Is TensorFlow Dying?

[VladKuzR]

Hey mates, I have a question to think about.

TensorBoard.dev has been shut down as of January 1, 2024. TensorFlow Hub was moved to Kaggle Models and abandoned (it seems), and the Tensorflow Developer Certification program will also be shut down soon.

• So, what do you think?
• Is TensorFlow dying?
• Are we better off transferring to PyTorch?
• Will it be developing in the future?
• Will it be supported in the future?

[huangssssx]

Well...this may be unfortunate news, but I think both pytorch and jax are just tools to solve problems. Learning tensorflow is just to use it to enter the palace of deep learning. When you get in, you can use this crowbar Choose to carry it but don't use it as a leg

[VladKuzR]

What do you mean? What can you use for deep learning if not tools for deep learning?

[rhuanbarros]

I think keras could dominate the field as with keras 3 we can "compile" the code to wherever framework we want like tensorflow, pytorch and jax, if I get it right. Also, I think as I gain experience in the field I should know how to use keras, tensorflow, pytorch and jax in a few years. So I just starting with tensorflow and keras.