Axolotl — fine-tuning as configuration

Everything so far scripts training in Python. Axolotl takes the infrastructure-as-code stance: an entire fine-tune — base model, dataset paths and formats, LoRA hyperparameters, sequence lengths, multi-GPU strategy (DeepSpeed/FSDP) — declared in one YAML file:

base_model: Qwen/Qwen2.5-7B-Instruct
load_in_4bit: true
adapter: qlora
lora_r: 16
datasets:
  - path: train_clean.jsonl
    type: chat_template
num_epochs: 2
learning_rate: 2e-4

axolotl train config.yml

The value proposition is threefold: reproducibility (the YAML is the experiment — diffable, reviewable, re-runnable), the community recipe library (proven configs for most model families, so you start from a working baseline instead of a blank page), and multi-GPU without tears (the DeepSpeed/FSDP incantations that make 70B fine-tunes work across 8 GPUs are config lines, not a systems project). It's built on the same PEFT/TRL stack underneath — a declarative skin over the machinery you already understand.

Summary

YAML-driven fine-tuning on the PEFT/TRL stack: reproducible experiments, community recipes, built-in multi-GPU.

Mental model

docker-compose for training runs — declare the run, don't script it.

Mistakes to avoid

Using it as a magic box before you've written the raw pipeline once (you have — Lesson 6 — which is why its YAML keys will all read as old friends), and copying a community config wholesale without rechecking the dataset type against your actual format: Topic 15's template mismatch wears YAML clothing too.

Exercise

Translate your entire Lesson 6 capstone script into an Axolotl YAML — every SFTConfig argument has a YAML twin. The act of translation is a review of everything Module 3 taught, in fifteen minutes.