Oct 22 5 days ago

1. 12:27

   Size GPU durations to measured times; cap frames at 65; better video example Running

   multimodalart HF Staff
   Verified signature

   0e2afa4
2. 12:17

   Fuse LoRA into the transformer

   multimodalart HF Staff
   Verified signature

   ee85eed
3. 12:08

   Timing logs; drop reference example

   multimodalart HF Staff
   Verified signature

   cc014f6
4. 11:56

   Disable VAE tile parallelism (single process)

   multimodalart HF Staff
   Verified signature

   cc8fb53
5. 11:51

   Upload app.py with huggingface_hub

   multimodalart HF Staff
   Verified signature

   d5ed3f4
6. 11:41

   Stub single-process parallel state (no DeviceMesh on ZeroGPU)

   multimodalart HF Staff
   Verified signature

   6d84135
7. 11:36

   Initial VitronEdit demo

   multimodalart HF Staff
   Verified signature

   886038b
8. 10:57

   Shim transformers.HybridCache = StaticCache (removed in transformers 5.x, imported by peft at module scope); revert to transformers 5.9.0 (4.x needs hub<1.0, incompatible with gradio 6.30)

   multimodalart HF Staff
   Verified signature

   a258166
9. 08:21

   Upload folder using huggingface_hub

   multimodalart HF Staff
   Verified signature

   9cbfb73
10. 07:50

    Upload folder using huggingface_hub

    multimodalart HF Staff
    Verified signature

    0617cf0