of `Policy` optionally took in a `tf.compat.v1.mixed_precision.LossScale` in the constructor, which defaulted to a dynamic loss scale for the `"mixed_float16"` policy and no loss scale for other policies. In `Model.compile`, if the model's policy had a loss scale, the optimizer would be wrapped with a `LossScaleOptimizer`. With the non-experimental `Policy`, there is no loss scale associated with the `Policy`, and `Model.compile` wraps the optimizer with a `LossScaleOptimizer` if and only if the...

expected as long as it is applied to points on the prime order subgroup, like // in Ed25519. In fact, it is lost to history why RFC 8032 adopted the // irrelevant RFC 7748 clamping, but it is now required for compatibility. func (s *Scalar) SetBytesWithClamping(x []byte) (*Scalar, error) { // The description above omits the purpose of the high bits of the clamping // for brevity, but those are also lost to reductions, and are also // irrelevant to edwards25519 as they protect against a specific // implementation...

expected as long as it is applied to points on the prime order subgroup, like // in Ed25519. In fact, it is lost to history why RFC 8032 adopted the // irrelevant RFC 7748 clamping, but it is now required for compatibility. func (s *Scalar) SetBytesWithClamping(x []byte) (*Scalar, error) { // The description above omits the purpose of the high bits of the clamping // for brevity, but those are also lost to reductions, and are also // irrelevant to edwards25519 as they protect against a specific // implementation...

## 0.101.1

### Fixes

* ✨ Add `ResponseValidationError` printable details, to show up in server error logs. PR [#10078](https://github.com/tiangolo/fastapi/pull/10078) by [@tiangolo](https://github.com/tiangolo).

### Refactors