mirror of https://github.com/razor-ai/soup.git
115f29df15
Visitors arrive from the Show HN headline and the first screen is all prose. Adds a 14.5s GIF directly under the 8B-on-4GB claim, above "Why Soup?". Segment 44.0-60.0s of the demo, chosen off the .vtt cue list and confirmed frame by frame rather than from the brief's estimate: the scene cut sits between 47 and 48s, so the clip opens on 3.5s of the static "Layer streaming BETA" pre-flight panel (3.60 GB store across 32 layers, 2 x 113 MB VRAM buffers, Training started!) and then runs the measurement card up to its settled 3.32 GB / 119.6 tok/s. Both halves read without sound or context. Encoding: two-pass palettegen/paletteuse so terminal colours survive, 960px wide, 10 fps, dither=none, diff_mode=rectangle -> 4.63 MB. Readability was the binding constraint, so the budget was met by cutting duration (16s -> 14.5s) and fps (12 -> 10) rather than width; a bayer-dithered cut of the same clip came to 5.08 MB and was dropped. Panel lines verified legible by opening the generated GIF, not assumed from the source resolution. Caption numbers are from benchmarks/gate-v0.72.2-nf4.md line 314 (the re-measurement through shipped code), matching the claim line above it. "What's New" untouched. README 440 -> 445 lines. |
||
|---|---|---|
| .. | ||
| assets | ||
| 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, reward-verifier synthesis |
| 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, layer streaming, 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.