docs(fp8): update FP8 Storage docs#9241
Open
Pfannkuchensack wants to merge 1 commit into
Open
Conversation
PR invoke-ai#9231 routed every nn.Module — including diffusers ModelMixin — through InvokeAI's `register_forward_pre_hook` / `register_forward_hook` path, but the FP8 Storage docs still described the old `enable_layerwise_casting` implementation. Also corrects two unrelated inaccuracies the rewrite surfaced: pre-Ampere CUDA cards are not a no-op (the FP8 path gates only on `device.type == "cuda"`, and `float8_e4m3fn` is a pure storage dtype that works on any CUDA device), and the UI does not grey out the toggle based on hardware. Tell users what to include when reporting an FP8 problem so triage isn't blocked on follow-up questions: repro steps, exact model + variant, LoRA stack, partner toggles (low-VRAM, cpu_only), GPU + VRAM, OS, and the relevant log lines.
JPPhoto
approved these changes
May 27, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Refresh FP8 Storage page after hook-based loader change.
PR #9231 routed every
nn.Module— including diffusersModelMixin— through InvokeAI'sregister_forward_pre_hook/register_forward_hookpath, but the FP8 Storage docs still described the oldenable_layerwise_castingimplementation. This PR rewrites the relevant sections to match the merged behavior.Also corrects two unrelated inaccuracies the rewrite surfaced:
device.type == "cuda", andfloat8_e4m3fnis a pure storage dtype that works on any CUDA device — so a 2080 Ti gets the same ~50% VRAM win as a 3090. The "no-op" tier now covers only MPS and CPU.Adds a "Reporting an FP8 issue" section to Troubleshooting so users include the info needed for triage on first contact: repro steps, exact model + variant (full-precision vs. GGUF/NF4/int8), LoRA stack, partner toggles (Low-VRAM,
cpu_only), GPU + VRAM, OS, and the relevant log lines.Related Issues / Discussions
Follow-up to #9231 — that PR's checklist had "Documentation added / updated" unchecked.
QA Instructions
Docs-only change. Build the Starlight docs locally (or preview on the docs deploy) and verify:
enable_layerwise_castingon the FP8 Storage page.Merge Plan
Straight merge. Docs only — no code, schema, or migration impact.
Checklist
What's Newcopy (if doing a release after this PR) — n/a