加速地址: https://www.flyai.com/m/glove.42B.300d.zip
Common Crawl (840B tokens, 2.2M vocab, cased, 300d vectors, 2.03 GB download)
加速地址: https://www.flyai.com/m/glove.840B.300d.zip
Wikipedia 2014 + Gigaword 5 (6B tokens, 400K vocab, uncased, 300d vectors, 822 MB download)
加速地址: https://www.flyai.com/m/glove.6B.zip
Twitter (2B tweets, 27B tokens, 1.2M vocab, uncased, 200d vectors, 1.42 GB download)
加速地址: https://www.flyai.com/m/glove.twitter.27B.zip

搜狗

原地址: http://www.sogou.com/labs/

加速地址: https://www.flyai.com/m/sgns.sogou.word.zip (312MB)