hzxie / citydreamer Goto Github PK

Hello, may I ask how did you process the OSM dataset into semantic maps and height fields? Did you use any open source Python API?

Hi,

thanks a lot for sharing this amazing project.
Following the README file, I successfully installed extrude_tensor and voxrender, but unfortunately the wheel for grid_encoder fails to be built all the time. Could you please provide me some solutions?
My environment is: Ubuntu 24.04, CUDA 12.2, Python 3.8.19, PyTorch 2.3.1.

Best,
Cecilia

Hi. Will the datasets be released? Where/when?
Thanks

FileNotFoundError: [Errno 2] No such file or directory: '/home/city-dreamer/output/sampler.pth'

When can I get this file?

      arch_list[-1] += '+PTX'
  IndexError: list index out of range
  [end of output]

note: This error originates from a subprocess, and is likely not a problem with pip.
ERROR: Failed building wheel for extrude_tensor
Running setup.py clean for extrude_tensor
Failed to build extrude_tensor
ERROR: Could not build wheels for extrude_tensor, which is required to install pyproject.toml-based projects

error when your install bash

cd $CITY_DREAMER_HOME/extensions
for e in `ls -d */`
do
  cd $CITY_DREAMER_HOME/extensions/$e
  pip install .
done

It seems like the pre-trained models and datasets cannot be downloaded. 404 not found

Hi, thanks for releasing this awesome project! Are there any plans to switch to an open source license?
Thanks!

Apologies for the too many questions below, but this is perfect for the project I am working on (and I feel our project will perfectly demonstrate the potentials of your work here):

Is it possible to export this to web 3d engines like BabylonJS? And if yes, will I be able to split one 3d map to many zones (for loading/unloading optimization) and will it be easy to apply collision on it?

Hi, thanks for releasing this awesome project! Are there any plans to open source train code ?
Thanks!

Hi, happen to see your wonderful work! Wondering is there an approximate release time of training code?

Thank you for your open source!

There's a problem here:
Weight download link not available

Great work! Wondering when you plan to release your demo on hugging face and the code to play with.

Thank you for your work! Regarding the section 4.1. of the paper The OSM Dataset, I was wondering if the height fields are originally present in the OpenStreetMap dataset? But I didn't find it, and I additionally used tools like osmnx and arcmap.

torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 5.62 GiB (GPU 0; 3.63 GiB total capacity; 904.46 MiB already allocated; 1.60 GiB free; 930.00 MiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

The output is video, can export the building structure,mesh obj or other formats...

RuntimeError: No CUDA GPUs are available

[INFO] 2024-04-02 14:55:42,701 192.168.31.42 - - [02/Apr/2024 14:55:42] "GET /image/e5fb56dc-5324-4875-b320-c69c81d52e0c.png/normalize.action HTTP/1.1" 200 -
[INFO] 2024-04-02 14:55:42,707 192.168.31.42 - - [02/Apr/2024 14:55:42] "GET /image/5dfef487-02fd-4725-a58e-1a4bae428e6c.png HTTP/1.1" 200 -
Rendering seg volume:   0%|                                                                                                                                                                 | 0/20 [00:00<?, ?it/s]
[INFO] 2024-04-02 14:55:44,741 192.168.31.42 - - [02/Apr/2024 14:55:44] "POST /trajectory/preview.action HTTP/1.1" 500 -
Traceback (most recent call last):
  File "/home/dastan/.virtualenvs/citydreamer/lib/python3.10/site-packages/flask/app.py", line 1488, in __call__
    return self.wsgi_app(environ, start_response)
  File "/home/dastan/.virtualenvs/citydreamer/lib/python3.10/site-packages/flask/app.py", line 1466, in wsgi_app
    response = self.handle_exception(e)
  File "/home/dastan/.virtualenvs/citydreamer/lib/python3.10/site-packages/flask/app.py", line 1463, in wsgi_app
    response = self.full_dispatch_request()
  File "/home/dastan/.virtualenvs/citydreamer/lib/python3.10/site-packages/flask/app.py", line 872, in full_dispatch_request
    rv = self.handle_user_exception(e)
  File "/home/dastan/.virtualenvs/citydreamer/lib/python3.10/site-packages/flask/app.py", line 870, in full_dispatch_request
    rv = self.dispatch_request()
  File "/home/dastan/.virtualenvs/citydreamer/lib/python3.10/site-packages/flask/app.py", line 855, in dispatch_request
    return self.ensure_sync(self.view_functions[rule.endpoint])(**view_args)  # type: ignore[no-any-return]
  File "/home/dastan/kas/projects/city-dreamer/demo/run.py", line 164, in get_trajectory_preview
    frames = get_seg_volume_rendering(hf_filepath, seg_filepath, trajectory)
  File "/home/dastan/kas/projects/city-dreamer/demo/run.py", line 204, in get_seg_volume_rendering
    seg_volume = inference.get_seg_volume(part_hf, part_seg)
  File "/home/dastan/kas/projects/city-dreamer/demo/../scripts/inference.py", line 376, in get_seg_volume
    return scripts.dataset_generator.get_seg_volume(part_seg, part_hf)
  File "/home/dastan/kas/projects/city-dreamer/demo/../scripts/dataset_generator.py", line 84, in get_seg_volume
    torch.from_numpy(part_seg_map[None, None, ...]).cuda(),
  File "/home/dastan/.virtualenvs/citydreamer/lib/python3.10/site-packages/torch/cuda/__init__.py", line 229, in _lazy_init
    torch._C._cuda_init()
RuntimeError: No CUDA GPUs are available

Dear Authors,

How long does your method need to train?

Thanks

Hi, I downloaded the OSM dataset in your shared link, but I found that the hf.png is different from Figure.3 (a).

demo.py

inference.py

Hi,
Is it possible to import a city into a gam engine (ie Unity or Unreal)?