You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
TensorFlow version (you are using): N/A
Are you willing to contribute it (Yes/No): Not immediately.
Describe the feature and the current behavior/state.
In some of the Keras Core examples JAX backend has been used. IIUC this flow uses jitting via XLA. Here I assume that the lowering must generate StableHLO as an IR before its consumed by XLA. If this is truly the case, is it viable to produce StableHLO while using JAX as the backend? It will be useful for compiling the same model using IREE instead of XLA.
Existing flow: Keras Core model --> JAX.JIT (XLA)
Desired flow: Keras Core model --> JAX.JIT --> Side outputs is StableHLO --> IREE
System information.
TensorFlow version (you are using): N/A
Are you willing to contribute it (Yes/No): Not immediately.
Describe the feature and the current behavior/state.
In some of the Keras Core examples JAX backend has been used. IIUC this flow uses jitting via XLA. Here I assume that the lowering must generate StableHLO as an IR before its consumed by XLA. If this is truly the case, is it viable to produce StableHLO while using JAX as the backend? It will be useful for compiling the same model using IREE instead of XLA.
Existing flow: Keras Core model --> JAX.JIT (XLA)
Desired flow: Keras Core model --> JAX.JIT --> Side outputs is StableHLO --> IREE
Will this change the current api? How?
I am not sure.
Who will benefit from this feature?
IREE users.
Contributing
The text was updated successfully, but these errors were encountered: