LLaMA Factory — Model fine-tuning

Overview

LLaMA Factory is a language-model fine-tuning project. Developers configure experiments using supported models, data formats and training methods.

Key features

Model fine-tuning: Configure training and data parameters
Experimental management: comparison of different training settings

Suitable uses

Developers and researchers who are suitable for language model training or fine-tuning experiments need to first master the relevant development environment.

Getting started

Read the corresponding version of the document first, confirming model licences, data formats and resource requirements. A small sample was used to validate the process and document configuration and results; after completion of the independent evaluation, the quality of representation was not improved through training.

Things to check

Training requires the rational management of data authorizations, computing resources and checkpoints. Models and tool versions will influence the recurrence results.

Pricing and availability

Account requirements, pricing, export limits and regional availability can change. Check the current official terms before using the service. This directory does not promise a particular free allowance or price.

References

Official website and product information