The best in LivePortrait ❤️

[medalawi]

LivePortrait ❤️ works good on GPUs 4GB VRAM this will make liveportrait work on all PCs , while ❌other projects ❌that have audio-driven feature require More VRAM 😞 I wish Liveportrait add audio-driven feature soon here is a list of audio-driven project that need high GPU VRAM to work

• Echomimic
• Hallo

[wfnian]

❤️

[Manamama]

FYI, after some fixes (many .gpu() to .cpu() ) , it started to work on a CPU only modest notebook:

Central Processing Unit: Intel i5-8265U (8) @ 3.9GHz 
CPU Usage: 35% 
Graphics Processing Unit: Intel WhiskeyLake-U GT2 [UHD Graphics 620] 
GPU Driver: Hewlett-Packard Company WhiskeyLake-U GT2 [UHD Graphics 620] [103c:84bb] 
Memory: 6612MiB / 7814MiB 
Disk (/): 108G / 116G (97%) 
Battery0: 5% [Charging] 
Uptime: 2 days, 51 mins 
------ 
Host: HP Pavilion Laptop

  model.load_state_dict(torch.load(ckpt_path, map_location=lambda storage, loc: storage))
[17:50:22] Load appearance_feature_extractor done.                                                                                                                                      live_portrait_wrapper.py:30
[17:50:23] Load motion_extractor done.                                                                                                                                                  live_portrait_wrapper.py:33
[17:50:24] Load warping_module done.                                                                                                                                                    live_portrait_wrapper.py:36
[17:50:25] Load spade_generator done.                                                                                                                                                   live_portrait_wrapper.py:39
           Load stitching_retargeting_module done.                                                                                                                                      live_portrait_wrapper.py:43
/home/xxx/.local/lib/python3.10/site-packages/onnxruntime/capi/onnxruntime_inference_collection.py:69: UserWarning: Specified provider 'CUDAExecutionProvider' is not in available provider names.Available providers: 'AzureExecutionProvider, CPUExecutionProvider'
  warnings.warn(
           LandmarkRunner warmup time: 0.082s                                                                                                                                                 landmark_runner.py:89
           FaceAnalysisDIY warmup time: 0.121s                                                                                                                                              face_analysis_diy.py:79
[17:50:26] Load source image from /home/xxx/Downloads/LivePortrait/src/config/../../assets/examples/source/s6.jpg                                                                    live_portrait_pipeline.py:46
[17:52:20] Load from video file (mp4 mov avi etc...): /home/xxx/Downloads/LivePortrait/src/config/../../assets/examples/driving/d0.mp4                                               live_portrait_pipeline.py:72
/home/xxx/.local/lib/python3.10/site-packages/imageio/plugins/pyav.py:779: MethodDeprecationWarning: SideData.to_bytes is deprecated.
  {item.type.name: item.to_bytes() for item in desired_frame.side_data}
Animating... ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━   0% -:--:--

I plan to leave it for a week or two, running, to see if by chance smth is indeed produced ;).