Add packages.txt: libopengl0/libegl1 so pymeshlab's libfilter_*.so plugins load (meshing_decimation_quadric_edge_collapse missing)

Bump GPU duration to 600s; force CPU onnxruntime (CUDA13 vs onnxruntime-gpu cu12 mismatch)

Swap gated benjamin-paine SD1.5 for the public stable-diffusion-v1-5/stable-diffusion-v1-5 mirror

Patch pytorch3d/renderer/__init__.py too (drop dangling pulsar import); force rebuild if previous install left broken renderer init

Retrigger build (probe OOM)

Strip pulsar from pytorch3d build (modern GCC link failure)

Unpin torch; use ZeroGPU default (2.11) and matching torch_scatter wheel

Pin torch to cu130 build to match ZeroGPU CUDA 13.0

Bump torch to 2.10.0 (cu130) to match ZeroGPU CUDA 13

Bump gradio sdk_version to 5.49.1 to fix HfFolder import

Support new ZeroGPU hardware: unpin torch, drop ABI-locked wheels, build pytorch3d/nvdiffrast from source

Update app.py

Update app.py

Update gradio_app/all_models.py

Update scripts/sd_model_zoo.py

Update gradio_app/gradio_3dgen.py

Update gradio_app/gradio_3dgen.py

Update app.py

Display the model from the front instead of the back (#5)

bug fix

bug fix, add examples

bug fix

Better ZeroGPU utilization

Update app.py

update readme

set time limit

bug fix

x

Upload nvdiffrast-0.3.1.torch-cp310-cp310-linux_x86_64.whl

bug fix

update requirements

add package to local

x

x

x

avoid memory leak

x

switch gl to cuda

upload prebuilt version

x

x

add prebuilt files

x

x

bug fix

bug fix

bug fix

fix bug

x

x