`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 policy is a `"mixed_float16"` policy. If you...

* ✨ Add support for wrapped functions (e.g. `@functools.wraps()`) used with forward references. PR [#5077](https://github.com/fastapi/fastapi/pull/5077) by [@lucaswiman](https://github.com/lucaswiman).
* ✨ Handle wrapped dependencies. PR [#9555](https://github.com/fastapi/fastapi/pull/9555) by [@phy1729](https://github.com/phy1729).

### Fixes

g.fixedName != byteorder.BEUint32(nonce[:4]) { panic("crypto/cipher: incorrect module name given to GCMWithCounterNonce") } counter -= g.start // Ensure the counter is monotonically increasing. if counter == math.MaxUint64 { panic("crypto/cipher: counter wrapped") } if counter < g.next { panic("crypto/cipher: counter decreased") } g.next = counter + 1 fips140.RecordApproved() return g.g.sealAfterIndicator(dst, nonce, plaintext, data) } func (g *GCMWithCounterNonce) Open(dst, nonce, ciphertext, data []byte)...