Principle:Mistralai Client python Finetuning Job Creation
| Knowledge Sources | |
|---|---|
| Domains | Fine_Tuning, Model_Training |
| Last Updated | 2026-02-15 14:00 GMT |
Overview
A job orchestration pattern that creates and configures a fine-tuning training job with a base model, training data, and hyperparameters.
Description
Finetuning Job Creation initiates a server-side training job that customizes a base Mistral model on user-provided training data. The job requires specifying a base model (e.g., "open-mistral-7b"), training_files (uploaded file IDs), and hyperparameters (learning rate, epochs, batch size). Optional settings include validation files, W&B integration, and a custom model name suffix.
Usage
Use this principle after uploading training data files. Configure hyperparameters based on dataset size and desired fine-tuning depth. The job can be set to auto_start or started manually after creation.
Theoretical Basis
Fine-tuning adapts a pre-trained model to a specific task:
- Base model provides general language understanding
- Training data provides task-specific examples
- Hyperparameters control the training process (learning rate, epochs)
- The result is a new model checkpoint that can be used for inference