Tilak

@sachinprasadhs, I was able to reproduce the issue on tensorflow v2.8, v2.9 and nightly. Kindly find the gist of [here](https://colab.research.google.com/gist/tilakrayal/fb8e4e27b072dd68cc4bc9ec89cc1499/56863.ipynb).

@7NoteDancing, We can see the file was loaded and saved on the drive. As the warning mentioned, please try to compile the file manually and try to load. Kindly find...

@hans-ekbrand, Could you please provide the source from where you are trying to install the tensorflow. It helps us to analyse and provide the CUDA, cudNN and the bazel version...

@eeDigitalSeal, I tried to provide the alternative approach for `stddevs` and executed the code without any issues with the expected result. Kindly find the gist of it [here](https://colab.research.google.com/gist/tilakrayal/9127572d30033834a944bd4152e4dc37/untitled450.ipynb). ``` The...

@gowthamkpr, I was able to reproduce the issue on tensorflow v2.8, v2.9 and nightly. Kindly find the gist of it [here](https://colab.research.google.com/gist/tilakrayal/fbe587381aa5e0b5658951552a275cd7/56842.ipynb).

@never-to-never, The `Dynamic Shape Mode` requires changes in many aspects regarding current XLA design & implementation. And MLIR provides an opportunity to make things easier. In order to support "Dynamic...

@never-to-never, The compilation realtion of TensorFlow graphs using XLA is as below. We can use `XLA` to compile and run parts of a `TensorFlow graph` at runtime. By compiling models...

@matteospanio, As suggested could you please try to use `tf.device("/gpu:0")` and also TensorFlow 2.9 is compatible with CUDA 11.2 and please take a look at tested build [configurations](https://www.tensorflow.org/install/source#gpu) and please...

Hi @aliencaocao, In order to expedite the trouble-shooting process, could you provide the complete code to reproduce the issue reported here.Thank you!

Hi @aliencaocao, Without the reproducible code, it would be difficult for us to debug the issue. In order to expedite the trouble-shooting process, could you provide a minimal code snippet...