Tensorflow(Deep Learning)

In the Tensorflow vedio you have used manual method of Linear regression and Gradient Descent and then used the tensorflow so Is it possible to do that manual Linear Regression and Gradient Descent with the help of Scikit Learn Package and then use tensorflow on it.
It will be really fast and easy to implement.
IS IT POSSIBLE OR WE HAVE TO APPLY MANUAL FORMULAE FOR THE SAME EVERYTIME?
PLEASE REPLY.

Hi Anubhav,

We don’t need to use manual methods. The whole point is to get the workaround for using scikit learn. After that, we have Deep Learning libraries like TensorFlow and Keras API on top of it.

The manual methods are for the fundamental and intuitive understanding of how the algorithms like Gradient Descent and Back-Propagation inner workings so that the mathematical constructs behind the algorithms are clear to us.

I hope it helps.

Regards

1 Like

So can we use scikit learn Linear Regression or other package and then tensorflow on it??

Hi Anubhav,

Scikit Learn Package has well documented functions for mostly everything for Machine Learning.
Same goes for TensorFlow and Keras, for Deep Learning. So, Yes Indeed. We use Scikit learn Linear Regression or other packages and then tensorflow on it.

Please switch to TensorFlow 2.x as soon as possible as it has shorter codes and its faster. For Example, there are multiple changes in TensorFlow 2.0 to make TensorFlow users more productive. TensorFlow 2.0 removes redundant APIs, makes APIs more consistent (Unified RNNs, Unified Optimizers), and better integrates with the Python runtime with Eager execution.

Happy Coding!

Regards

1 Like

so that meansThe vedios of Tensorflow are not updated, it should be updated.

Hi Anubhav,

TensorFlow 1.x is still being used in most of the industries and still very relevant. So the videos tutorials are also relevant. Switching to Tf 2.x won’t be difficult at all after knowing the Tf 1.x.

The technology is changing so fast that after every few months, a new version of almost every library is released addressing the shortcomings of the previous version. This is how the industry works and we have to keep up with the pace ourselves.

I hope it helps

Regards

Yes, I agree but you should also keep us updated is’nt it?

Hi Anubhav,

I am also a co-learner just like you and the others. I have been around in ML and DL for a long time now and I still can’t call myself an expert. No one can.

The codes changes so fast that we have to keep ourselves updated at a very fast pace.

Its sort of an experimental computer science and the progress and the updates are way too rapid. I totally agree that everything should be updated but I think its not possible at all for any of the courses including this and any other to keep updating everything in such a short notice. The hard way is the only way I guess.

I hope it helps

Happy Coding and Regards

Absolutely I agree with you, and Can you please do me a favour if you are in cloudxlab team then please tell them to add vedios on recommendation system also because it is one of the main and practical use case for ML.

Thankyou.

Hi Anubhav,

Unfortunately, No. I don’t work for Cloudxlab. Please search through youtube. There are a lot of exceptional Recommender Systems tutorials over there. I think you will definitely find the one which will get your job done.

Happy Coding and Regards.