Autoaugment [data preprocessing]

[remc (Rémi Caland)]

Hello everyone,
I am currently working on object identification using tensorflow 2.0 rc, and I just implemented autoaugment for data augmentation.
Did someone use it already?
Did it give you better result than before?

For those who did not know it, you can take a look at fast auto-augment here (made with pytorch) & auto-augment here !

[fila.pro (Danila)]

Hi @remc,

Could you please share the code of AutoAugment working with tf2.0? Unfortunately, original implementation is compatible only with tf1.*.

Thank you.

[remc (Rémi Caland)]

Hello ! I didnt upgrade it for now, I was hopping for some reviews before working on it.

[Nilanjan (Nilanjan Chattopadhyay)]

Hi! I want to implement the fast autoaugment in Fastai? Does anyone know how to do that?

[juvian (julian)]

Applying Google’s AutoAugment data augmentation policies to a Fastai object detection model

In 2019, the Google Brain Team published a paper entitled, “AutoAugment: Learning Augmentation Policies from Data”, by Ekin D. Cubuk, Barret Zoph, Dandelion Mane, Vijay Vasudevan, and Quoc V. Le. AutoAugment is an automatic method of designing custom data augmentation policies for computer vision data sets and is especially effective in training object detection models. The derived policies apply to ImageNet, COCO, MNIST, and other data sets We have taken the ImageNet policy and developed an …