Axolotl

In this catalog it is the Fine-tuning option when you need to adapt base weights to domain data. Keep Axolotl on the observe/propose side of the loop: useful signals and drafts, with human or policy authority still required before anything irreversible runs.

Axolotl fine-tunes LLMs from YAML configs (LoRA/QLoRA/full) across many model families.

1. Author a training YAML (model, data, method).
2. Launch training; checkpoints write to disk.
3. Evaluate against held-out metrics.
4. Export adapters/weights for serving.

Related: unsloth · llamafactory · vllm · topics/04-data

• Use when: Axolotl is a popular open toolkit for fine-tuning LLMs with YAML configs covering LoRA, QLoRA, full FT, and many model families matches the bottleneck and you specifically need to adapt base weights to domain data.
• Skip when: another pack already covers the need with less ops surface, or Axolotl would become a second place people retype truth.
• Prefer instead: compare Unsloth, LlamaFactory, vLLM before locking the cell.

• Ops: confirm license, hosting, and release cadence for Axolotl (axolotl-ai-cloud/axolotl) before relying on it in a live corridor.
• Integration: bad credentials, scopes, or schema drift produce empty retrievals or runaway tool/model spend without an obvious UI alarm.
• Scope: Fine-tuning components support agents; they are not the company ledger or the approval authority.