Repository navigation
feat(krea2): run Krea-2 attention on SageAttention when enabled - #9662
Draft
Pfannkuchensack wants to merge 21 commits into
Draft
Pfannkuchensack wants to merge 21 commits into
Pfannkuchensack wants to merge 21 commits into
Conversation
New `attention_backend: auto | sage` setting; with `sage`, eligible PyTorch SDPA calls inside a denoising scope run on SageAttention 2.2 and every other call keeps SDPA unchanged. Each GPU's first SageAttention call is checked against SDPA and a failing GPU falls back for the session; no model family enters the scope yet. Adds the SageAttention docs page and the regenerated config schema.
With `attention_backend: sage`, FLUX.1 denoising, including ControlNets inside the loop, enters the SageAttention scope. RTX 4090, FLUX.1 dev FP8: 16.1 s → 14.0 s at 1024², 84.2 s → 54.8 s at 2048², peak VRAM unchanged; a blind review judged no image worse.
With `attention_backend: sage`, FLUX.2 denoising enters the SageAttention scope. RTX 4090, FLUX.2 Klein 9B FP8: 15.6 s → 15.1 s at 1024², 27.1 s → 23.1 s at 2048², peak VRAM unchanged; a blind review judged no image worse.
With `attention_backend: sage`, Z-Image denoising enters the SageAttention scope; Z-Image Control masks its attention and keeps SDPA. RTX 4090, Z-Image FP8: 10.4 s → 10.0 s at 1024², 33.7 s → 25.9 s at 2048², peak VRAM unchanged; a blind review judged no image worse.
With `attention_backend: sage`, Qwen-Image denoising enters the SageAttention scope. RTX 4090, Qwen-Image 2512 GGUF Q4_K_M: 78.8 s → 73.5 s at 1024², 208.9 s → 150.3 s at 2048², peak VRAM unchanged; a blind review judged no image worse.
With `attention_backend: sage`, Krea-2 denoising enters the SageAttention scope unless INVOKE_KREA2_SDPA_BACKEND pins a PyTorch kernel. RTX 4090, Krea-2 Turbo GGUF Q4_K_M: 11.5 s → 11.0 s at 1024², 42.7 s → 36.2 s at 2048², peak VRAM unchanged; a blind review judged no image worse.
This was referenced Oct 3, 2026
github-actions
Bot
added
python
…ent SageAttention failures sm86 now runs the Triton kernel upstream's sageattn picks there, and sm87 gets none; the Windows build's CUDA routing for both is unexplained and unmeasured here. Out-of-memory and Triton cache errors fall back to SDPA for that call (three retire the device), and the first-call check runs per precision and head size. The kernel runs with the tensors' device current, so a single-device install on cuda:1 uses SageAttention too.
An out-of-memory error now moves only the rest of that generation to SDPA, logged once, instead of counting toward retiring the device; out-of-memory is recognised by the shared is_oom_error. Only Triton cache failures in a row retire a device, and a success resets the count.
# Conflicts: # invokeai/frontend/api/openapi.json
# Conflicts: # invokeai/frontend/api/openapi.json # tests/test_config.py
This branch has not been deployed
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
3 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.
Note
Stacked PR 5 of 8: merge after #9661. It targets
main, so until the PRs before it merge, its diff also shows their commits and the merges that carry them forward. This PR's own change is one commit,482e7f348f.Summary
With
attention_backend: sage, Krea-2 denoising enters the SageAttention scope from #9657, so its eligible attention calls run on SageAttention; its grouped-query attention (48 query and 12 K/V heads) runs natively, without expanding K/V. WhileINVOKE_KREA2_SDPA_BACKENDpins a PyTorch kernel for measurement, the scope stays closed so the pinned kernel is what runs. The defaultautois unchanged. The docs page lists Krea-2 under Supported models.Related Issues / Discussions
Stacked on #9657 (SageAttention backend) and #9661, the PR before it.
QA Instructions
Krea-2 Turbo GGUF Q4_K_M, model loaded, arms as described in #9657:
To try it: install SageAttention as the docs page from #9657 describes, set
attention_backend: sage, restart and generate with Krea-2. The first generation logsSageAttention ... serves diffusion-model attention on cuda:0 (...)with the GPU and kernel name, and withlog_level: debugevery generation logs how many calls ran on SageAttention and why the others did not. Withattention_backend: autothe output is bit-identical tomain.Checks: ruff clean;
pytest tests/app/invocationspasses with the whole stack applied (2090 passed). No call-site test was added: this change only enters the scope, whose behavior #9657 tests.Review
Resolved: with
INVOKE_KREA2_SDPA_BACKENDset, SageAttention served the calls the override was meant to measure; the scope now stays closed while it is set.Compatibility / Rollout
Merge after #9661. No API or persisted-state change; only
attention_backend: sagebehaves differently.Checklist
What's Newcopy (if doing a release after this PR)