RuntimeError: "clamp_scalar_cpu" not implemented for 'Half'

[Tany-g]

This error occurred when I tried to run it.
/home/ubuntu/anaconda3/envs/streamdiffusion/lib/python3.10/site-packages/diffusers/image_processor.py:339: FutureWarning: Passing imageas torch tensor with value range in [-1,1] is deprecated. The expected value range for image tensor is [0,1] when passing as pytorch tensor or numpy Array. You passedimage with value range [-0.050994873046875,1.0] warnings.warn( Exception in thread Thread-1 (_receive_images): Traceback (most recent call last): File "/home/ubuntu/anaconda3/envs/streamdiffusion/lib/python3.10/threading.py", line 1016, in _bootstrap_inner self.run() File "/home/ubuntu/anaconda3/envs/streamdiffusion/lib/python3.10/threading.py", line 953, in run self._target(*self._args, **self._kwargs) File "/home/ubuntu/GITHUG/StreamDiffusion/examples/screen/../../utils/viewer.py", line 54, in _receive_images postprocess_image(queue.get(block=False), output_type="pil")[0], File "/home/ubuntu/anaconda3/envs/streamdiffusion/lib/python3.10/site-packages/streamdiffusion/image_utils.py", line 60, in postprocess_image [ File "/home/ubuntu/anaconda3/envs/streamdiffusion/lib/python3.10/site-packages/streamdiffusion/image_utils.py", line 61, in <listcomp> denormalize(image[i]) if do_denormalize[i] else image[i] File "/home/ubuntu/anaconda3/envs/streamdiffusion/lib/python3.10/site-packages/streamdiffusion/image_utils.py", line 13, in denormalize return (images / 2 + 0.5).clamp(0, 1) RuntimeError: "clamp_scalar_cpu" not implemented for 'Half'

Env:ubuntu1804
GPU:RTX3090*2
cuda:11.4

[cvv-student]

I also encountered this problem. The issue is that the CPU model does not support FP16 calculations. I switched it to FP32 calculations, and edited the error message in 'image_utils.py' in the function 'denormalize'. I added the following code:
'if images.dtype==torch.float16 and images.device.type=="cpu": images = images.float()
and that worked