Skip to content

Conversation

@asutermo
Copy link

@asutermo asutermo commented Feb 14, 2025

Several of the models are using removed or bad urls. I updated those and ran the scripts successfully. in standalone, and the lip syncer node in comfyui

This should fix some of the 'fake' cv errors popping up.

I added int codes for when pre_check fails and print messages to let you know when pre_check fails. Because there is no error handling for pre_check, the scripts continue trying to execute even when intermediate outputs were never created. This results in cases like LipSyncer failing with a cv error.

Also removed pycache and fixed a typo in gitignore

Sample Output:

(comfyui-work) PS D:\Work\src\ComfyUI-DeepFuze> python ./run.py --frame-processors face_enhancer --face-enhancer-model codeformer -t D:\\Work\\src\\ComfyUI\\output\\deepfuze\\17395604431507955.mp4 -o D:\\Work\\src\\ComfyUI\\output\\deepfuze\\enhanced_17395604431507955.mp4 --headless
Download https://github.com/facefusion/facefusion-assets/releases/download/models-3.0.0/open_nsfw.onnx to D:\Work\models\deepfuze
Download https://github.com/facefusion/facefusion-assets/releases/download/models-3.0.0/2dfan4.onnx to D:\Work\models\deepfuze
Download https://huggingface.co/JackCui/facefusion/resolve/main/face_landmarker_68_5.onnx to D:\Work\models\deepfuze
Download https://huggingface.co/JackCui/facefusion/resolve/main/gender_age.onnx to D:\Work\models\deepfuze
Download https://github.com/facefusion/facefusion-assets/releases/download/models-3.0.0/yoloface_8n.onnx to D:\Work\models\deepfuze
Download https://github.com/facefusion/facefusion-assets/releases/download/models-3.0.0/arcface_w600k_r50.onnx to D:\Work\models\deepfuze
Download https://huggingface.co/JackCui/facefusion/resolve/main/face_occluder.onnx to D:\Work\models\deepfuze
Download https://huggingface.co/JackCui/facefusion/resolve/main/face_parser.onnx to D:\Work\models\deepfuze
Download https://huggingface.co/JackCui/facefusion/resolve/main/voice_extractor.onnx to D:\Work\models\deepfuze
Download https://github.com/facefusion/facefusion-assets/releases/download/models-3.0.0/codeformer.onnx to D:\Work\models\deepfuze
Downloading: 100%|==================================================================| 359M/359M [00:08<00:00, 43.3MB/s]
Analysing: 100%|===========================================================| 94/94 [00:00<00:00, 698.79frame/s, rate=0]
[DEEPFUZE.CORE] Extracting frames with a resolution of 640x360 and 29.97002997002997 frames per second
[DEEPFUZE.PROCESSORS.FRAME.MODULES.FACE_ENHANCER] Processing
Processing: 100%|=| 94/94 [01:32<00:00,  1.02frame/s, execution_providers=['cpu'], execution_thread_count=4, execution_
[DEEPFUZE.CORE] Merging video with a resolution of 640x360 and 29.97002997002997 frames per second
[DEEPFUZE.CORE] Processing to video succeed in 93.53 seconds

@asutermo asutermo changed the title Fix pre_check calls by updating model paths Fix pre_check calls by updating model paths, fix red herring 'cv' errors. Feb 14, 2025
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant