* If you use masked losses with Keras the loss values may be different in TensorFlow `2.12` compared to previous versions.
 * In cases where the mask was previously ignored, you will now get an error if you pass a mask with an incompatible shape.

### Major Features and Improvements

`tf.keras`:

necessarily // corresponding) index. The memory beyond the slice length is ignored. func AnyOverlap(x, y []byte) bool { return len(x) > 0 && len(y) > 0 && uintptr(unsafe.Pointer(&x[0])) <= uintptr(unsafe.Pointer(&y[len(y)-1])) && uintptr(unsafe.Pointer(&y[0])) <= uintptr(unsafe.Pointer(&x[len(x)-1])) } // InexactOverlap reports whether x and y share memory at any non-corresponding // index. The memory beyond the slice length is ignored. Note that x and y can // have different lengths and still not have...

necessarily // corresponding) index. The memory beyond the slice length is ignored. func AnyOverlap(x, y []byte) bool { return len(x) > 0 && len(y) > 0 && uintptr(unsafe.Pointer(&x[0])) <= uintptr(unsafe.Pointer(&y[len(y)-1])) && uintptr(unsafe.Pointer(&y[0])) <= uintptr(unsafe.Pointer(&x[len(x)-1])) } // InexactOverlap reports whether x and y share memory at any non-corresponding // index. The memory beyond the slice length is ignored. Note that x and y can // have different lengths and still not have...

* ⬆ Update sqlalchemy requirement from <1.4.43,>=1.3.18 to >=1.3.18,<2.0.33. PR [#11979](https://github.com/fastapi/fastapi/pull/11979) by [@dependabot[bot]](https://github.com/apps/dependabot).
* 🔊 Remove old ignore warnings. PR [#11950](https://github.com/fastapi/fastapi/pull/11950) by [@tiangolo](https://github.com/tiangolo).