Compile with TORCH_USE_CUDA_DSA to enable device-side assertions.

[chensyo]

新版本0.14更新后报这个错误，之前0.13是可以正常使用的，python3.11，torch2.60+cuda12.6
!!! Exception during processing !!! CUDA error: operation not supported
Compile with TORCH_USE_CUDA_DSA to enable device-side assertions.

Traceback (most recent call last):
File "D:\Program Files\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 "D:\Program Files\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 "D:\Program Files\ComfyUI\execution.py", line 174, in _map_node_over_list
process_inputs(input_dict, i)
File "D:\Program Files\ComfyUI\execution.py", line 163, in process_inputs
results.append(getattr(obj, func)(**inputs))
^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "D:\Program Files\ComfyUI\custom_nodes\svdquant\nodes\models\flux.py", line 134, in load_model
transformer = transformer.to(device)
^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\Mingzhenwang\AppData\Roaming\Python\Python311\site-packages\diffusers\models\modeling_utils.py", line 1077, in to
return super().to(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\Mingzhenwang\AppData\Roaming\Python\Python311\site-packages\torch\nn\modules\module.py", line 1343, in to
return self._apply(convert)
^^^^^^^^^^^^^^^^^^^^
File "C:\Users\Mingzhenwang\AppData\Roaming\Python\Python311\site-packages\torch\nn\modules\module.py", line 903, in _apply
module._apply(fn)
File "C:\Users\Mingzhenwang\AppData\Roaming\Python\Python311\site-packages\torch\nn\modules\module.py", line 903, in _apply
module._apply(fn)
File "C:\Users\Mingzhenwang\AppData\Roaming\Python\Python311\site-packages\torch\nn\modules\module.py", line 903, in _apply
module._apply(fn)
File "C:\Users\Mingzhenwang\AppData\Roaming\Python\Python311\site-packages\torch\nn\modules\module.py", line 930, in _apply
param_applied = fn(param)
^^^^^^^^^
File "C:\Users\Mingzhenwang\AppData\Roaming\Python\Python311\site-packages\torch\nn\modules\module.py", line 1329, in convert
return t.to(
^^^^^
RuntimeError: CUDA error: operation not supported
Compile with TORCH_USE_CUDA_DSA to enable device-side assertions.

Prompt executed in 5.77 seconds
FETCH ComfyRegistry Data: 10/36
FETCH ComfyRegistry Data: 15/36
got prompt
[2025-03-08 21:26:51.585] [info] Initializing QuantizedFluxModel
[2025-03-08 21:26:52.465] [info] Loading weights from D:\Program Files\ComfyUI\models\diffusion_models\svdq-int4-flux.1-schnell\transformer_blocks.safetensors
[2025-03-08 21:26:52.466] [warning] Failed to load safetensors using method MIO: CUDA error: operation not supported (at D:\Program Files\BuildWhl\nunchaku\src\Serialization.cpp:130)
[2025-03-08 21:26:57.272] [info] Done.
!!! Exception during processing !!! CUDA error: operation not supported
Compile with TORCH_USE_CUDA_DSA to enable device-side assertions.

Traceback (most recent call last):
File "D:\Program Files\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 "D:\Program Files\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 "D:\Program Files\ComfyUI\execution.py", line 174, in _map_node_over_list
process_inputs(input_dict, i)
File "D:\Program Files\ComfyUI\execution.py", line 163, in process_inputs
results.append(getattr(obj, func)(**inputs))
^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "D:\Program Files\ComfyUI\custom_nodes\svdquant\nodes\models\flux.py", line 134, in load_model
transformer = transformer.to(device)
^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\Mingzhenwang\AppData\Roaming\Python\Python311\site-packages\diffusers\models\modeling_utils.py", line 1077, in to
return super().to(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\Mingzhenwang\AppData\Roaming\Python\Python311\site-packages\torch\nn\modules\module.py", line 1343, in to
return self._apply(convert)
^^^^^^^^^^^^^^^^^^^^
File "C:\Users\Mingzhenwang\AppData\Roaming\Python\Python311\site-packages\torch\nn\modules\module.py", line 903, in _apply
module._apply(fn)
File "C:\Users\Mingzhenwang\AppData\Roaming\Python\Python311\site-packages\torch\nn\modules\module.py", line 903, in _apply
module._apply(fn)
File "C:\Users\Mingzhenwang\AppData\Roaming\Python\Python311\site-packages\torch\nn\modules\module.py", line 903, in _apply
module._apply(fn)
File "C:\Users\Mingzhenwang\AppData\Roaming\Python\Python311\site-packages\torch\nn\modules\module.py", line 930, in _apply
param_applied = fn(param)
^^^^^^^^^
File "C:\Users\Mingzhenwang\AppData\Roaming\Python\Python311\site-packages\torch\nn\modules\module.py", line 1329, in convert
return t.to(
^^^^^
RuntimeError: CUDA error: operation not supported
Compile with TORCH_USE_CUDA_DSA to enable device-side assertions.

