# Axolotl v0.19.0 — ternary BitNet training and a 16.4x tokenizer speedup

> Axolotl can now train weights down to {-1, 0, 1} with no extra dependency, and stops hanging on long chat datasets.

- Org: Axolotl AI
- Date: 2026-09-10
- Category: repo, tool
- Source: https://github.com/axolotl-ai-cloud/axolotl/releases/tag/v0.19.0
- Canonical page: https://ai-tldr.dev/releases/axolotl-v0-19-0/

## In short

Axolotl v0.19.0 adds native BitNet b1.58 ternary quantization-aware training, a declarative way to add model architectures, and two new optimizers. Tokenizing an 80-turn chat drops from 1,577.6 ms to 96.4 ms.

## What it is

Ternary quantization-aware training arrives in Axolotl v0.19.0 behind a single config key: `qat: weight_dtype: ternary` restricts a model's weights to {-1, 0, 1} in the BitNet b1.58 style, natively and with no extra dependency. The release ships 72 commits since v0.18.0 and includes `examples/neutrino/`, the two-stage recipe Fermion used for Neutrino-1 — ternary continued pretraining, then instruction tuning on top.

## How it works

Adding an architecture used to mean edits scattered across the loaders. A declarative model-support system replaces that: a model picks a family template such as `VANILLA_CAUSAL_LM` or `IMAGE_TEXT_TO_TEXT` and declares only what makes it different, with existing descriptors and configs unchanged. Two optimizers join as well — PoLoRA, a Muon-like optimizer for LoRA and QLoRA on single GPU, DDP or FSDP2, and AdamC, which scales weight decay by learning rate.

## Why it matters

The tokenization fix removes a long-standing annoyance: preprocessing an 80-turn Qwen2.5 conversation now takes 96.4 ms instead of 1,577.6 ms, a 16.4x speedup that grows with conversation length, so long multi-turn datasets no longer look like a hung job. Fine-tuning coverage also widens — Qwen3.8-Flash-Next, a 176.94B multimodal mixture-of-experts, fits on a single B300 in as little as 120 GiB with `ple_cpu_offload: true`, and Ling 3.0, Muse Glimmer 30B, Cohere North Micro Vision Instruct, Shieldstral, Qwen3.8 27B and Nemotron 3.5 Lightning 30B-A3B each ship example configs.

## Who it's for

fine-tuning practitioners

## Try it

pip install -U axolotl

## Links

- [Release Notes](https://github.com/axolotl-ai-cloud/axolotl/releases/tag/v0.19.0)
- [Repository](https://github.com/axolotl-ai-cloud/axolotl)
- [QAT Docs](https://docs.axolotl.ai/docs/qat.html)
- [GRPO Docs](https://docs.axolotl.ai/docs/grpo.html)