Spaces:
Running on Zero
Running on Zero
Commit History
Fix turn_into_video: normalize Wan return shape + serve video locally f074c20
Mount /outputs statically so generated PNGs are served (404 fix) 0d0014f
Serve outputs/ (not tempfiles) + harden frontend URL extraction & logging b27b052
Fix endpoint output decl + url=None: add return type hints, serve via temp FileData 17bf644
Fix transformer pos_embed crash and frontend submit().on() error e6b651d
Add type hint to progress params for gradio.Server api() validator 291a410
Temporarily fall back to base Qwen-Image-Edit-2511 (v18 transformer repo 401) 068e315
Trim UI to original feature set (remove zoom, multi-gen, dice buttons, tabs) 702b146
Bump gradio to 6.20.0 in README and requirements a0143a9
Switch to gradio.Server with Photoshop-style studio UI 6f0b1e3
Pin kernels<0.15.1 to dodge required revision/version arg break ca7f544
moose commited on
Remove temporary dimension-flow debug prints ad7b834
moose Claude Opus 4.7 (1M context) commited on
Derive output aspect ratio from first image, not last 09cc589
moose Claude Opus 4.7 (1M context) commited on
Add temporary debug prints for dimension flow (will revert) 7a46cb6
moose commited on
Revert "Preserve input image aspect ratio when sliders are at default" 593798f
moose commited on
Preserve input image aspect ratio when sliders are at default efe093f
moose Claude Opus 4.7 (1M context) commited on
Expose a second image input slot in the UI b34fbef
moose Claude Opus 4.7 (1M context) commited on
Add debug diagnostic dashboard 486537f
moose Claude Opus 4.7 (1M context) commited on
Lower @spaces.GPU duration from 300s to 60s 7654ca1
moose Claude Opus 4.7 (1M context) commited on
Replace external vllm-flash-attn3 kernel with PyTorch cuDNN FA3 backend 1543d6f
moose Claude Opus 4.7 (1M context) commited on
Disable AOTI compile to dodge spaces==0.49.3 NVML regression 3e914c4
moose Claude Opus 4.7 (1M context) commited on
Pin spaces==0.50.0 to dodge today's regression e64cecc
moose Claude Opus 4.7 (1M context) commited on
Bisect AOTI compile duration upward: 600 → 1000 09d4474
moose Claude Opus 4.7 (1M context) commited on
Lower AOTI compile duration 1200 → 600 (cap probe) 9ce1f2c
moose Claude Opus 4.7 (1M context) commited on
Lower AOTI compile duration 1500 → 1200 (cap probe) 64a4e11
moose Claude Opus 4.7 (1M context) commited on
Clean up for public release fcd1c32
moose Claude Opus 4.6 commited on
Add .claude/ to gitignore 076f484
moose Claude Opus 4.6 commited on
Add backup file patterns to gitignore 973d4b0
moose Claude commited on
Fix device mismatch during AOT compilation warmup bca4977
moose Claude Opus 4.5 commited on
Upgrade to Qwen-Image-Edit-2511 with Phr00t v18 transformer 07e1710
moose Claude Opus 4.5 commited on
Reduce memory during AOT compilation by offloading text encoder bbac3b5
moose Claude Opus 4.5 commited on
Revert "Fix OOM by deferring model loading and AOT compilation to runtime" 73a98d6
moose commited on
Fix model link in UI description 6c84590
moose Claude Opus 4.5 commited on
Fix OOM by deferring model loading and AOT compilation to runtime 49afc98
moose Claude Opus 4.5 commited on
Add *.log to gitignore 25b29ac
moose Claude Opus 4.5 commited on
routine commit b333e18
moose commited on
Agent guidelines c47a9b9
moose commited on
Add custom folder to gitignore for local development 70b9077
moose Claude commited on
Remove demo files and fix image serving for proper download support 2525dc6
moose Claude commited on