Cannot build mesh with error about `CLIPVisionModelWithProjection` and `torch`

[ArchieV1]

Hi,

I just downloaded Modly on Ubuntu by cloning the repo but when I try to build a mesh I get the error below.
I have tried both Hunyuan3D 2 Mini Turbo and Hunyuan3D 2 Mini Fast. The only configuration I have done is set the HuggingFace API key

I do not see any mention of CUDA 13 here which may be the issue:
// Derive max supported CUDA version from driver version
Is there a way for me to set which version of CUDA Modly should use? nvidia-smi that it calls returns 13.0 on my system.
I tried adding (driverMajor >= 580) cudaVersion = 130 but it didn't work so it may be more complex than that

The error seems to be about Torch based on AttributeError: module 'torch' has no attribute 'float8_e8m0fnu'.

Thanks :)

Specs:
AMD Ryzen 7 7800X3D 8-Core Processor
NVIDIA Corporation AD104 [GeForce RTX 4070 SUPER] (rev a1)

Ubuntu 24.04.4 LTS
Python 3.12.3
NPM 11.15.0
CUDA Version: 13.0

I do have a few CUDA versions installed

> ls -l /usr/local | grep cuda
lrwxrwxrwx  1 root root   22 Sep  4  2024 cuda -> /etc/alternatives/cuda
lrwxrwxrwx  1 root root   25 Sep  4  2024 cuda-12 -> /etc/alternatives/cuda-12
drwxr-xr-x 15 root root 4096 Mar 18  2025 cuda-12.8
lrwxrwxrwx  1 root root   25 Aug 12  2025 cuda-13 -> /etc/alternatives/cuda-13
drwxr-xr-x 14 root root 4096 Aug 12  2025 cuda-13.0

Error:

Error: Traceback (most recent call last):
  File "/home/archie/Programs/modly/api/routers/generation.py", line 153, in _run_generation
    gen = await loop.run_in_executor(None, generator_registry.get_active)
          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/lib/python3.12/concurrent/futures/thread.py", line 58, in run
    result = self.fn(*self.args, **self.kwargs)
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/archie/Programs/modly/api/services/generator_registry.py", line 236, in get_active
    gen.load()
  File "/home/archie/Programs/modly/api/services/extension_process.py", line 180, in load
    raise RuntimeError(msg.get("traceback") or msg.get("message"))
RuntimeError: Traceback (most recent call last):
  File "/home/archie/Documents/Modly/extensions/hunyuan3d-mini-turbo/venv/lib/python3.12/site-packages/transformers/utils/import_utils.py", line 2254, in __getattr__
    module = self._get_module(self._class_to_module[name])
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/archie/Documents/Modly/extensions/hunyuan3d-mini-turbo/venv/lib/python3.12/site-packages/transformers/utils/import_utils.py", line 2488, in _get_module
    raise e
  File "/home/archie/Documents/Modly/extensions/hunyuan3d-mini-turbo/venv/lib/python3.12/site-packages/transformers/utils/import_utils.py", line 2486, in _get_module
    return importlib.import_module("." + module_name, self.__name__)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/lib/python3.12/importlib/__init__.py", line 90, in import_module
    return _bootstrap._gcd_import(name[level:], package, level)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "<frozen importlib._bootstrap>", line 1387, in _gcd_import
  File "<frozen importlib._bootstrap>", line 1360, in _find_and_load
  File "<frozen importlib._bootstrap>", line 1331, in _find_and_load_unlocked
  File "<frozen importlib._bootstrap>", line 935, in _load_unlocked
  File "<frozen importlib._bootstrap_external>", line 995, in exec_module
  File "<frozen importlib._bootstrap>", line 488, in _call_with_frames_removed
  File "/home/archie/Documents/Modly/extensions/hunyuan3d-mini-turbo/venv/lib/python3.12/site-packages/transformers/models/clip/modeling_clip.py", line 26, in <module>
    from ...modeling_layers import GradientCheckpointingLayer
  File "/home/archie/Documents/Modly/extensions/hunyuan3d-mini-turbo/venv/lib/python3.12/site-packages/transformers/modeling_layers.py", line 27, in <module>
    from .processing_utils import Unpack
  File "/home/archie/Documents/Modly/extensions/hunyuan3d-mini-turbo/venv/lib/python3.12/site-packages/transformers/processing_utils.py", line 83, in <module>
    from .modeling_utils import PreTrainedAudioTokenizerBase
  File "/home/archie/Documents/Modly/extensions/hunyuan3d-mini-turbo/venv/lib/python3.12/site-packages/transformers/modeling_utils.py", line 69, in <module>
    from .integrations.finegrained_fp8 import ALL_FP8_EXPERTS_FUNCTIONS
  File "/home/archie/Documents/Modly/extensions/hunyuan3d-mini-turbo/venv/lib/python3.12/site-packages/transformers/integrations/finegrained_fp8.py", line 45, in <module>
    _UE8M0_SF_DTYPE = torch.float8_e8m0fnu
                      ^^^^^^^^^^^^^^^^^^^^
  File "/home/archie/Documents/Modly/extensions/hunyuan3d-mini-turbo/venv/lib/python3.12/site-packages/torch/__init__.py", line 2681, in __getattr__
    raise AttributeError(f"module '{__name__}' has no attribute '{name}'")
AttributeError: module 'torch' has no attribute 'float8_e8m0fnu'

