报错 RuntimeError: "replication_pad3d_cuda" not implemented for 'BFloat16'

VAE解码出现报错

loaded completely 9.5367431640625e+25 470.1210079193115 True
!!! Exception during processing !!! "replication_pad3d_cuda" not implemented for 'BFloat16'
Traceback (most recent call last):
  File "/home/ruiy/sdb/Q/ComfyUI/execution.py", line 327, in execute
    output_data, output_ui, has_subgraph = get_output_data(obj, input_data_all, execution_block_cb=execution_block_cb, pre_execute_cb=pre_execute_cb)
  File "/home/ruiy/sdb/Q/ComfyUI/execution.py", line 202, in get_output_data
    return_values = _map_node_over_list(obj, input_data_all, obj.FUNCTION, allow_interrupt=True, execution_block_cb=execution_block_cb, pre_execute_cb=pre_execute_cb)
  File "/home/ruiy/sdb/Q/ComfyUI/execution.py", line 174, in _map_node_over_list
    process_inputs(input_dict, i)
  File "/home/ruiy/sdb/Q/ComfyUI/execution.py", line 163, in process_inputs
    results.append(getattr(obj, func)(**inputs))
  File "/home/ruiy/sdb/Q/ComfyUI/nodes.py", line 318, in decode
    images = vae.decode_tiled(samples["samples"], tile_x=tile_size // compression, tile_y=tile_size // compression, overlap=overlap // compression, tile_t=temporal_size, overlap_t=temporal_overlap)
  File "/home/ruiy/sdb/Q/ComfyUI/comfy/sd.py", line 531, in decode_tiled
    output = self.decode_tiled_3d(samples, **args)
  File "/home/ruiy/sdb/Q/ComfyUI/comfy/sd.py", line 453, in decode_tiled_3d
    return self.process_output(comfy.utils.tiled_scale_multidim(samples, decode_fn, tile=(tile_t, tile_x, tile_y), overlap=overlap, upscale_amount=self.upscale_ratio, out_channels=self.output_channels, index_formulas=self.upscale_index_formula, output_device=self.output_device))
  File "/home/ruiy/sdb/anaconda3/chy/lib/python3.9/site-packages/torch/utils/_contextlib.py", line 115, in decorate_context
    return func(*args, **kwargs)
  File "/home/ruiy/sdb/Q/ComfyUI/comfy/utils.py", line 926, in tiled_scale_multidim
    ps = function(s_in).to(output_device)
  File "/home/ruiy/sdb/Q/ComfyUI/comfy/sd.py", line 452, in <lambda>
    decode_fn = lambda a: self.first_stage_model.decode(a.to(self.vae_dtype).to(self.device)).float()
  File "/home/ruiy/sdb/Q/ComfyUI/comfy/ldm/models/autoencoder.py", line 209, in decode
    dec = self.decoder(dec, **decoder_kwargs)
  File "/home/ruiy/sdb/anaconda3/chy/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl
    return self._call_impl(*args, **kwargs)
  File "/home/ruiy/sdb/anaconda3/chy/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl
    return forward_call(*args, **kwargs)
  File "/home/ruiy/sdb/Q/ComfyUI/comfy/ldm/modules/diffusionmodules/model.py", line 709, in forward
    h = self.conv_in(z)
  File "/home/ruiy/sdb/anaconda3/chy/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl
    return self._call_impl(*args, **kwargs)
  File "/home/ruiy/sdb/anaconda3/chy/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl
    return forward_call(*args, **kwargs)
  File "/home/ruiy/sdb/Q/ComfyUI/comfy/ldm/modules/diffusionmodules/model.py", line 61, in forward
    x = torch.nn.functional.pad(x, self.padding, mode=self.padding_mode)
RuntimeError: "replication_pad3d_cuda" not implemented for 'BFloat16'

解决办法：

升级/降级torch, 

Previous PyTorch Versions | PyTorch

个人cuda12.1

升级torch至2.3.1解决问题

pip install torch==2.3.1 torchvision==0.18.1 torchaudio==2.3.1 --index-url https://download.pytorch.org/whl/cu121