## Otimizar Desempenho da Resposta - Modelo de Resposta - Tipo de Retorno { #optimize-response-performance-response-model-return-type }

Cuando creas una *path operation* en **FastAPI**, normalmente puedes devolver cualquier dato desde ella: un `dict`, una `list`, un modelo de Pydantic, un modelo de base de datos, etc.

Si declaras un [Response Model](../tutorial/response-model.md) FastAPI lo usará para serializar los datos a JSON, usando Pydantic.

Si no declaras un response model, FastAPI usará el `jsonable_encoder` explicado en [JSON Compatible Encoder](../tutorial/encoder.md) y lo pondrá en un `JSONResponse`.

## Retorno Adicional com `model` { #additional-response-with-model }

Você pode fornecer o parâmetro `responses` aos seus *decoradores de caminho*.

Este parâmetro recebe um `dict`, as chaves são os códigos de status para cada retorno, como por exemplo `200`, e os valores são um outro `dict` com a informação de cada um deles.

## Optimiser la performance des réponses - Modèle de réponse - Type de retour { #optimize-response-performance-response-model-return-type }

      <plugin>
        <groupId>org.codehaus.modello</groupId>
        <artifactId>modello-maven-plugin</artifactId>
        <configuration>
          <version>1.1.0</version>
          <basedir>${project.basedir}/../../api/maven-api-toolchain</basedir>
          <velocityBasedir>${project.basedir}/../../src/mdo</velocityBasedir>
          <models>
            <model>src/main/mdo/toolchains.mdo</model>
          </models>
        </configuration>

      <plugin>
        <groupId>org.codehaus.modello</groupId>
        <artifactId>modello-maven-plugin</artifactId>
        <configuration>
          <basedir>${project.basedir}/../../api/maven-api-model</basedir>
          <velocityBasedir>${project.basedir}/../../src/mdo</velocityBasedir>
          <version>4.1.0</version>
          <models>
            <model>src/main/mdo/maven.mdo</model>
          </models>
          <params>

      <plugin>
        <groupId>org.codehaus.modello</groupId>
        <artifactId>modello-maven-plugin</artifactId>
        <configuration>
          <version>2.0.0</version>
          <velocityBasedir>${project.basedir}/../../src/mdo</velocityBasedir>
          <models>
            <model>src/main/mdo/settings.mdo</model>
          </models>
          <templates>
            <template>model.vm</template>
          </templates>

      <plugin>
        <groupId>org.codehaus.modello</groupId>
        <artifactId>modello-maven-plugin</artifactId>
        <configuration>
          <version>1.2.0</version>
          <velocityBasedir>${project.basedir}/../../src/mdo</velocityBasedir>
          <models>
            <model>src/main/mdo/toolchains.mdo</model>
          </models>
          <templates>
            <template>model.vm</template>
          </templates>

"""Experimental Bzlmod support for TensorFlow"""

module(
    name = "tensorflow",
    repo_name = "org_tensorflow",
)

bazel_dep(name = "bazel_features", version = "1.36.0")
bazel_dep(name = "abseil-cpp", version = "20250814.1", repo_name = "com_google_absl")
single_version_override(
    module_name = "abseil-cpp",
    patch_strip = 1,
    patches = [
        "//third_party/absl:build_dll.patch",
        "//third_party/absl:endian.patch",

## Use Case { #use-case }

Let's start with an example **use case** and then see how to solve it with this.

Let's imagine that you have some **machine learning models** that you want to use to handle requests. 🤖

The same models are shared among requests, so, it's not one model per request, or one per user or something similar.