Fine-tuning
Fine-tuning job management and model import. These endpoints let you create and end training jobs, track training progress, import models from third-party providers, and retrieve training metadata.
Jobs
Create fine-tuning job
Starts a new fine-tuning job that trains a LoRA adapter on top of a supported base model using your training logs.
Stop fine-tuning job
Stops a running fine-tuning job before it completes all configured training steps. The partially trained model checkpoint is preserved.
Get training info
Returns training metadata and progress information for a fine-tuned model, including loss curves, step counts, and hyperparameter configuration.
Get training log count
Returns the number of training log entries that match the specified filter criteria.
Model Import
Import model from provider
Imports a LoRA adapter from Hugging Face into AI Studio for inference.
Get model license
Returns the license terms for a base model from the specified provider. Some models require license acceptance before use.
Preview third-party model
Returns metadata and compatibility details for a Hugging Face model before committing to a full import.
Get model import progress
Returns the current status and progress of a third-party model import job.