Large models. Smaller machines.
Fine-tune Mixture-of-Experts models on a single GPU. Start with Loggetta: check the hardware, plan the run, and keep a report of what happened.
pip install loggetta
loggetta inspect
loggetta plan Qwen/Qwen3-30B-A3B --seq 2048A result you can check
Unsloth spends 1.92× e4b's GPU time per training step on Qwen3-30B-A3B / RTX 5090, and 2.80× its wall-clock time on an AMD EPYC 7713 host. Matched adapters, initialization, tokens, and software; comparable held-out loss. Unsloth peaked lower (24.27 vs 26.16 GB).
GPU time is the host-independent figure: Unsloth spends more of each step on the CPU, so its wall-clock ratio grows on a slower host. Measured on torch 2.12.1+cu130 / transformers 5.5.0.
What you can use today
Loggetta 0.7.0 plans and runs supported single-GPU MoE training on your own data, and keeps reusable adapters (guide). It also plans serving placement; server launch uses the runtime separately.
Check your model
Start with the supported and tested model matrix. Memory estimates are checked against real runs; they are estimates, not a guarantee that every workload fits.
One install, three projects
| Project | Job |
|---|---|
| Loggetta | Inspect, plan, run, and save reports. |
| experts4bit-qlora | Load, fine-tune, serve, and offload model weights. |
| grouped-nf4-gemm | Compute on packed weights and move expert data. |
Contributions upstream
Unsloth Zoo merged PR #852 on 15 October 2026. It adds opt-in mask_out_tokens support to response-only training: chosen token sequences, such as </think>, are excluded from loss without altering default behavior. The change covers list/tensor data, vision collators, and MLX delegation, with regression tests. This records an upstream merge, not a packaged-release claim.
Choose your next step
- Plan and run training
- Solve a memory, training, or kernel problem
- Check model and hardware requirements
- Discuss an integration or research project
Versions, evidence, and source details
experts4bit-qlora v0.58.1 · current status → — claims tiered in the public register: 465 measured · 8 measured-private · 3 open · 19 superseded · 9 retired. grouped-nf4-gemm v0.51.0 · current status → — claims tiered in the public register: 11 confirmed · 36 measured · 3 measured-private · 1 projected · 2 open · 1 superseded · 4 retired.
Evidence vocabulary
- confirmed
- a public receipt plus an independent confirmation cell (kernel register)
- verified
- a public receipt plus an independent check (runtime register)
- measured
- the run happened and the receipt is public in the repository
- measured-private
- the run happened and the receipt lives outside the repository; NOT publicly reproducible from the repository alone
- projected
- arithmetic from measured inputs, not a run
- open
- a statement the documents make that has no evidence either way yet; it backs no capability
- superseded
- replaced by a later claim that names it; kept as the record, never current
- retired
- withdrawn; kept as the record, never current
Runtime/kernel compatibility: experts4bit-qlora >=0.55.0 → grouped-nf4-gemm >=0.43.0 (extra: fast).
Task guides
Data and documentation
Site index · Package data · Compatibility manifest · Source record · Latest research data · AI reading guide
OpenVINO upstream fixes · Research notes · Hugging Face
rendered from experts4bit-qlora 0.58.1 · source: pjordanandrsn/experts4bit-qlora@605d21e67958 · checked 2026-10-25
rendered from grouped-nf4-gemm 0.51.0 · source: pjordanandrsn/grouped-nf4-gemm@30adb1bfa76f · checked 2026-10-25