Pytorch vs Tensorflow vs Keras | Deep Learning Tutorial 6 (Tensorflow Tutorial, Keras & Python)

Key Takeaways

i want to discuss the difference between pytorch tensorflow and keras in this video pytorch and tensorflow are two most popular deep learning frameworks there is a third framework called cntk by microsoft but it is not as popular as the other two pytorch is by facebook and tensorflow is by google now you will ask me what is keras then keras is not a full-fledged deep learning framework just like pytorch and tensorflow keras is just a nice wrapper around tensorflow cntk and theano so previously when people were using tensorflow or cntk directly or the programming uh in these frameworks was not that easy it was a little difficult so then keras was created as just a nice wrapper around these libraries and it just provided a convenience so it is not a full-fledged framework but it just provides your convenience here i have a code snippet where you can see if you import keras by the way if you want to install keras you can just do paper install keras and the code snippet will look something like this if you're directly using keras and nowadays with tensorflow 2.0 they have made it a part of tensorflow library itself so here is another code snippet where you can use keras directly from tensorflow so in all our tutorials we are not going to install keras separately we will just use tensorflow and then we will use keras which is inbuilt into tensorflow to make use of their convenient apis now if you're using a keras previously like you can specify a backend so in the code snippet of keras you you have you would have seen the back end was tensorflow by default and you can change the backend you can change it to cnk cntk or theano but we're not going to go into all of that we'll just use tensorflow and use keras within the tensorflow or to write our programs so i hope that clarifies the difference between these three i will see you in the next tutorial thank you