Implementation:Mistralai Client python JSONL Data Format

Knowledge Sources	Mistral Client Python Mistral Fine-tuning
Domains	Fine_Tuning, Data_Preparation
Last Updated	2026-02-15 14:00 GMT

Overview

Interface specification for preparing JSONL training data files for Mistral fine-tuning jobs.

Description

This is a Pattern Doc — it documents the user-implemented data preparation pattern. Training data must be formatted as JSONL (JSON Lines) where each line is a valid JSON object containing a messages array. The messages follow the standard chat format with role and content fields.

Usage

Prepare JSONL files before calling client.files.upload(). Each line should be a self-contained training example.

Code Reference

Source Location

• Repository: client-python
• File: N/A (user code pattern)
• Reference: examples/mistral/jobs/jobs.py

Interface Specification

import json

def create_training_jsonl(examples: list, output_path: str) -> str:
    """
    Create a JSONL file from training examples.

    Args:
        examples: List of conversation dicts with 'messages' key
        output_path: Path to write the .jsonl file

    Returns:
        Path to the created file
    """
    with open(output_path, "w") as f:
        for example in examples:
            f.write(json.dumps(example) + "\n")
    return output_path

I/O Contract

Inputs

Name	Type	Required	Description
examples	List[Dict]	Yes	Training examples with "messages" key
output_path	str	Yes	Path to write the .jsonl file

Outputs

Name	Type	Description
.jsonl file	File	JSONL file ready for upload

Usage Examples

Create Training Data

import json

# Prepare examples
examples = [
    {
        "messages": [
            {"role": "user", "content": "What is machine learning?"},
            {"role": "assistant", "content": "Machine learning is a subset of AI..."}
        ]
    },
    {
        "messages": [
            {"role": "system", "content": "You are a coding expert."},
            {"role": "user", "content": "Write a Python hello world."},
            {"role": "assistant", "content": "print('Hello, World!')"}
        ]
    },
]

# Write JSONL file
with open("train.jsonl", "w") as f:
    for ex in examples:
        f.write(json.dumps(ex) + "\n")

Related Pages

Implements Principle

• Principle:Mistralai_Client_python_Training_Data_Preparation