Using callbacks in training, getting started in TF 2.0, & more! #AskTensorFlow

preview_player
Показать описание

Learn about using callbacks to cancel training once you’ve reached your desired accuracy, how to get started with TensorFlow 2.0 if you’re new to machine learning, and we’ll show you a fun example of image classification in the browser.

Remember to use #AskTensorFlow to have your questions answered in a future episode!

This video is also subtitled in Chinese, Indonesian, Italian, French, German, Japanese, Korean, Portuguese, and Spanish.

event: TensorFlow Dev Summit 2019; re_ty: Publish; product: TensorFlow - General;
Рекомендации по теме
Комментарии
Автор

1:05 -- if(logs.get('accuracy') > 0.6):
'accuracy' instead of 'acc' in TensorFlow 2.0 nightly build
Otherwise it will be error: TypeError: '>' not supported between instances of 'NoneType' and 'float'

foobar
Автор

early_stop = tf.keras.callbacks.EarlyStopping(monitor="val_accuracy", patience=10)
model.fit(train_ds, validation_data=val_ds, epochs=50, callbacks=[early_stop])

i.e. get as good of results as you can out of your selected data, without wasting compute cycles.

contractorwolf
Автор

Awesome format of Q&A. Loved it! Is there a similar one for those interested in Google's AutoML API?

TheCuriousPrimate
Автор

Thank you for the call example. Does the callback return/logs the ETA's? I need to use it so that if the training surpasses a limited time, stops the training.

chekadsarami
Автор

Yes, his class on Coursera is amazing, and it is free. I suggest you guys take his class.

jonathansum
Автор

Thank's guys, really loving TensorFlow! A little tricky to install to be fair, but once is up and running is great!

nachoridesbikes
Автор

#AskTensorFlow please make a video on performing different types of clustering using tensorflow platform.

ranjeetjha
Автор

#AskTensorFlow How to start with tensorflow. I am new to machine learning.

ashishranjanpatel
Автор

Yeah, with keras everything is easy peasy. How do you do it with e.g. estimators?

ChadieRahimian
Автор

Unlike TF2.0 (cpu version), I find TF2.0-gpu is not easy to install. Am I the only one who experienced it?

anuragbhatia