cuda not accepting

[raochinmay6]

Error occurred when executing ImageConcatMulti:

[enforce fail at alloc_cpu.cpp:114] data. DefaultCPUAllocator: not enough memory: you tried to allocate 33288192000 bytes.

File "E:\ComfyUI_windows_portable\ComfyUI\execution.py", line 151, in recursive_execute output_data, output_ui = get_output_data(obj, input_data_all) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "E:\ComfyUI_windows_portable\ComfyUI\execution.py", line 81, in get_output_data return_values = map_node_over_list(obj, input_data_all, obj.FUNCTION, allow_interrupt=True) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "E:\ComfyUI_windows_portable\ComfyUI\execution.py", line 74, in map_node_over_list results.append(getattr(obj, func)(**slice_dict(input_data_all, i))) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "E:\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-KJNodes\nodes\image_nodes.py", line 1288, in combine image, = ImageConcanate.concanate(self, image, new_image, direction, match_image_size, first_image_shape=first_image_shape) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "E:\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-KJNodes\nodes\image_nodes.py", line 257, in concanate concatenated_image = torch.cat((image1, image2_resized), dim=2) # Concatenate along width ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

[kijai]

The concat node can be pretty memory hungry with larger resolutins/frame counts, I recommend simply not using it in those cases, it's only to make the comparison videos.

[lyonnel4]

Hi. I had that issue and I solved it by adding a Save image node at the output of the LivePortrait process node. Instead of using a lot of ram for concatenate the pictures, it search for the images on the drive and there's no more RAM issue. that solved also a freeze I had once reaching 100% (I guess it was the video combine VHS node that used a lot of RAM too)

[raochinmay6]

Ok let me try that