What is the main difference between Trax and Tensorflow?
Both of them are deep learning libraries and implemented by Google.
$\begingroup$
$\endgroup$
2
asked Jan 12, 2021 at 10:09
-
$\begingroup$ They aren't necessarily competing; it looks more like they are complementary. You can work with both of them together. I don't know why they maintain two different libraries though, my guess is that they originated from opensourcing different projects by different google research teams. $\endgroup$– Djib2011Commented Jan 12, 2021 at 10:36
-
$\begingroup$ Related thread at Stack Overflow: What is the difference between JAX, Trax, and TensorRT, in simple terms? $\endgroup$– desertnautCommented Oct 27, 2021 at 18:51
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
Trax is a high-level interface to make it easy to apply existing models (like ResNet, LSTM, Transformer). Trax focuses almost exclusively on deep learning.
TensorFlow is a data flow computational framework that can be applied to a wide variety of problems. TensorFlow allows fine grain control over every step in the modeling process. TensorFlow is often used for deep learning, but given its flexibility it can be used for any numerical computation.
answered Mar 14, 2021 at 15:16