Introduction to TensorFlow 2.0: Easier for beginners, and more powerful for experts (TF World '19)
[ad_1]
TensorFlow 2.0 is all about ease of use, and there has never been a better time to get started. In this talk, we will introduce model-building styles for beginners and experts, including the Sequential, Functional, and Subclassing APIs. We will share complete, end-to-end code examples in each style, covering topics from “Hello World” all the way up to advanced examples. At the end, we will point you to educational resources you can use to learn more.
Presented by: Josh Gordon
View the website → https://goo.gle/36smBfW
#TFWorld All Sessions → https://goo.gle/TFWorld19
Subscribe to the TensorFlow channel → https://goo.gle/TensorFlow
Source
[ad_2]
The tutorials for beginners in the website do not explain most things (what are all the lines doing). I code in python for 15 years. I have never have a trouble understanding any library anything until now.
That's a super cool presentation! I believe even someone who is not familiar with TF and ML will have the motivation to try it out after this talk.
Anyone know where can I find the share collab projects?
But early stopping shouldn't be used because it will lead orthogonalization
Is keras 100 percent comparible with tensorflow 2.0?
Nothing simple conceptually? What does that mean? 9:35
Google programmers are so incredibly well spoken and good at presenting, either there must be some training internally or they select for this at interviews
Tensorflow 2.0 will make modelling so much expressive/artistic with the added touch of lower level control if needed , plus easier deployment? Wow! just wow! ! What an exciting time to exist.😂😂😂
Excellent presentation. Thanks!
Absolutely fantastic talk, thank you
Wonderful presentation! Thanks so much.
I was a lot skeptical about tensorflow 2.0 when I heard that it has been made like Keras. Thb, I absolutely hate keras as something always feels missing in the framework (maybe a lack of control). But the introduction of subclassing really exites me now. Its good that TF is a lot like Torch now, the flexibility looks amazing. Good work tensorflow!
Cannot install tensorflow as wrapt is not getting installed
HELP!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
nice!
Where can we find the slides please ?
Very well explained ! I am impressed with this guy's articulation skills.
IMPRESSED
need help in prediction from trained & saved model, same as 36:36 ,,, any one expert here ?
Like this new way of writing custom stuff in the low level with tf.Module instead of keras
Excellent!
slides for this presentation: it has lots of useful links… https://drive.google.com/file/d/1YG3QX8syS-t8WyWOjGaGrtrsvSDqSTLx/view
When is CSV data going to be done? It's still in the experimental section, and TF has been out 4 years. Secondly, there is no instructional coverage of CSV at all in the brand new Tensorflow Specialization on coursera by Andrew Ng and Google's trainer. The whole UCI ML dataset lib is CSV, in case you didn't notice yet somehow. Unlike software app developers, data scientists don't get to create whatever dataset we want just like app developers seem to do. We have to do work on the data that the organization has provided and in case data science is a new concept for you, that is largely CSV for example Higgs Boson dataset. But both Microsoft and Google still seem to treat data scientists as if software apps with unstructured vision/NLP are all that really matters, and the structured datasets (ie, CSV) are for science nerds and don't really matter much so lets let our summer interns take care of CSV support while we're on vacation. "It should be easy." Did Microsoft actually try to use the CNTKTextFormatReader (later simply CTF Reader)? DUDES go to Kaggle and try the BOSCH dataset (not Titanic enough of that) and get back to me stories about easy CSVs again, bro. Try to use your own tf.data.experimental.CsvDataset and get back to us.
Great presentation. Are there slides available for this?
17:18 Can we use callbacks on train_on_batch? Any references for that?
wow! super cool presentation! Thanks, Josh!!
TensorFlow 2.0 is all about ease of use, and there has never been a better time to get started. In this talk, we will introduce model-building styles for beginners and experts, including the Sequential, Functional, and Subclassing APIs. We will share complete, end-to-end code examples in each style, covering topics from "Hello World" all the way up to advanced examples. At the end, we will point you to educational resources you can use to learn more.
#TFWorld
Introduction to TensorFlow 2.0: Easier for beginners, and more powerful for experts (TF World '19)
Nice job, Josh!
Nice job team
WOW ! i am very excited to try …these new features of TF 2.0
Thanks for an excellent presentation.
Nice presentation, thaks Josh!
👍👍👍👍👍
Great talk, josh!
I am learning