How to pass the TensorFlow Developer Certificate : Everything you need to know

#tensorflow #deeplearning #machinelearning
A while back I decided to take the TensorFlow Developer Certification. So I did find some time over the christmas break, studied hard and did it.
In this video I tell my experience and the resources i have used.
*TensorFlow Developer Certification homepage - www.tensorflow.org/certificate
*Coursera - DeepLearning.AI TensorFlow Developer Professional Certificate - www.coursera.org/professional...
*Hands-On Machine Learning Book - www.oreilly.com/library/view/...
* Deep Learning Book - www.deeplearningbook.org/

[@mikhailbelov6900]

Thank you very much! very informative

[@Sebpv2006]

Una clase magistral ! Muchas gracias !

[@alexeiwmartinez8686]

Thank you! I usually work with tensorflow and I did not know that there was a certification, very interesting ...

[@fullscan]

Glad it was helpful!!

[@garrettosborne4364]

Thanks this is very helpful.

[@danielmejia8530]

Great!

[@narendraprasath5745]

Thanks a lot for sharing your experience for certification. Is google search or tensorflow document reference allowed during tensorflow developer exam time for syntax or function name to check?

[@UzorchukwuOkonkwo]

Great content and delivery ❤‍🔥 I am considering taking the exam and this has been informative. However, I believe there is a little discrepancy with the usage of the Functional API. Using : "model.compile" after building the functional api code did not result in an error because the code simply ran the earlier defined "Sequential model", which should not be the case. So for others like me, I suggest using this just for the functional api section of this video: inputs = tf.keras.Input(shape=(250,250,3)) x = pretrained_model (inputs) x= tf.keras.layers.Flatten() (x) x= tf.keras.layers.Dense(128) (x) outputs = tf.keras.layers.Dense(3, activation='softmax') (x) model = tf.keras.Model(inputs, outputs) Then run the model.compile and model.fit. Thanks once again for this content and I hope this was was helpful as well.

[@chaoukimachreki6422]

You are awesome and your accent is swell!

[@emrecicekyurt3700]

Thanks for the informative video but I wonder why you didn't check the loss curves. The accuracy of the model seems pretty high and there is a chance that overfitting appears because whenever you applied data augmentation and dropout layers, the accuracy has dramatically decreased by 30 percent. Don't you need to check the overfitting or underfitting for the exam? Is it okay if we give a high accuracy resulting model?

[@ihr]

Way to go with videos in English, Luis!

[@narendraprasath5745]

test_ds_scaled = train_ds.map(preprocess) validation_ds_scaled = train_ds.map(preprocess) The above lines of code is not correct. You need to change it as below: test_ds_scaled = test_ds.map(preprocess) validation_ds_scaled = validation_ds.map(preprocess) Thanks

[@Kedoink18]

You mention attention layers not being covered in the course, i guess we can still use it for the nlp portion of the exam?

[@deepaksaini4158]

the documentation suggests that it is an open book exam. Can someone suggest what all kind of resources I can access during the exam. How about physical books, are these allowed?

[@BB-ko3fh]

The TF developer course does not have multivariate time series forecasting (all examples are for univariate). Are there any materials you can refer me to, including some of your old exercises, if any?

[@ShubhamSingh-cw5pd]

Imagine multivariate as using LSTMs with multiple inputs and outputs.. That's the only difference..

[@BB-ko3fh]

@@ShubhamSingh-cw5pd @full scan but how would that work during data preprocessing as you are using windowing to separate the xs and ys. Or does that not come in the exam?

[@danielmejia8530]

Can you share some others exercises to practice? Thanks!

[@fullscan]

Check this out! Public repos from the deeplearning.ai couses github.com/https-deeplearning-ai/tensorflow-1-public

[@BB-ko3fh]

@@fullscan The TF developer course does not have multivariate time series forecasting (all examples are for univariate). Are there any materials you can refer me to, including some of your old exercises, if any? Or does that not come in the exam?

[@ajbssb]

In the last part, you are not using the pretrained model, you are not using x, you are just using model, the one you defined before

[@user-sd4yd9us3x]

why do I get mutch lower accuracy than you? (60-70%)

[@hoanglong4911]

I was frustrated too. Found out he used train dataset for both testing and validation

[@ratnasambhav7338]

Now with the onset of chatGPT, what steps tensorflow is taking to ensure that examiner is not cheating using chatGPT. What if a group of people is solving those problems in a distributed way, the exam is not proctored. If all of this is happening, wouldn't it make this certificate not so useful.

[@indigo1417]

Ive been wondering this too.

[@Grounded441]

more English content would be nice