4 ms·
One possible explanation is to make PyTorch victory over TensorFlow/JAX even more apparent: with Google ceasing to release interesting open-source models, and F
by krasin 3y ago
One possible explanation is to make PyTorch victory over TensorFlow/JAX even more apparent: with Google ceasing to release interesting open-source models, and Facebook / Meta continuing doing so (even if under a less permissive license than usual), it means that there's nothing interesting in the TensorFlow/JAX ecosystem, and a lot in the PyTorch one.
(context: I work for Google DeepMind, opinions are my own)
- iudqnolq 3y agoAnother possible reason might be to attract and retain talent? Speculating wildly but I'd guess some top researchers would value having their work be the industry standard rather than internal.
- desperate 3y agoYour speculation is pretty accurate afaik. Mark commented that releasing open source models is what is needed to retain the best so it is what they will do (paraphrasing) on his interview with Lex Friedman.
- whimsicalism 3y agoHmm, but I feel like Jax/Pytorch interop is only getting better. Tensorflow abstraction has already firmly lost.
- krasin 3y ago> Hmm, but I feel like Jax/Pytorch interop is only getting better There's definitely some convergence. It's good. It still stands for a fact that most people do/think about PyTorch nowadays, not JAX (speaking as someone who likes JAX).