Google translation offerings
Google Cloud translation offerings enable your websites and applications to translate text with state-of-the-art translation models. Depending on your use case, you can use the Cloud Translation API to apply those models programmatically, or access them through Vertex AI.
Here is a summary of the different translation options:
Cloud Translation - Basic API provides quick, plug-and-play access to the standard Neural Machine Translation (NMT) model.
Cloud Translation - Advanced API provides access to the standard Neural Machine Translation (NMT) model and to an advanced Translation LLM (TLLM) model. It supports customizations of the NTM model, and Adaptive Translation with the TLLM model that can achieve customization and localization with a limited number of example translation pairs, optionally in multi-sentence context windows. Cloud Translation - Advanced API also has features such as glossary support, and document, batch, and real-time translation.
Vertex AI Model Garden Platform lets you access and customize other models directly. This platform is for those who want to build their own end-to-end translation MLOps pipeline. You can use it to work with foundation models for translation-adjacent tasks, and customize them through Vertex fine tuning and embeddings. It supports Neural Machine Translation (NMT) custom models and a Translation LLM (TLLM) supervised fine-tuning model.
Vertex AI Studio (UI experience) provides a user-friendly interface for rapid prototyping, quick experimentation, and fine-tuning. It lets you choose between the NMT and Translation LLM (TLLM) models, and use the UI to upload example translation pairs for Adaptive Translation or to fine-tune and customize your models.
This diagram further illustrates Cloud Translation options:
See the Cloud Translation API overview for more information about API operations.