The above exception was the direct cause of the following exception:

Traceback (most recent call last):
  File "/home/archie/Documents/Modly/extensions/hunyuan3d-mini-turbo/generator.py", line 296, in _ensure_hy3dgen
    from hy3dgen.shapegen import Hunyuan3DDiTFlowMatchingPipeline  # noqa: F401
    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/archie/Documents/Modly/models/hunyuan3d-mini-turbo/generate/_hy3dgen/hy3dgen/shapegen/__init__.py", line 15, in <module>
    from .pipelines import Hunyuan3DDiTPipeline, Hunyuan3DDiTFlowMatchingPipeline
  File "/home/archie/Documents/Modly/models/hunyuan3d-mini-turbo/generate/_hy3dgen/hy3dgen/shapegen/pipelines.py", line 30, in <module>
    from .models.autoencoders import ShapeVAE
  File "/home/archie/Documents/Modly/models/hunyuan3d-mini-turbo/generate/_hy3dgen/hy3dgen/shapegen/models/__init__.py", line 27, in <module>
    from .conditioner import DualImageEncoder, SingleImageEncoder, DinoImageEncoder, CLIPImageEncoder
  File "/home/archie/Documents/Modly/models/hunyuan3d-mini-turbo/generate/_hy3dgen/hy3dgen/shapegen/models/conditioner.py", line 29, in <module>
    from transformers import (
  File "/home/archie/Documents/Modly/extensions/hunyuan3d-mini-turbo/venv/lib/python3.12/site-packages/transformers/utils/import_utils.py", line 2342, in __getattr__
    raise ModuleNotFoundError(
ModuleNotFoundError: Could not import module 'CLIPVisionModelWithProjection'. Are this object's requirements defined correctly?

The above exception was the direct cause of the following exception:

Traceback (most recent call last):
  File "/home/archie/Programs/modly/api/runner.py", line 143, in main
    gen.load()
  File "/home/archie/Documents/Modly/extensions/hunyuan3d-mini-turbo/generator.py", line 49, in load
    self._ensure_hy3dgen()
  File "/home/archie/Documents/Modly/extensions/hunyuan3d-mini-turbo/generator.py", line 298, in _ensure_hy3dgen
    raise RuntimeError(
RuntimeError: hy3dgen still not importable after extraction to /home/archie/Documents/Modly/models/hunyuan3d-mini-turbo/generate/_hy3dgen.
Check the folder contents.
Could not import module 'CLIPVisionModelWithProjection'. Are this object's requirements defined correctly?

