* build artifacts. When Maven cannot find an artifact in the local
 * repository, it attempts to retrieve it from one or more remote
 * repositories.</p>
 *
 * <p>There are several types of remote repositories:</p><ul>

import static com.google.common.collect.testing.Helpers.assertEmpty;
import static com.google.common.collect.testing.Helpers.mapEntry;
import static com.google.common.collect.testing.features.CollectionSize.SEVERAL;
import static com.google.common.collect.testing.features.CollectionSize.ZERO;
import static com.google.common.collect.testing.features.MapFeature.ALLOWS_NULL_VALUES;

Using Linux containers has several advantages including **security**, **replicability**, **simplicity**, and others.

/// tip

In a hurry and already know this stuff? Jump to the [`Dockerfile` below 👇](#build-a-docker-image-for-fastapi).

///

import static com.google.common.collect.testing.features.CollectionSize.ONE;
import static com.google.common.collect.testing.features.CollectionSize.SEVERAL;
import static com.google.common.collect.testing.features.CollectionSize.ZERO;
import static com.google.common.collect.testing.features.MapFeature.ALLOWS_NULL_KEYS;

import static com.google.common.collect.testing.features.CollectionSize.ONE;
import static com.google.common.collect.testing.features.CollectionSize.SEVERAL;
import static com.google.common.collect.testing.features.CollectionSize.ZERO;
import static com.google.common.collect.testing.features.MapFeature.ALLOWS_NULL_KEYS;

# Custom Response Classes - File, HTML, Redirect, Streaming, etc.

There are several custom response classes you can use to create an instance and return them directly from your *path operations*.

Read more about it in the [FastAPI docs for Custom Response - HTML, Stream, File, others](https://fastapi.tiangolo.com/advanced/custom-response/).

You can import them directly from `fastapi.responses`:

```python
from fastapi.responses import (
    FileResponse,

Inconsistency in spelling conventions can lead to several issues:

* **Confusion for users and contributors**: Mixed spellings (e.g., "color" and "colour", "initialize" and "initialise") in the codebase create confusion about which variant to use

import static com.google.common.collect.testing.features.CollectionFeature.SUPPORTS_REMOVE;
import static com.google.common.collect.testing.features.CollectionSize.SEVERAL;
import static com.google.common.collect.testing.features.CollectionSize.ZERO;
import static com.google.common.collect.testing.testers.ReflectionFreeAssertThrows.assertThrows;
import static java.util.Collections.singleton;

running the `tf.train.Server`.

## Untrusted inputs during training and prediction

TensorFlow supports a wide range of input data formats. For example it can
process images, audio, videos, and text. There are several modules specialized
in taking those formats, modifying them, and/or converting them to intermediate
formats that can be processed by TensorFlow.

These modifications and conversions are handled by a variety of libraries that

import org.apache.maven.project.MavenProject;

/**
 * A build context that matches a Maven project to a given task segment, and the session to be used.
 * <p>
 * A note to the reader;
 * </p>
 * <p>
 * There are several issues/discussions regarding how "aggregator" plugins should be handled.
 * Read for instance http://docs.codehaus.org/display/MAVEN/Deterministic+Lifecycle+Planning
 * </p>
 * <p>