Here's a little bit of that history.

## Alternatives

I have been creating APIs with complex requirements for several years (Machine Learning, distributed systems, asynchronous jobs, NoSQL databases, etc), leading several teams of developers.

As part of that, I needed to investigate, test and use many alternatives.

MinIO uses [`klauspost/compress/s2`](https://github.com/klauspost/compress/tree/master/s2)
streaming compression due to its stability and performance.

This algorithm is specifically optimized for machine generated content.
Write throughput is typically at least 500MB/s per CPU core,
and scales with the number of available CPU cores.
Decompression speed is typically at least 1GB/s.

	app := cli.NewApp()
	app.Name = name
	app.Author = "MinIO, Inc."
	app.Version = ReleaseTag
	app.Usage = "High Performance Object Storage"
	app.Description = `Build high performance data infrastructure for machine learning, analytics and application data workloads with MinIO`
	app.Flags = GlobalFlags
	app.HideHelpCommand = true // Hide `help, h` command, we already have `minio --help`.
	app.Commands = commands

</p>

<p>
Reads of memory locations larger than a single machine word
are encouraged but not required to meet the same semantics
as word-sized memory locations,
observing a single allowed write <i>w</i>.
For performance reasons,
implementations may instead treat larger operations
as a set of individual machine-word-sized operations
in an unspecified order.

st ``` The MinIO deployment starts using default root credentials `minioadmin:minioadmin`. You can test the deployment using the MinIO Console, an embedded web-based object browser built into MinIO Server. Point a web browser running on the host machine to <http://127.0.0.1:9000> and log in with the root credentials. You can use the Browser to create buckets, upload objects, and browse the contents of the MinIO server. You can also connect using any S3-compatible tool, such as the MinIO Client `mc`...

### Serverspeicher

Wenn Ihr Code beispielsweise ein Machine-Learning-Modell mit **1 GB Größe** lädt und Sie einen Prozess mit Ihrer API ausführen, verbraucht dieser mindestens 1 GB RAM. Und wenn Sie **4 Prozesse** (4 Worker) starten, verbraucht jeder 1 GB RAM. Insgesamt verbraucht Ihre API also **4 GB RAM**.

!!! tip
    If you are wondering, "AlexNet", "ResNet", and "LeNet" are just names of Machine Learning <abbr title="Technically, Deep Learning model architectures">models</abbr>.

### Declare a *path parameter*

Then create a *path parameter* with a type annotation using the enum class you created (`ModelName`):

!!! tip "Astuce"
    Pour ceux qui se demandent, "AlexNet", "ResNet", et "LeNet" sont juste des noms de <abbr title="Techniquement, des architectures de modèles">modèles</abbr> de Machine Learning.

### Déclarer un paramètre de chemin

Créez ensuite un *paramètre de chemin* avec une annotation de type désignant l'énumération créée précédemment (`ModelName`) :

```Python hl_lines="16"

build:rbe_cross_compile_base --remote_instance_name=projects/tensorflow-testing/instances/default_instance

# XLA depends on some local Python headers that are configured as Genrule. They
# are present on the local host machine but not on the remote execution machine,
# leading to build failures. To resolve the issue, the following line is added
# to make sure all Genrule targets are excuted locally.
build:rbe_cross_compile_base_xla --config=rbe_cross_compile_base

!!! tip "Consejo"
    Si lo estás dudando, "AlexNet", "ResNet", y "LeNet" son solo nombres de <abbr title="Técnicamente, arquitecturas de modelos de Deep Learning">modelos</abbr> de Machine Learning.

### Declara un *parámetro de path*

Luego, crea un *parámetro de path* con anotaciones de tipos usando la clase enum que creaste (`ModelName`):

```Python hl_lines="16"