search

daswer123 / hallo-webui

Webui for Hallo: Hierarchical Audio-Driven Visual Synthesis for Portrait Image Animation
https://fudan-generative-vision.github.io/hallo/
82 stars 29 forks source link

Issue when running inference : Protobuf parsing failed. #3

Open moda20 opened 3 months ago

moda20 commented 3 months ago

Hi whenever I am running the inference I am getting this error : 

Load model from ./pretrained_models/face_analysis/models/1k3d68.onnx failed:Protobuf parsing failed.

I want to know if this is related to the model size as i only have 8GB of VRAM ? if not is there a way to fix it ?

rvsh2 commented 3 months ago

Got the same issue but I have 24gb VRAM. Used docker. How to fix?

moda20 commented 3 months ago

@rvsh2 i created the docker container for this repo, if you used that it might be the source of our problem

chenav commented 2 months ago

same problem in a docker container. Same error, impossible to run. at all

TemporalLabsLLC-SOL commented 1 month ago

This is because the models inside '/pretrained_models` folder are outdated when pulling from what this repo was originally built with.

FOR NOW : https://huggingface.co/fudan-generative-ai/hallo/tree/main and get the latest to replace manually.

It works.

moda20 commented 1 month ago

@TemporalLabsLLC-SOL can you share a full example of how to use the new models ? they are not total replacement as i just tested them and i am getting the same error : 

onnxruntime.capi.onnxruntime_pybind11_state.InvalidProtobuf: [ONNXRuntimeError] : 7 : INVALID_PROTOBUF : Load model from ./pretrained_models/face_analysis/models/1k3d68.onnx failed:Protobuf parsing failed.
bensonbs commented 1 month ago

For me use command

  1. apt-get install git-lfs
  2. git lfs install
  3. git lfs clone https://huggingface.co/fudan-generative-ai/hallo pretrained_models

to fix it.

moda20 commented 3 weeks ago

@bensonbs how much VRAM you have ?mine iss 8GB