Expand description
The Fine-tuning API: manage fine-tuning jobs via /fine_tuning.
![warn] This module is untested! If you encounter any issues, please report them on the repository.
A fine-tuning job trains a model on a JSONL training file (and an
optional validation file) uploaded through the
files API. See
the OpenAI Fine-tuning guide.
Submodules: jobs (create / list / retrieve / cancel, plus the
events and checkpoints sub-resources) and models (restore a
fine-tuned model after a timeout).
The POST /fine_tuning/alpha/permissions endpoints (GRPO reward
functions) are not implemented; they are reserved for the
enterprise-only alpha program.
Modules§
- jobs
- The
fine_tuning::jobsendpoint family module tree. - models
- Fine-tuning model endpoints under
/fine_tuning/models.
Structs§
- Fine
Tune Metrics - The per-step training/validation loss statistics of a fine-tuning job.
- Fine
Tuning Job - The result files of a completed fine-tuning job.
- Hyperparameters
- The hyperparameters of a fine-tuning job.
Enums§
- Fine
Tuning JobStatus - The lifecycle status of a fine-tuning job.