Tensorflow 2.0 support

[Ryan Wheeler]

Thanks for the excellent bindings, we have had great success using them. Does anyone know when tensorflow 2 saved models will be supported?

I have inquired in other channels about the stability of the tensorflow 2.0 C API but not gotten a response. The docs https://www.tensorflow.org/install/lang_c still show `Note: There is no libtensorflow support for TensorFlow 2 yet. It is expected in a future release`

[Chip Collier]

Howdy!

I'm currently using this crate at work with a saved model created with tensorflow 2. We use yolo for object detection and then a second model for 3d orientation estimation and everything appears to be working out.

So thankfully for at least some tasks it'll work but really I'm just plumbing and couldn't tell you about any particular drawbacks from doing this.

---

From: Ryan Wheeler <ryanmw...@gmail.com>
Sent: Wednesday, May 6, 2020 3:47:45 PM
To: Rust for TensorFlow <ru...@tensorflow.org>
Subject: Tensorflow 2.0 support

 

Thanks for the excellent bindings, we have had great success using them. Does anyone know when tensorflow 2 saved models will be supported?

I have inquired in other channels about the stability of the tensorflow 2.0 C API but not gotten a response. The docs https://www.tensorflow.org/install/lang_c still show `Note: There is no libtensorflow support for TensorFlow 2 yet. It is expected in a future release`

--
You received this message because you are subscribed to the Google Groups "Rust for TensorFlow" group.
To unsubscribe from this group and stop receiving emails from it, send an email to rust+uns...@tensorflow.org.
To view this discussion on the web visit https://groups.google.com/a/tensorflow.org/d/msgid/rust/b07da82d-bd41-44fc-bd4a-41f5794aa455%40tensorflow.org.

[Adam Crume]

I'm glad the crate is useful!  The Rust bindings rely heavily on the C API, so we can't upgrade until the C API for TensorFlow 2 is released, and I don't know when that will be.  I can't guarantee that saved models created in TF 2 will work with the crate, which is currently using TF 1.15, but as long as you avoid using any ops or saved model features which are new in TF 2, it may work.  The official compatibility guarantees are at https://www.tensorflow.org/guide/versions.

To view this discussion on the web visit https://groups.google.com/a/tensorflow.org/d/msgid/rust/DBAPR01MB67268F81FA7F437057AC8F82A8A40%40DBAPR01MB6726.eurprd01.prod.exchangelabs.com.