Maolin Li
Maolin Li
railway版本也缺少google.generativeai模块，如何手动在railway安装？ PS:已经尝试修改了requirements.txt 添加了google.generativeai,仍旧出现相同的错误。
Hi, does the latest version of nerfacc support torch 1.10?
I changed another conda env. env info:' cuda 11.3, Torch 1.11' I encounter some errors:'Traceback (most recent call last): File "/gpfs/home/mli/banmo/scripts/visualize/nvs.py", line 201, in app.run(main) File "/home/mli/.conda/envs/torch-110/lib/python3.9/site-packages/absl/app.py", line 308, in...
traverse_grids inputs: {'rays_o': tensor([[-0.2437, -0.0069, 0.0255], [-0.2437, -0.0069, 0.0255], [-0.2437, -0.0069, 0.0255], ..., [-0.2437, -0.0069, 0.0255], [-0.2437, -0.0069, 0.0255], [-0.2437, -0.0069, 0.0255]], device='cuda:0'), 'rays_d': tensor([[ 0.9839, 0.1560, -0.2869], [ 0.9840,...
Thank you for checking! [inputs.pth.zip](https://github.com/nerfstudio-project/nerfacc/files/14423244/inputs.pth.zip)
I have checked the NaN values, shape, and type of input arguments, and there are no existing issues. Can you provide any suggestions on how to handle it?
I have checked the GPU memory usage, it is normal.
Thank you for your reply, I closed the issue.
> Hey folks, currently published version only supports whatever condacolab runs on. However, the development version on `main` _should_ support arbitrary Python runtimes, but this hasn't been tested widely yet...