Modelis:
Attēlu iekodētājs (torchvision) - pretrained on ImageNet
ConvNet VGG
ResNet
DenseNet
EfficentNet
ViT
RegNet vai vēl kādi
Laika rindu modelis (viņa galā klasifikators)
LSTM (RNN)
Transformer
Rādītājs: F1, ACC
Video lekcijas
Video: https://youtube.com/live/RAN8By8leYY?feature=share
Jamboard: https://jamboard.google.com/d/1cqybN158FOYY7HnQyozed7CQ5hc3W4cbEXkodndCcxE/edit?usp=sharing
Materiāli: https://cs231n.github.io/convolutional-networks/ http://www.cs.utoronto.ca/~rgrosse/cacm2011-cdbn.pdf https://ikhlestov.github.io/pages/machine-learning/convolutions-types/ https://github.com/vdumoulin/conv_arithmetic/blob/master/README.md
Video: https://youtube.com/live/iuf33QP8T0s?feature=share
Jamboard: https://jamboard.google.com/d/1gHCHbXnb9adphDPC2090g2amOSMlviptuXhNQu9zEbc/edit?usp=sharing
Materials:
ResNet: https://arxiv.org/abs/1512.03385
DenseNet: https://arxiv.org/abs/1608.06993
Video: https://youtube.com/live/paCJ6SNWw7s?feature=share
Jamboard: https://jamboard.google.com/d/15_fiAoHrma00h963OWvEbr1qY1QGzRvGVKQpaAU0Yxw/edit?usp=sharing
Materials:
https://calvinfeng.gitbook.io/machine-learning-notebook/supervised-learning/recurrent-neural-network/recurrent_neural_networks/ https://lanwuwei.github.io/courses/SP19/3521_slides/11-Recurrent_Neural_Networks_2.pdf https://danijar.com/tips-for-training-recurrent-neural-networks/ https://medium.com/datadriveninvestor/attention-in-rnns-321fbcd64f05 https://arxiv.org/abs/1610.09513 https://wiki.pathmind.com/word2vec
Video: https://youtube.com/live/A_BN7XkZKbs?feature=share
Jamboard: https://jamboard.google.com/d/1NNp8mYKyPHURPtF8Z9RISGSmRl_QEEsIb2q2-CxeGQE/edit?usp=sharing
Preparation materials: http://jalammar.github.io/illustrated-transformer/ https://arxiv.org/abs/1706.03762
Video: https://youtube.com/live/BYz6Uc-twSw?feature=share
Jamboard: https://jamboard.google.com/d/1GXHTFpnrXIVzBnAISrZBoM-r5IUx58PN70BYvK6dYbI/edit?usp=sharing
Materials: