RomanNum3ral
/
soup
mirror of https://github.com/razor-ai/soup.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

docs: fix two broken anchor links in the layer-streaming pages 

The v0.72.2 heading change ("Layer Streaming (BETA, v0.72.0)" ->
"... v0.72.0; NF4 v0.72.2") moved its anchor, leaving the inbound pointer from
docs/training.md dead. Repointed, and took the opportunity to mention that
`quantization: 4bit` is what makes 8B fit a 4 GB card.

Also fixes a PRE-EXISTING dead self-link: the TOC entry for "FP8 Attention +
NVFP4 + Native unsloth_bnb_4bit" carried a "(v0.53.0)" suffix its heading never
had, so that anchor has never resolved.

Verified by walking every `performance-and-quantization.md#...` reference in
docs/ against the file's real headings, using GitHub's slug rule (each space
becomes one hyphen, so punctuation leaves a double hyphen — a naive
whitespace-collapsing slugger reports seven false positives here). 0 broken.

Docs-only: no version bump, no tag, does not ship to PyPI.
This commit is contained in:
Alpamys 2026-07-28 16:29:09 +05:00
parent 7c398ccdfd
commit 1c7540c314
2 changed files with 4 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
docs/performance-and-quantization.md
View File

@ -24,7 +24,7 @@
- [MoE Expert Quantization + Router-Only Training (live in v0.71.20)](#moe-expert-quantization--router-only-training-live-in-v07120)
- [Unsloth Dynamic 2.0 GGUF Ladder (v0.53.0)](#unsloth-dynamic-20-gguf-ladder-v0530)
- [KV Cache Types (v0.53.0)](#kv-cache-types-v0530)
- [FP8 Attention + NVFP4 + Native `unsloth_bnb_4bit` (v0.53.0)](#fp8-attention--nvfp4--native-unsloth_bnb_4bit-v0530)
- [FP8 Attention + NVFP4 + Native `unsloth_bnb_4bit` (v0.53.0)](#fp8-attention--nvfp4--native-unsloth_bnb_4bit)
- [LF / Axolotl Quant Parity (v0.53.0)](#lf--axolotl-quant-parity-v0530)
- [Advanced Save Formats (v0.53.0)](#advanced-save-formats-v0530)
- [Quant Menu II + Export Pipeline (v0.53.1)](#quant-menu-ii--export-pipeline-v0531)

5
docs/training.md
View File

@ -6,8 +6,9 @@
> **Training a model bigger than your GPU?** `training.stream_layers: true` streams the
> frozen base from CPU RAM one decoder layer at a time, so peak VRAM is bounded by one
> layer instead of the whole model —
> see [Layer Streaming](performance-and-quantization.md#layer-streaming-beta-v0720).
> layer instead of the whole model. Add `quantization: 4bit` and an 8B base fits a 4 GB
> card — see
> [Layer Streaming](performance-and-quantization.md#layer-streaming-beta-v0720-nf4-v0722).
**Contents:**