Skip to content

Pull requests: ostris/ai-toolkit

New pull request New
41 Open 212 Closed
41 Open 212 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Reviews
Filter by reviews
No reviews Review required Approved review Changes requested
Assignee
Filter by who’s assigned
Assigned to nobody Loading
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

Auto-caption QoL change, build errors cleared.
#876 opened Jun 10, 2026 by ElBiggus Loading…
Fix progress bar iteration speed stability
#873 opened Jun 9, 2026 by PlagueKind Contributor Loading…
Add block-level compile and qcompile torch.compile whole model modes
#866 opened Jun 7, 2026 by PlagueKind Contributor Loading…
Add a PixlStash Remote plugin
#864 opened Jun 5, 2026 by lindkvis Loading…
Add support for remote dataset source plugins
#863 opened Jun 5, 2026 by lindkvis Loading…
feat: Add Anima support
#860 opened Jun 2, 2026 by rmatif Loading…
1
2
Use TextArea for job name entry
#856 opened May 31, 2026 by fatalis Loading…
Fixed macstats loading on Mac OS in the UI
#851 opened May 29, 2026 by kinnrai Loading…
1
2
feat: Stop with checkpoint save
#850 opened May 29, 2026 by entmike Loading…
fix: restart queue when a job is queued
#846 opened May 26, 2026 by SidneyAnderson Loading…
Flux2: support local safetensors paths and stream-load to fix RAM blowup
#842 opened May 22, 2026 by deepglugs Loading…
feat(ui): persist JobLossGraph settings per job + fix chart height on control toggle
#841 opened May 22, 2026 by MarcBate Loading…
5 tasks
fix: warm latent cache loading does an unnecessary model swap
#840 opened May 22, 2026 by Mandrakia Loading…
fix(flux2-klein): support community Qwen3 text encoders with missing chat_template
#838 opened May 20, 2026 by smhc Loading…
fix(flux2-klein-base): clamp transformer residual stream to bf16-safe range
#837 opened May 20, 2026 by fengcunhan Loading…
1
Update dgx_instructions.md
#830 opened May 12, 2026 by NeoTech Loading…
1
Add dual-GPU model-parallel paths for 8 LoRA-training architectures (FLUX.2 / Wan 2.2 / LTX-2 / Qwen-Image / Chroma / FLUX.1-Kontext / HiDream-I1 / Z-Image)
#829 opened May 11, 2026 by genno-whittlery Loading…
1
8
Fix Ace Step 1.5 Batching error
#825 opened May 9, 2026 by AdamNizol Loading…
Enable VAE tiling for Wan to avoid OOM at sample-time
#819 opened May 7, 2026 by jeremy-linchao Loading…
fix: ACE Step Training float32 saving and shape
#814 opened May 2, 2026 by DefinitlyEvil Loading…
perf: enable pin_memory on CUDA DataLoaders for faster CPU->GPU transfer
#813 opened May 1, 2026 by Booyaka101 Loading…
2
Add Florence-2 captioner to Auto Caption
#812 opened Apr 30, 2026 by Zorgsd Loading…
fix(ui): generate CSS for grid-cols-13 through grid-cols-40
#811 opened Apr 29, 2026 by kyleinprogress Loading…
RLHF Flow-GRPO implementation POC
#808 opened Apr 28, 2026 by ifilipis Loading…
8 of 10 tasks
4
fix: handle missing Flux VAE config values and None prompt inputs
#807 opened Apr 28, 2026 by octaviomejiadiaz Loading…
ProTip! Exclude everything labeled bug with -label:bug.