[sxtyzhangzk]

Could you check if the latest commit fixes this issue? Thanks.

[chensyo]

git pull 最新的main，看到应该是更新了common.h，重新编译安装了一下，仍然报错
RuntimeError: CUDA error: operation not supported
CUDA kernel errors might be asynchronously reported at some other API call, so the stacktrace below might be incorrect.
For debugging consider passing CUDA_LAUNCH_BLOCKING=1
Compile with TORCH_USE_CUDA_DSA to enable device-side assertions.

[hdfhssg]

git pull 最新的main，看到应该是更新了common.h，重新编译安装了一下，仍然报错 RuntimeError: CUDA error: operation not supported CUDA kernel errors might be asynchronously reported at some other API call, so the stacktrace below might be incorrect. For debugging consider passing CUDA_LAUNCH_BLOCKING=1 Compile with to enable device-side assertions.TORCH_USE_CUDA_DSA

你是Windows10吧，我Windows10也有这个报错，我修改文件已经解决它了，我再测试测试没问题我把我改的放上来吧，或者你等他们官方看看有没有更好的修改方式

[sxtyzhangzk]

git pull 最新的main，看到应该是更新了common.h，重新编译安装了一下，仍然报错 RuntimeError: CUDA error: operation not supported CUDA kernel errors might be asynchronously reported at some other API call, so the stacktrace below might be incorrect. For debugging consider passing CUDA_LAUNCH_BLOCKING=1 Compile with TORCH_USE_CUDA_DSA to enable device-side assertions.

Could you try setting the environment variable NUNCHAKU_LOAD_METHOD?
Try
set NUNCHAKU_LOAD_METHOD=READ
or
set NUNCHAKU_LOAD_METHOD=READNOPIN

[chensyo]

git pull 最新的main，看到应该是更新了common.h，重新编译安装了一下，仍然报错 RuntimeError: CUDA error: operation not supported CUDA kernel errors might be asynchronously reported at some other API call, so the stacktrace below might be incorrect. For debugging consider passing CUDA_LAUNCH_BLOCKING=1 Compile with to enable device-side assertions.TORCH_USE_CUDA_DSA

你是Windows10吧，我Windows10也有这个报错，我修改文件已经解决它了，我再测试测试没问题我把我改的放上来吧，或者你等他们官方看看有没有更好的修改方式

是的，我是win10.之前0.1.3版本报那个Unable to pin memory: operation not supported错误然后闪退，后面拿你改的Serialization.cpp编译完虽然也报错但是可以正常使用，现在更新了又不行了

[hdfhssg]

git pull 最新的main，看到应该是更新了common.h，重新编译安装了一下，仍然报错 RuntimeError: CUDA error: operation not supported CUDA kernel errors might be asynchronously reported at some other API call, so the stacktrace below might be incorrect. For debugging consider passing CUDA_LAUNCH_BLOCKING=1 Compile with to enable device-side assertions.TORCH_USE_CUDA_DSA

Could you try setting the environment variable ? Try orNUNCHAKU_LOAD_METHOD``set NUNCHAKU_LOAD_METHOD=READ set NUNCHAKU_LOAD_METHOD=READNOPIN

OK, I'll try later.

[chensyo]

git pull 最新的main，看到应该是更新了common.h，重新编译安装了一下，仍然报错 RuntimeError: CUDA error: operation not supported CUDA kernel errors might be asynchronously reported at some other API call, so the stacktrace below might be incorrect. For debugging consider passing CUDA_LAUNCH_BLOCKING=1 Compile with TORCH_USE_CUDA_DSA to enable device-side assertions.

Could you try setting the environment variable NUNCHAKU_LOAD_METHOD? Try set NUNCHAKU_LOAD_METHOD=READ or set NUNCHAKU_LOAD_METHOD=READNOPIN

这个可以，恢复正常了，感谢