Cohere Releases North Small Translate: A 218B MoE Translation Model That Scores 83.6 on WMT26 Across 50 Languages

Introduction

In this tutorial, you'll learn how to use Cohere's North Small Translate model for machine translation. This open-weight model can translate between 50 languages and achieves impressive scores on translation benchmarks. We'll walk through setting up the environment, installing the required libraries, and making translation requests using the Cohere API.

By the end of this tutorial, you'll be able to translate text between multiple languages using Cohere's powerful translation model, even if you're new to AI and machine translation.

Prerequisites

Before starting this tutorial, you'll need:

• A computer with internet access
• A free Cohere account (sign up at cohere.ai)
• Python 3.6 or higher installed on your system
• Basic understanding of how to open a terminal/command prompt

Step-by-Step Instructions

1. Create a Cohere Account and Get Your API Key

The first step is to sign up for a free Cohere account. Visit cohere.ai and click on "Sign Up". After creating your account, navigate to the dashboard to find your API key. This key is essential for accessing Cohere's models.

2. Install Required Python Libraries

We'll use the cohere Python library to interact with Cohere's API. Open your terminal or command prompt and run:

pip install cohere

This command installs the Cohere Python SDK, which provides an easy way to make API calls to Cohere's models.

3. Set Up Your Python Environment

Create a new Python file called translation_demo.py. In this file, we'll first import the necessary libraries and set up our API key:

import cohere

# Replace 'YOUR_API_KEY' with your actual Cohere API key
co = cohere.Client('YOUR_API_KEY')

This code creates a connection to Cohere's API using your unique API key. The cohere.Client object is how we'll interact with the translation model.

4. Test Basic Translation

Let's start with a simple translation from English to Spanish. Add the following code to your Python file:

response = co.translate(
  texts=['Hello, how are you?'],
  target_lang='es'
)

print(response.translations[0])

When you run this code, you should see "Hola, ¿cómo estás?" printed to the console. This demonstrates how easy it is to translate text using Cohere's API.

5. Translate Multiple Texts at Once

One of the advantages of using Cohere's API is that you can translate multiple sentences in a single request:

response = co.translate(
  texts=['Hello world', 'How are you today?', 'Goodbye for now'],
  target_lang='fr'
)

for translation in response.translations:
    print(translation)

This code translates three English sentences into French. You'll see the French translations printed one by one.

6. Try Different Languages

North Small Translate supports 50 languages. Let's try translating from English to German:

response = co.translate(
  texts=['The weather is nice today', 'I like to read books'],
  target_lang='de'
)

for translation in response.translations:
    print(translation)

This will translate your English text into German. Try changing the target_lang parameter to other language codes like 'it' (Italian), 'pt' (Portuguese), or 'ja' (Japanese).

7. Explore Language Codes

For the full list of supported languages and their codes, you can check Cohere's documentation. Common codes include:

• en - English
• es - Spanish
• fr - French
• de - German
• it - Italian
• pt - Portuguese
• ja - Japanese
• ko - Korean
• zh - Chinese

Remember to replace 'YOUR_API_KEY' in your code with your actual Cohere API key before running any translation requests.

8. Run Your Translation Program

Save your Python file and run it using:

python translation_demo.py

If everything is set up correctly, you should see translations printed to your console.

Summary

In this tutorial, you've learned how to use Cohere's North Small Translate model for machine translation. You've installed the necessary Python libraries, set up your API key, and made translation requests in multiple languages. The model supports 50 languages and can be used for both simple and complex translation tasks.

Remember that while the weights are free for non-commercial use, commercial applications should consider using Cohere Model Vault or RWS Language Weaver for access. This hands-on experience gives you a foundation for working with advanced translation models and can be extended to more complex applications like building translation tools or integrating translation into larger software projects.