# Has Google completely abandoned TensorFlow?

**URL:** <https://discuss.ai.google.dev/t/has-google-completely-abandoned-tensorflow/38874>\
**Category:** TensorFlow\
**Tags:** tfdata\
**Created:** [September 9, 2024, 12:56pm UTC](https://discuss.ai.google.dev/t/has-google-completely-abandoned-tensorflow/38874 "2024-09-09T12:56:37Z")\
**Posts on this page:** 5\
**Page:** 1

<div class="post-metadata">

**Author:** ![yehuang](https://d1dlmcr85iqnpo.cloudfront.net/user_avatar/discuss.ai.google.dev/yehuang/32/17317_2.png) [@yehuang](https://discuss.ai.google.dev/u/yehuang)\
**Post date:** [September 9, 2024, 12:56pm UTC](https://discuss.ai.google.dev/t/has-google-completely-abandoned-tensorflow/38874/1 "2024-09-09T12:56:37Z")

</div>

Starting this year, I have raised some issues related to TensorFlow, but no one has addressed them at all. Recently, I have also submitted some pull requests to the TensorFlow repository, but no one has reviewed them. Does this mean that Google has abandoned TensorFlow? If so, I will shift my development focus to Jax.

---

<div class="post-metadata">

**Author:** ![Jeff\_Corpac](https://d1dlmcr85iqnpo.cloudfront.net/user_avatar/discuss.ai.google.dev/jeff_corpac/32/6955_2.png) [@Jeff\_Corpac](https://discuss.ai.google.dev/u/Jeff_Corpac)\
**Post date:** [September 9, 2024, 10:41pm UTC](https://discuss.ai.google.dev/t/has-google-completely-abandoned-tensorflow/38874/3 "2024-09-09T22:41:07Z")

</div>

It looks like Google’s attention has shifted over to Keras, which can use several libraries such as TF, PyTorch, or JAX as its back-end. TensorFlow still seems to be the default back-end, but Google is focusing more on the developer-facing tools.

---

<div class="post-metadata">

**Author:** ![yehuang](https://d1dlmcr85iqnpo.cloudfront.net/user_avatar/discuss.ai.google.dev/yehuang/32/17317_2.png) [@yehuang](https://discuss.ai.google.dev/u/yehuang)\
**Post date:** [September 9, 2024, 11:50pm UTC](https://discuss.ai.google.dev/t/has-google-completely-abandoned-tensorflow/38874/4 "2024-09-09T23:50:35Z")

</div>

Maybe not, Keras’ commit frequency is too low, even lower than TensorFlow.

---

<div class="post-metadata">

**Author:** ![Jeff\_Corpac](https://d1dlmcr85iqnpo.cloudfront.net/user_avatar/discuss.ai.google.dev/jeff_corpac/32/6955_2.png) [@Jeff\_Corpac](https://discuss.ai.google.dev/u/Jeff_Corpac)\
**Post date:** [September 10, 2024, 12:50am UTC](https://discuss.ai.google.dev/t/has-google-completely-abandoned-tensorflow/38874/5 "2024-09-10T00:50:24Z")

</div>

Yeah. It’s pretty telling that Gemini API and Google AI Studio are listed ahead of both of them on this forum.

---

<div class="post-metadata">

**Author:** ![Remy\_Wehrung](https://d1dlmcr85iqnpo.cloudfront.net/user_avatar/discuss.ai.google.dev/remy_wehrung/32/20714_2.png) [@Remy\_Wehrung](https://discuss.ai.google.dev/u/Remy_Wehrung)\
**Post date:** [September 10, 2024, 5:18am UTC](https://discuss.ai.google.dev/t/has-google-completely-abandoned-tensorflow/38874/6 "2024-09-10T05:18:04Z")

</div>

I don’t think so at all, TensorFlow is evolving, we have gone from a “simple” ML framework to a meta interpreter, which is very good, then, we have to look beyond the current LLMs, we have problems with code optimization, frugality, now mandatory, this is not achieved with an annual release like Apple, it’s long term