Chevron Left
Volver a Custom Models, Layers, and Loss Functions with TensorFlow

Opiniones y comentarios de aprendices correspondientes a Custom Models, Layers, and Loss Functions with TensorFlow por parte de deeplearning.ai

4.9
estrellas
732 calificaciones

Acerca del Curso

In this course, you will: • Compare Functional and Sequential APIs, discover new models you can build with the Functional API, and build a model that produces multiple outputs including a Siamese network. • Build custom loss functions (including the contrastive loss function used in a Siamese network) in order to measure how well a model is doing and help your neural network learn from training data. • Build off of existing standard layers to create custom layers for your models, customize a network layer with a lambda layer, understand the differences between them, learn what makes up a custom layer, and explore activation functions. • Build off of existing models to add custom functionality, learn how to define your own custom class instead of using the Functional or Sequential APIs, build models that can be inherited from the TensorFlow Model class, and build a residual network (ResNet) through defining a custom model class. The DeepLearning.AI TensorFlow: Advanced Techniques Specialization introduces the features of TensorFlow that provide learners with more control over their model architecture and tools that help them create and train advanced ML models. This Specialization is for early and mid-career software and machine learning engineers with a foundational understanding of TensorFlow who are looking to expand their knowledge and skill set by learning advanced TensorFlow features to build powerful models....

Principales reseñas

PK

3 de feb. de 2021

It is advanced TF specialization and the way contents are presented in the course are very systematically. Definitely recommended for developers already familiar with TF and wanted to explore further.

MS

24 de nov. de 2020

Really great course, it teaches you all about the TF API and how to customize it for your needs, i thought only pytorch can make that as it's really pythonic, but i am a nieve noob what can i say.

Filtrar por:

26 - 50 de 162 revisiones para Custom Models, Layers, and Loss Functions with TensorFlow

por Enzo D G M

23 de nov. de 2020

Excellent course!!! very well explained

por onkar s p

24 de nov. de 2020

Lawrence you are just awesome !!

por Tim C

22 de dic. de 2020

Powerful stuff

por Giora S

7 de ene. de 2021

I expected more, but maybe this will come in the following courses. Anyone can subclass Layer/Model with a short stack overflow post. But how do I loop a special fitting procedure? How do I make TF take into consideration external variables? What about some more complex loss functions with TF math operations? What about sparse input?

por Dmitry D

8 de ene. de 2021

Great course! Trainer clearly explains necessary features of keras, that are widely used. Very good code examples.

Just one issue: a dataset used in week 5 is unavailable in early january 2021 -- at least couple of days

por Fadillah M

6 de ene. de 2021

Great course for you who want to know how flexible Keras is. From this course, I realize that both Tensorflow & Keras are flexible and simple to use with.

por Ryan A

7 de ene. de 2021

I started this course with the intention of learning the syntax needed to implement VAEs. This course satisfied that requirement perfectly! Thank you :)

por Imad y

8 de ene. de 2021

It was a very interesting course and the teacher delivered it very well. Thanks to Coursera and its members.

por Teemu K

1 de dic. de 2020

Interesting content in easy to digest sized parts. Maybe a bit too much repetation (lectures and coding screencasts). Some typos like units="1", which look quite amateurish. Homeworks were too easy. Too much fill this part kinda homeworks.

por Sayak P

18 de nov. de 2020

The quiz questions seemed a bit fuzzy sometimes. For example, there was a question on why do we loop through the residual blocks and the answer was to reduce the network depth. What's the context of reduction here?

por Franklin S

3 de ene. de 2021

Very useful and complete course for those who want to go deeper into tensorflow tools for make customize models in their own problem.

por brahmendra

25 de dic. de 2020

Very useful for developing our own networks

por Igor B

7 de ene. de 2021

Ok, but Andrew NG Courses are much better

por Asif A

21 de ene. de 2021

Perfectly structured course with clear explanations and very easy to follow examples. Learned inception network structure, vgg, block based architectures, removed a lot confusion on I had on resnets that I learned earlier. I had some prior knowledge of functional API multi IO, but learning about model subclassing, custom losses, layers, lambda layers will make it easy to implement various custom model by myself.

por Ernest W

12 de nov. de 2021

The first course of Advanced Techniques specialization is great! I was afraid it could be too difficult after finishing only the deep learning specialization. It is an overview of creating custom functions and models, someone might think the assignments are too trivial but I think it's great as the author explains the concepts in understandable way. More complex tasks would be overwhelming.

por Marco Z

21 de feb. de 2021

The course is built really step by step with many clear examples and repetitions not only about tensor flow but also to improve the way one writes code, efficiently and clean. Quiz and Programming Tasks are not difficult as all is very well formulated and the clear. Check capabilities reduce debugging stress ...

por Akshay K

14 de ene. de 2021

I enjoyed this course thoroughly. The video explanations were simple and easy to understand. I really liked the part where a video is dedicated to go through the code. Along with having hands-on access to the code which is taught in the videos, the assignments were also of great help for getting our hands dirty.

por Homayoun

23 de abr. de 2021

My favorite part of this course and other courses in this and other TensorFlow specializations offer by Laurence and Deep learnign.AI is the recaps at the beginning of every video; He connects all the videos and concepts together and makes the learner understand where they are and where they're going and why.

por Andrea G

3 de sep. de 2021

The lessons and examples are clear and entertaining. The overall content is interesting. The only flaws are the extremely detailed instructions and large code chunks provided in the practical tests. These help to concentrate on the topic, but tests result a little straightforward.

por Abd-elrahman S

1 de feb. de 2022

overall great course if you have solid foundation in tensorflow ,Don't be fooled by the easy assignments in week 1 ,because from there the course gets into object oriented programming with customization of a lot of things and you will encounter building the complex architectures

por zaky r

31 de ene. de 2021

Excellent introduction to customizing your neural network. its an eye opener for people who wants to have more customizable layers, loss function, callbacks and even the architecture itself. Also the instructor themself are amazing. I really recommend the course! definitely 5/5

por Florian C

4 de ago. de 2021

E​xcellent course! Thank you to the instructors and everyone which contributed to it!

The theory is very well explained and I also had the opportunity to practice meaningful hands-on coding. It gives a boost to anyone wanting to work with more complex machine learning models.

por Mohamed M

14 de ene. de 2021

Great course with a well structured plan and flow between topics. I really enjoyed the way the programming assignments were delivered. Completing this course has made me eager to apply my new knowledge of the TensorFlow 2 Functional API to my data science pipelines!

por Ozan G

15 de feb. de 2021

If you are going to be doing research with Tensorflow, you will most likely need this course. Functional API and custom layers, loss functions and models are essential for going above and beyond training MNIST classifiers and building something novel.

por Pranjal J

23 de dic. de 2021

The Course provides easy understanding to explore tensorflow layers and APIs to create complex models. The Custom Loss and Custom Callback features along with model subclassing will help students to build powerful models very efficiently.