Overview of AutoML Translation
AutoML Translation is a machine learning service that allows you to create custom translation models for your specific domain. You can train these models on your own data, incorporating specific terminology, phrasing, and domain-specific language. This allows for a more accurate and tailored translation solution for your business needs.
Translation API vs AutoML Translation
If you do not require a custom solution, the Translation API provides translation for over 100 languages. However, if you need a more specific translation service, AutoML Translation is the better choice.
Getting Started with AutoML Translation
Before you can use the system, you will need to set up a Google Cloud project. Here are the steps you will need to follow:
- Create a Google Cloud Account
If you are new to Google Cloud, you will need to create an account to evaluate how the products perform in real-world scenarios. At the time of writing this article, new customers are eligible to receive $300 in free credits to run, test, and deploy workloads on Google Cloud. This offer provides a great opportunity to explore the capabilities of Google Cloud’s products and services without incurring any costs upfront.
- Enable AutoML Translation and Cloud Storage APIs
To get started with Translation, you will need to enable the AutoML and Cloud Storage APIs.There are several ways to accomplish this task, but one common approach is to run the following command:
gcloud services enable storage-component.googleapis.com automl.googleapis.com storage-api.googleapis.com
- Create a Google Cloud Project
Next, create a new Google Cloud project or select an existing one. Make sure billing is enabled for your project.
- Set Up gcloud CLI
To run the commands on this page, set up the gcloud CLI in one of the following development environments:
- Cloud Shell
- Local shell
- Create a Cloud Storage Bucket
Translation requires a Google Cloud Storage bucket to store the sentence pairs that you will use to train your custom model. The bucket name must be in the format: project-id-vcm. You can create the storage bucket in the us-central1 region using the following command:
gsutil mb -p project-id -c regional -l us-central1 gs://project-id-vcm/
Preparing Your Training Data
In order to train your custom model, you will need to provide training data. This data should be specific to your domain and contain translations that you want your custom model to learn from.
Creating Datasets, Training Models, and Evaluating Results
You can begin creating your datasets, training your models, and evaluating the results once you format your training data correctly. You will accomplish this using the intuitive interface provided by GOOGLEAutoML Translation UI, which allows you to upload your data, select your features, and build your models.
Once you train and evaluate your model, you can utilize the Translation UI or the Translation API to translate content.
REST Reference and RPC Reference
For more information on the AutoML Translation REST and RPC APIs, visit the REST reference and RPC reference pages.
If you have any questions or feedback about AutoML Translation, you can visit the AutoML Translation discussion group to view discussions or post your own questions and feedback.