mirror of https://github.com/razor-ai/soup.git
aa347222a6
Docs-only — no src/ changes, so no version bump (per the CI-only/docs-only rule).
Audit of the v0.71.35 release against the checklist found four misses:
* README's docs/ index table was missing the docs/compliance.md row — a new
topic page unreachable from the front door (step 9: "if the release adds a
whole new topic area, add/adjust the row in the README's docs/ index table").
It was only added to docs/README.md.
* The GGUF toolchain was never documented, despite the slot's own plan calling
it out ("Risk: MSVC build variance — document the exact toolchain; keep the
llama.cpp tag pinned"). Added the exact, verified build commands + the
VS2022 component actually required, the single- vs multi-config binary
layouts, and an explicit warning never to install llama.cpp's
requirements.txt (it pins torch~=2.2.1 CPU and downgrades a CUDA torch —
the bug fixed in v0.71.35).
* docs/peft-and-efficiency.md still said "16 built-in templates" — stale as a
direct result of the 4 compliance templates (now 21).
* Stale counts CONTRIBUTING "Test Files (307 files)" -> 313 and "existing 15
templates" -> 21 (that step also still pointed at schema.py rather than the
templates/ YAML + manifest.json registry).
The documented build command and binary path are the ones actually executed
during the release validation, not idealised.
|
||
|---|---|---|
| .. | ||
| README.md | ||
| adapters-and-governance.md | ||
| backends-and-ops.md | ||
| commands.md | ||
| compliance.md | ||
| data.md | ||
| evaluation.md | ||
| models.md | ||
| peft-and-efficiency.md | ||
| performance-and-quantization.md | ||
| serving-and-export.md | ||
| training.md | ||
README.md
Soup Documentation
The main README is the 5-minute front door. This directory holds the full
feature reference — every soup capability, grouped by area.
| Guide | Covers |
|---|---|
| Training tasks & methods | SFT, DPO/GRPO/PPO/KTO/ORPO/SimPO/IPO/BCO, tool-calling, PRM, pre-training, distillation, classification, vision/audio/TTS, unlearning, RAFT/RA-DIT, loop-hardening detectors |
| PEFT, long context & efficiency | DoRA, LoRA+, rsLoRA, VeRA, OLoRA, NEFTune, PiSSA, ReLoRA, optimizer & PEFT zoo, LLaMA Pro, GaLore, YaRN/LongLoRA, packing, curriculum, auto-tuning, depth pruning + distill-heal (soup shrink) |
| Performance & quantization | QAT, FP8, Quant Menu (I + II), KV-cache, NVFP4, save formats, Cut Cross-Entropy, gradient checkpointing, kernels, activation offloading, multi-GPU / DeepSpeed / FSDP |
| Data engineering | Formats, the Axolotl/LF-parity pipeline, data tools, synthetic generation & forge, quality scorecards, trace tooling, remote datasets, mixing, recipe DAGs |
| Evaluation & probes | Eval design/gate, eval-gated training, benchmarks, NLG metrics, calibration, Elo arena, diagnose, soup ship verdict, post-train X-ray probes, A/B, drift, tunability, soup advise |
| Serving & export | OpenAI-compatible server, batch inference, benchmarking, merge/export, Anthropic Messages endpoint, speculative decoding (train + measure your own draft), deploy autopilot, Web UI, Agent Forge |
| Adapters, registry & governance | Adapter lifecycle/management, model registry, Soup Cans, the data flywheel (soup loop), knowledge editing, steering, supply-chain controls |
| Compliance & governance quickstart | HIPAA/SOC2/EU-AI-Act/SR-11-7 init templates, provenance (BOM/attest/repro-receipt), audit log, air-gap, model-card autogen (soup card), CI gate (soup ci init) |
| Backends, platform & ops | MLX/Unsloth backends, Modal cloud GPU training, alternative hubs, HF Hub integration, autopilot, experiment tracking, plan/apply, env lockfiles, hardware-fit, completions, plugins, utility commands |
| Command reference | The full soup command list |
| Supported models & extras | Recommended model families, the VRAM size guide, the pip extras matrix |
Per-release notes live on the GitHub Releases page; see also the repo-root CHANGELOG.md.