Innovate Futures @ Benji
patreon
Upcoming video, performance comparison on RAVE and Animatediff
🕑 Added 2024-01-25 21:20:40 +0000 UTCThroughout the test on this few day. I have recorded some from running queue.
Here is what I came out in result, hope this help when you decide to apply RAVE or not. This is really depends on situation.
===RAVE + AnimateDiff===
- 3 or 4 times memories (Vram, gpu power and RAM) consume compare with AnimateDiff only. Therefore, it limited less frame images are able to gen. On each queue.
- Text Propmt do play the key role on styling , even I have use IPA , the styling will not follow the same as IPA result.
- Depth map control net do better for styling totally different outfit. If Lineart , sometime keep the sources video shadows, and shape of the source outfit.
===AnimateDiff only===
- It can do consistent style with text and image prompts, but 90% apply on the output video.
RAVE , 95% consistency following the text prompt.
- It fast an lightweight with LCM.
- able to use IPAdapter for character (s) and background styling.
Hope this helps.
Comments
Benjamin Law
Hi, Looks like this is the last part upscale ? It mentioned the path value should be string or number, have you miss something in setting? It cannot leave it empty in one of the setting. Thanks
yohance modeste
hi sorry to bother you so much but i keep having this error before i get my final video. ERROR:root:!!! Exception during processing !!! ERROR:root:Traceback (most recent call last): File "C:\ComfyUI_windows_portable\ComfyUI\execution.py", line 155, in recursive_execute output_data, output_ui = get_output_data(obj, input_data_all) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "C:\ComfyUI_windows_portable\ComfyUI\execution.py", line 85, in get_output_data return_values = map_node_over_list(obj, input_data_all, obj.FUNCTION, allow_interrupt=True) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "C:\ComfyUI_windows_portable\ComfyUI\execution.py", line 78, in map_node_over_list results.append(getattr(obj, func)(**slice_dict(input_data_all, i))) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "C:\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-Impact-Pack\modules\impact\impact_pack.py", line 109, in load_model size = os.path.getsize(modelname) ^^^^^^^^^^^^^^^^^^^^^^^^^^ File "", line 50, in getsize TypeError: stat: path should be string, bytes, os.PathLike or integer, not NoneType
