General Question about Trax

Course Q&A Natural Language Processing NLP Resources
1

Hi everyone,

I was wondering if at the time of writing this it is actually recommended to get deeper into trax?

I saw in another post that trax hasn’t been updated for about 2 years and that still seems to be the case. When I checked for online resources (documentation, tutorials…) there was not a lot to find either. Installation on Windows and M1 Macs also seems to be rather difficult. So, after finishing this specialization, it feels like concentrating on Tensorflow and/or PyTorch would be more useful, or am I wrong?

I honestly like the way how to handle things with Trax. But if models, resources, tutorials and so on on the internet are little common, it does not feel like the right path to continue.

Thanks for your input in advance! :slight_smile:

2

You certainly have a point. My guess is because it takes a lot of resources to develop a new (be it higher level than existing) frameworks, google has probably concentrated on Tensorflow.

When I was doing the specialization trax seemed to have many advantages like easier to write, fewer commands, faster on some computations etc. But that all came with certain investment, and probably Tensorflow is catching up with speed and so forth nowdays.

So they probably decided to move forward with TF only, Keras now is also part of TF.

1 Like
3

Notice also that Lukasz Kaiser, the inventor and original author, left Google in June of 2021. Coincidence?

ps: he moved over to OpenAI

1 Like
4

Well that seals it. :grin:

1 Like
5

yep, seems like it! :smiley:

1 Like