function is outlined to make the allocations inline in the caller // rather than happen on the heap. var out [{{ .Prefix }}ElementLen]byte return e.bytes(&out) } func (e *{{ .Element }}) bytes(out *[{{ .Prefix }}ElementLen]byte) []byte { var tmp {{ .Prefix }}NonMontgomeryDomainF {{ .Prefix }}FromMontgomery(&tmp, &e.x) {{ .Prefix }}ToBytes(out, (*{{ .Prefix }}UntypedFieldElement)(&tmp)) {{ .Prefix }}InvertEndianness(out[:]) return out[:] } // SetBytes sets e = v, where v is a big-endian {{ .BytesLen }}-byte...

function is outlined to make the allocations inline in the caller // rather than happen on the heap. var out [{{ .Prefix }}ElementLen]byte return e.bytes(&out) } func (e *{{ .Element }}) bytes(out *[{{ .Prefix }}ElementLen]byte) []byte { var tmp {{ .Prefix }}NonMontgomeryDomainF {{ .Prefix }}FromMontgomery(&tmp, &e.x) {{ .Prefix }}ToBytes(out, (*{{ .Prefix }}UntypedFieldElement)(&tmp)) {{ .Prefix }}InvertEndianness(out[:]) return out[:] } // SetBytes sets e = v, where v is a big-endian {{ .BytesLen }}-byte...

* `prefix`: to set the path prefix for a router. Up to now, this was only possible when calling `include_router`.
* `tags`: OpenAPI tags to apply to all the *path operations* in this router.

For example:

```Python
from fastapi import APIRouter, Depends


async def some_dependency():
    return

        `padded_batch()`. This functionality can be enabled through
        `tf.data.Options()`.
    *   Support cancellation of long-running `reduce`.
    *   Now we use `dataset` node name as prefix instead of the op name, to
        identify the component correctly in metrics, for pipelines with repeated
        components.
    *   Improve the performance of datasets using `from_tensors()`.