* «the form model»: «das Formularmodell»
* «the form body»: «der Formularbody»
* «the header»: «der Header»
* «the headers» (plural): «die Header»
* «in headers» (plural): «in Headern»
* «the forwarded header»: «der Forwarded-Header»
* «the lifespan event»: «das Lifespan-Event»
* «the lock»: «der Lock»
* «the locking»: «das Locking»
* «the mobile application»: «die Mobile-Anwendung»
* «the model object»: «das Modellobjekt»

### Docs

* 📝 Add documentation for Behind a Proxy - Proxy Forwarded Headers, using `--forwarded-allow-ips="*"`. PR [#14028](https://github.com/fastapi/fastapi/pull/14028) by [@tiangolo](https://github.com/tiangolo).

                                "application/json": {
                                    "schema": {
                                        "title": "Response No Response Model Annotation Forward Ref List Of Model No Response Model Annotation Forward Ref List Of Model Get",
                                        "type": "array",
                                        "items": {"$ref": "#/components/schemas/User"},

from contextlib import AsyncExitStack, contextmanager
from copy import copy, deepcopy
from dataclasses import dataclass
from typing import (
    Annotated,
    Any,
    Callable,
    ForwardRef,
    Optional,
    Union,
    cast,
)

import anyio
from fastapi import params
from fastapi._compat import (
    ModelField,
    RequiredParam,
    Undefined,
    _regenerate_error_with_loc,

# This is the same as the official TensorFlow builds.
# See https://developer.nvidia.com/cuda-gpus#compute
# `compute_XY` enables PTX embedding in addition to SASS. PTX
# is forward compatible beyond the current compute capability major
# release while SASS is only forward compatible inside the current
# major release. Example: sm_80 kernels can run on sm_89 GPUs but
# not on sm_90 GPUs. compute_80 kernels though can also run on sm_90 GPUs.