Réglage Fin (Fine-tuning)
Le fine-tuning est le processus qui consiste à prendre un modèle d'apprentissage automatique pré-entraîné et à l'entraîner davantage sur un ensemble de données spécifique afin d'améliorer ses performances sur une tâche particulière.
Le fine-tuning exploite l'apprentissage par transfert, en adaptant un modèle initialement entraîné sur un vaste ensemble de données (par exemple, un grand modèle linguistique entraîné sur un corpus de texte massif) à une tâche ou un ensemble de données plus spécialisés. Cela implique d'ajuster les pondérations du modèle à l'aide des nouvelles données, souvent avec un taux d'apprentissage plus faible que lors de l'entraînement initial. Le fine-tuning permet d'obtenir des performances élevées avec moins de données et de ressources informatiques par rapport à l'entraînement d'un modèle à partir de zéro.
graph LR
Center["Réglage Fin (Fine-tuning)"]:::main
Pre_machine_learning["machine-learning"]:::pre --> Center
click Pre_machine_learning "/terms/machine-learning"
Pre_large_language_model["large-language-model"]:::pre --> Center
click Pre_large_language_model "/terms/large-language-model"
Center --> Child_lora["lora"]:::child
click Child_lora "/terms/lora"
Center --> Child_rlhf["rlhf"]:::child
click Child_rlhf "/terms/rlhf"
Rel_front_running["front-running"]:::related -.-> Center
click Rel_front_running "/terms/front-running"
Rel_inference["inference"]:::related -.-> Center
click Rel_inference "/terms/inference"
classDef main fill:#7c3aed,stroke:#8b5cf6,stroke-width:2px,color:white,font-weight:bold,rx:5,ry:5;
classDef pre fill:#0f172a,stroke:#3b82f6,color:#94a3b8,rx:5,ry:5;
classDef child fill:#0f172a,stroke:#10b981,color:#94a3b8,rx:5,ry:5;
classDef related fill:#0f172a,stroke:#8b5cf6,stroke-dasharray: 5 5,color:#94a3b8,rx:5,ry:5;
linkStyle default stroke:#4b5563,stroke-width:2px;
🧠 Test de connaissances
🧒 Explique-moi comme si j'avais 5 ans
It's like taking a chef who knows how to cook many things (pre-trained model) and teaching them your specific family recipes (new dataset) so they become great at cooking just your favorite dishes.
🤓 Expert Deep Dive
Fine-tuning operates on the principle that representations learned on large-scale, diverse datasets capture fundamental patterns applicable to related tasks. In deep learning, this typically involves adjusting the weights of a pre-trained network (e.g., ResNet, BERT) using backpropagation on a target dataset. The learning rate is often set significantly lower than during pre-training to avoid drastic weight updates that could disrupt the learned features. Layer freezing is a common strategy: earlier layers capturing low-level features (e.g., edges, textures in images; word embeddings in text) are often frozen, while later layers capturing more task-specific features are fine-tuned. Alternatively, adapter modules can be inserted between layers, allowing task-specific parameters to be learned while keeping the original model weights fixed. The effectiveness relies heavily on the similarity between the pre-training and fine-tuning data distributions and tasks. Domain shift can necessitate more extensive fine-tuning or different adaptation strategies. Overfitting remains a primary concern, especially with very small target datasets, often mitigated by regularization techniques or early stopping.