Fine-tune LLMs from one YAML. Layer streaming trains an 8B model on a 4 GB laptop GPU.
Python cliconsumer-gpudpofine-tuningggufhuggingfacellmllmopslocal-ailocal-llmloralow-vrammachine-learningollamapeftpythonpytorchqlorasfttransformers
docs(contributors): #2140 [skip ci]
Alpamysdbce10eebb13
diff --git a/CONTRIBUTORS.md b/CONTRIBUTORS.mdindex bd4908d2..fb332610 100644--- a/CONTRIBUTORS.md+++ b/CONTRIBUTORS.md@@ -940,6 +940,7 @@ Listed by first contribution. PR numbers link the work.- Took item 2 of #1048 (the issue stays open for save and resume and for the both-spellings test): under `training.stream_layers` a LoRA target on the streamed `lm_head` or `embed_tokens` was refused at config load by literal name only, so a regex or dotted spelling loaded and then died at the first default save, while a tied model, which round-trips cleanly, was refused. They claimed the item on the thread with a plan, asked before writing the PR how the dotted `model.embed_tokens` spelling should go (it saves only through an accident of peft's literal-name check) and pinned the maintainer's answer as an intentional refusal, then moved the check into `build_streamed_model` after `get_peft_model()`, where it looks at the modules peft actually wrapped and fires only when the model streams them. **The part worth copying is the controls**: the tied cases train one real step, save with the defaults and reload every LoRA tensor `torch.equal` after moving the fresh build off its init values, and the untargeted untied build is kept as a control. 12 tests (11 fail on `main`), 10 of 10 mutations killed under peft 0.20.0 and 0.21.2 ([#1048](https://github.com/MakazhanAlpamys/Soup/issues/1048), [#2082](https://github.com/MakazhanAlpamys/Soup/pull/2082))- Took the harness half of #1693 (the issue stays open for the measured table): `benchmarks/harness/stream_loss_chunk_vram.py` builds a chunked and an unchunked arm through the real `SFTTrainerWrapper.setup`, runs one streamed step each, and prints the pre-flight's own prediction, the free VRAM it compared against and the real peak per length. **They said up front they have no GPU, so it carries a CPU `--self-test`** that runs the same protocol on a tiny local Llama and checks 13 things (the realised length, the pre-flight's inputs and arithmetic, the hook on the chunked arm only, the verdicts and exit codes), and every figure in the new sizing text is labelled as the pre-flight's arithmetic, not a measurement. Review found the teardown of each arm unpinned and one repo scanner red on a `print()` assertion; the follow-up commit added a test that counts both teardown calls and the scanner's marker the same day. ([#1693](https://github.com/MakazhanAlpamys/Soup/issues/1693), [#2102](https://github.com/MakazhanAlpamys/Soup/pull/2102))- Took #2047: GRPO could only score with a rule, so a task with no checkable answer had no scorer. They posted the plan first, took the maintainer's twelve corrections, then added `training.reward_fn: judge_rank`: one judge call sees all `num_generations` completions of a prompt, in a fresh seeded order, and the group scores `None` instead of 0.0 when the judge gives no usable answer. A required `grpo_judge_max_calls` is checked after the data loads and before the model does, and failed or all-equal groups are logged under `judge/`. **The part worth copying is the questions**: the PR lists the five choices it made without a ruling and the one mutation that survives, and the follow-up commit closed that survivor by deleting the dead flag instead of arguing for it. ([#2047](https://github.com/MakazhanAlpamys/Soup/issues/2047), [#2117](https://github.com/MakazhanAlpamys/Soup/pull/2117))+ - Took item 3 of #797 (the issue stays open for the vision/audio k-bit test): `training.gradient_checkpointing: true` or a tier did nothing on `distill`, `classifier`, `asr` and `unlearn`, which never read it. They passed it to the HF arguments of the first three and, for unlearn's own loop, called `gradient_checkpointing_enable()` on the LoRA policy model, with tests that drive each real `setup()` on CPU and count the backward recompute, and a table of opposite-value mutations per site. 36 tests (26 fail on `main`), 10 of 10 mutations killed by named tests, and the adapter weights and first-step loss check out with checkpointing on. ([#797](https://github.com/MakazhanAlpamys/Soup/issues/797), [#2140](https://github.com/MakazhanAlpamys/Soup/pull/2140))- **Nariman Abdikarim** ([@Nari-Ab](https://github.com/Nari-Ab))- Sanitized model and dataset paths across CLI display sinks and both short-name branches, with 17 portable regressions covering refusals, adapter loading, saved outputs and real Elo/leaderboard table consumers. Independent maintainer validation passed all 17 cases before and after five source undos that caused 1/1/1/3/4 failures; all eight required hosted checks passed ([#1864](https://github.com/MakazhanAlpamys/Soup/issues/1864), [#1868](https://github.com/MakazhanAlpamys/Soup/pull/1868))