This is a maintenance release of InvokeAI focused on bug fixes and stability. Version 6.14.0 will be the next major feature release, featuring video generation, multiple GPU support, the Wan 2.2 image and video models, pressure-sensitive canvas, HiDiffusion integration, workflow to workflow calls, and much more.
Here are the highlights of the current release:
Updated Libraries
- React version 19 (frontend support) @joshistoast
- ROCm version 7.1 (AMD GPU support) @lstein @heathen711
- Transformers version 5.5.4 (AI inference and training support) @kappacommit
Major Bug Fixes
- Fix the image viewer progress preview behavior when selecting a gallery image during image generation. @JPPhoto
- Save and restore both positive and negative prompt history for SD images. @DustyShoe
- Fix crash when a LoRA was applied to a Flux.2 4B/9B Diffusers model loaded with FP8 storage. @Pfannkuchensack
- Restore the X/Y queue badges when running in multiuser mode. @lstein
- Prevent main panel from being squeezed out of existence on small screens. @lstein
- Gallery maintenance script updated to correctly handle nested image subfolders when structured image directory activated. @lstein
- Fix crash when using a Z-Image model with regional guidance. @Pfannkuchensack
- Prevent crash when Heun (2d order) sampler applied to a FLUX.2 Klein model with reference images. @Pfannkuchensack
- Prevent hang when an unknown wildcard was passed to the dynamic prompt processor. @Pfannkuchensack
New Features
- Support PEFT named-adapter LoRAs. @Pfannkuchensack
- Add Qwen Image single-file checkpoint loader with fp8 support. @Pfannkuchensack
Documentation
- Added information on 3d party GPU providers that host InvokeAI. @lstein @michaelmanly
- Fix uvicorn URL. @fishd72
- Add graph execution example to API guide. @wunianze666-netizen
Full List of Changes
- Prep PR for v6.13.0 by @lstein in #9134
- chore(version): bump to 6.13.0.post1 by @lstein in #9242
- docs(fp8): update FP8 Storage docs by @Pfannkuchensack in #9241
- fix(ui): require metadata handler success for some mode by @YUHAO-corn in #9197
- chore(github): update actions from deprecated node-20 versions by @keturn in #9149
- fix(ci/cd): fix typegen test failure introduced by PR 9209 by @lstein in #9260
- fix(openapi): normalize Path defaults to POSIX in schema generation by @Pfannkuchensack in #9209
- chore(ci): update cached-lfs-checkout from deprecated node 20 version by @keturn in #9261
- feat(model manager): external providers UI/UX tweaks by @joshistoast in #9237
- Support Hugging Face blob file URLs by @C2015 in #9167
- docs: Realign External Models by @Pfannkuchensack in #9222
- Fix(model install): wait for startup restore before imports by @JPPhoto in #9239
- feat(anima): bundle T5-XXL tokenizer, remove t5_encoder_model input by @kappacommit in #9244
- feat(lora): support PEFT named-adapter LoRAs (e.g. Klein 9B) by @Pfannkuchensack in #9252
- Fix progress preview gallery selection by @JPPhoto in #9217
- Fix: add SD negative prompt graph wiring and history by @DustyShoe in #9243
- fix(ci): repair wrapped ENABLE_ANALYTICS expression in deploy-docs workflow by @lstein in #9271
- Fix workflow execution state reconciliation by @JPPhoto in #9199
- ci: use uv.lock when running tests by @keturn in #9214
- ci: run pytest via uv run to fix venv PATH breakage from #9214 by @lstein in #9286
- fix(lora): sidecar-patch fp8 weights to avoid float8 add crash by @lstein in #9246
- docs: Updated uvicorn URL in documentation by @fishd72 in #9230
- docs: add graph execution example to API guide by @wunianze666-netizen in #9276
- feat: ⬆️ Upgrade to React 19 by @joshistoast in #8708
- fix(multiuser): restore X/Y queue badge and redacted queue entries for non-admins by @lstein in #9262
- Fix/z image controlnet logging (supersedes #9289) by @lstein in #9298
- Update ROCm to 7.1 (supersedes #8867) by @lstein in #9297
- fix(db): repair model_relationships foreign keys broken by migration 22 by @Pfannkuchensack in #9272
- Fix python tests on CUDA by @JPPhoto in #9215
- feat(mm): add Qwen Image single-file checkpoint loader with fp8 support by @Pfannkuchensack in #9253
- fix(flux2): strip reference-image tokens from Heun preview latents by @Pfannkuchensack in #9274
- fix(ui): keep main panel reachable on small screens by @lstein in #9144
- fix(ui): add missing canvas-workflow-integration log namespace translation by @Pfannkuchensack in #9275
- fix(gallery): traverse image subfolders when finding orphaned db entries by @lstein in #9291
- feat(api): append mode for recalled reference images (+ fix recall socket double-emit) by @lstein in #9288
- fix: honor empty-string metadata overrides (supersedes 9283) by @lstein in #9309
- Fix: z-image regional guidance split mismatch by @Pfannkuchensack in #9273
- docs: add GitHub Sponsors button and README sponsors section by @lstein in #9312
- chore: Remove misplaced claude .md file by @lstein in #9315
- fix(dynamicprompts): prevent hang on unknown wildcards by @Pfannkuchensack in #9307
- feat - Migrate to Transformers 5.5.4 by @kappacommit in #9248
New Contributors
- @C2015 made their first contribution in #9167
- @fishd72 made their first contribution in #9230
- @wunianze666-netizen made their first contribution in #9276
Full Changelog: v6.13.0...v6.13.5.rc1