Ir al contenido
Menú
Español

Caso de uso

Customise a model

Adapt a general model to a specific task or style, and keep the cost of doing so proportionate.

Ver el stack recomendado, con motivos y alternativas →

Lo que se necesita

En el orden en que las adoptarías. Cada paso se abre en el grafo de conocimiento.

  1. 1 Base model

    GPT

    model

    Family of decoder-only Transformer language models trained with a generative objective.

  2. 2 Method

    Fine-tuning

    technique

    Continue training a pre-trained model on task-specific data to adapt its behaviour.

  3. 3 Efficient alternative

    LoRA

    technique

    Fine-tuning by training small low-rank adapter matrices instead of all weights.

  4. 4 Serving

    KV Cache

    technique

    Cached attention keys and values that make generating one token at a time affordable.