Commit History
4 days ago
-
Update requirements.txt
Verified signaturefaa438c -
Update requirements.txt
Verified signature2f7a5a7 -
Update requirements.txt
Verified signature52651a3 -
Update requirements.txt
Verified signature2df70d6
-
use xlarge ZeroGPU (full Blackwell) for VAE-decode headroom
Adam Oster +1: Claude Opus 4.8 (1M context)Crash is in VAE-decode conv3d only, after the transformer completes, on every torch version. On the default 'large' size (half a Blackwell, ~48GB) the large full-res decode allocation fails and PyTorch's OOM path dies in the forked process's NVML query, surfacing as the nvmlInit assert. Built-in VAE tiling drops the decoder's timestep conditioning so it's unsafe here; request the full 96GB Blackwell instead (2x quota cost). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>8a42e93 -
pin torch==2.8.0 to match working LTX-2-3 ZeroGPU Space
Adam Oster +1: Claude Opus 4.8 (1M context)Lightricks/LTX-2-3 runs fine on the same ZeroGPU Blackwell hardware with torch==2.8.0. torch 2.9.1, 2.11.0, and ZeroGPU's default all hit the same nvmlInit assert in VAE decode here, but 2.8.0 was never tried. Match the known-good sibling Space. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>64b2479 -
remove torch pin; use ZeroGPU-provided torch
Adam Oster +1: Claude Opus 4.8 (1M context)The NVML_SUCCESS / nvmlInit assert in CUDACachingAllocator fires on the first heavy VAE-decode allocation at ALL resolutions (incl. 360x640), so it is NVML init failing in the ZeroGPU forked GPU process, not OOM and not app logic. Both pinned PyPI torch builds (2.9.1, 2.11.0) crash identically. ZeroGPU ships its own patched torch and discourages pinning, so drop the pin and let the platform supply a torch matched to the Blackwell runtime. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>d30bc58 -
bump torch to 2.11.0 for ZeroGPU Blackwell NVML fix
Adam Oster +1: Claude Opus 4.8 (1M context)The NVML_SUCCESS assert in CUDACachingAllocator was a torch/driver mismatch after ZeroGPU moved to RTX Pro 6000 Blackwell GPUs. torch 2.11.0 ships CUDA libraries matching the Blackwell driver. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>a1ccfc0
-
pin torch version to fix CUBLAS error (#34)
- pin torch version to fix CUBLAS error (7d0ff05e62d749c93a9032628f7bc5e034f7e644) Co-authored-by: Linoy Tsaban <linoyts@users.noreply.huggingface.co>072c816
-
update gradio version (#29)
Verified signature- update gradio version (d12b1c430db24f734c5280c1339929c4a6e32c35) Co-authored-by: Linoy Tsaban <linoyts@users.noreply.huggingface.co>6dd10bd
-
add note about the update to v0.9.8 (#24)
Verified signature- add note about the update to v0.9.8 (f256d9380013e0ecbac7f17d2e91f018b1f8df5f) Co-authored-by: Linoy Tsaban <linoyts@users.noreply.huggingface.co>12a74c8
-
Update README.md
Verified signature9c24b47 -
update to 0.9.8 (#23)
Verified signature- update to 0.9.8 (0135c48cef03cb1cbdefffc47d83a85095518e10) - Update app.py (e1c8d1db758dbb5663f2b31af5be19c7aaf5b5c3) - update code base with main (27e0be50ac121b4b63239a5b5dc6a11eddac9209) - Update app.py (c3db712c6a92b978a2feb71ccfd70038b6ee3108) - Update app.py (ab760cfbc5339b4551b6e5a02020981b4c8affe3) Co-authored-by: Linoy Tsaban <linoyts@users.noreply.huggingface.co>68008ac
-
mcp-improve (#22)
Verified signature- docstring (34da4adf433380c683a14b9ce895f39611ca456f) - docstring (e6fdb5dbc9a7972f90cc20d5f4149c3af60d64f9) - Fix MCP client parameter validation for LTX video generation (08451a637e822ca3685f05fefadceee64075969b) Co-authored-by: Victor Mustar <victor@users.noreply.huggingface.co>77756bc
-
Update app.py
Verified signaturee1afb2d
-
Update README.md
Verified signature9ac909b -
disable deeplink temporarily
Verified signaturea1fc814
-
Update requirements.txt
Verified signaturee81457f -
update-inference (#4)
Verified signature- Update inference.py (7b9f053b980d37b4af43b808e6cfada81caee869) - Update app.py (9742f923e6f1247a0c029af3dce3613733f5530c)190cbef -
Update app.py
Verified signature7d1232d -
oops
Verified signaturef62fc75 -
get rid of gr.State
Verified signature6243da9
-
Update app.py
Verified signature71c64ba -
Randomize seeds sby default
Verified signature826eb28
-
Fix `improve_texture` and hide video-to-video
Verified signatureb4e4e06 -
Update app.py
Verified signature05ab75c -
Update app.py
Verified signaturedf03e5b -
Add DeepLinkButton (#2)
Verified signature- Update app.py (9c078325886ca74e05d7b3afc868d8095cd25438) Co-authored-by: Freddy Boulton <freddyaboulton@users.noreply.huggingface.co>045eb77 -
MCP support? (#1)
Verified signature- MCP support? (44513f731f3dd4174d73f2f39f023b20da05e20b) Co-authored-by: Victor Mustar <victor@users.noreply.huggingface.co>654e0c3 -
Update README.md
Verified signature230e1d9 -
Update configs/ltxv-13b-0.9.7-distilled.yaml
Verified signature21e81c3 -
Update app.py
Verified signaturec4fd703 -
Update app.py
Verified signaturea930ff9 -
Update app.py
Verified signature4d68dfd -
Update app.py
Verified signature6ad4062 -
Update app.py
Verified signature3d79b08 -
Update app.py
Verified signaturecbdec18 -
Update app.py
Verified signatureb972f40 -
Update app.py
Verified signaturede0b990 -
Update app.py
Verified signatureaebf56b -
Update app.py
Verified signature626b672 -
Update app.py
Verified signature70c8ddd -
Update app.py
Verified signatured8bb216 -
Update app.py
Verified signature3947f33 -
Update app.py
Verified signature2a3fbc3 -
Update app.py
Verified signaturec942f44 -
Update app.py
Verified signature17b56a5 -
Update app.py
Verified signaturedff15f5 -
Update configs/ltxv-13b-0.9.7-distilled.yaml
Verified signature9235701 -
Update app.py
Verified signature65d68c9