PIP dependency tree inside the VENV:

cryptography==48.0.0
└── cffi [required: >=2.0.0, installed: 2.0.0]
    └── pycparser [required: Any, installed: 3.0]
fastapi==0.115.6
├── pydantic [required: >=1.7.4,<3.0.0,!=2.1.0,!=2.0.1,!=2.0.0,!=1.8.1,!=1.8, installed: 2.13.4]
│   ├── annotated-types [required: >=0.6.0, installed: 0.7.0]
│   ├── pydantic_core [required: ==2.46.4, installed: 2.46.4]
│   │   └── typing_extensions [required: >=4.14.1, installed: 4.15.0]
│   ├── typing_extensions [required: >=4.14.1, installed: 4.15.0]
│   └── typing-inspection [required: >=0.4.2, installed: 0.4.2]
│       └── typing_extensions [required: >=4.12.0, installed: 4.15.0]
├── starlette [required: >=0.40.0,<0.42.0, installed: 0.41.3]
│   └── anyio [required: >=3.4.0,<5, installed: 4.13.0]
│       ├── idna [required: >=2.8, installed: 3.18]
│       └── typing_extensions [required: >=4.5, installed: 4.15.0]
└── typing_extensions [required: >=4.8.0, installed: 4.15.0]
httptools==0.8.0
huggingface_hub==1.18.0
├── click [required: >=8.4.0, installed: 8.4.1]
├── filelock [required: >=3.10.0, installed: 3.29.1]
├── fsspec [required: >=2023.5.0, installed: 2026.4.0]
├── hf-xet [required: >=1.4.3,<2.0.0, installed: 1.5.0]
├── httpx [required: >=0.23.0,<1, installed: 0.28.1]
│   ├── anyio [required: Any, installed: 4.13.0]
│   │   ├── idna [required: >=2.8, installed: 3.18]
│   │   └── typing_extensions [required: >=4.5, installed: 4.15.0]
│   ├── certifi [required: Any, installed: 2026.5.20]
│   ├── httpcore [required: ==1.*, installed: 1.0.9]
│   │   ├── certifi [required: Any, installed: 2026.5.20]
│   │   └── h11 [required: >=0.16, installed: 0.16.0]
│   └── idna [required: Any, installed: 3.18]
├── packaging [required: >=20.9, installed: 26.2]
├── PyYAML [required: >=5.1, installed: 6.0.3]
├── tqdm [required: >=4.42.1, installed: 4.68.1]
├── typer [required: >=0.20.0,<0.26.0, installed: 0.25.1]
│   ├── annotated-doc [required: >=0.0.2, installed: 0.0.4]
│   ├── click [required: >=8.2.1, installed: 8.4.1]
│   ├── rich [required: >=13.8.0, installed: 15.0.0]
│   │   ├── markdown-it-py [required: >=2.2.0, installed: 4.2.0]
│   │   │   └── mdurl [required: ~=0.1, installed: 0.1.2]
│   │   └── Pygments [required: >=2.13.0,<3.0.0, installed: 2.20.0]
│   └── shellingham [required: >=1.3.0, installed: 1.5.4]
└── typing_extensions [required: >=4.1.0, installed: 4.15.0]
pip==24.0
pipdeptree==3.1.0
└── packaging [required: >=26, installed: 26.2]
pymeshlab==2025.7.post1
└── numpy [required: Any, installed: 2.4.6]
python-dotenv==1.2.2
python-multipart==0.0.20
trimesh==4.12.2
└── numpy [required: >=1.20, installed: 2.4.6]
uvicorn==0.34.0
├── click [required: >=7.0, installed: 8.4.1]
└── h11 [required: >=0.8, installed: 0.16.0]
uvloop==0.22.1
watchfiles==1.2.0
└── anyio [required: >=3.0.0, installed: 4.13.0]
    ├── idna [required: >=2.8, installed: 3.18]
    └── typing_extensions [required: >=4.5, installed: 4.15.0]
websockets==16.0