search

ashawkey / RAD-NeRF

Real-time Neural Radiance Talking Portrait Synthesis via Audio-spatial Decomposition
MIT License
878 stars 153 forks source link

How do I get json file?? #12

Closed merif closed 1 year ago

merif commented 1 year ago

I'm going to train to make the video into a portrait. How do I get the files I need for learning?

For example, the Obama video.

  1. transforms_train(val).json
  2. track_params.pt
  3. additionally-need

What more resources do I need to learn custom videos? There are too many obstacles to learning through scripts. Thank you.

ashawkey commented 1 year ago

@merif Hi, please check the dara processing part in readme.

merif commented 1 year ago

The above problem occurs when i run this part {python data_utils/process.py data//.mp4} in the Data-pre-processing part.

[Errno 2] No such file or directory: 'data/obama/track_params.pt' The problem occurs when I finish Preparation of Data-pre-processing(Download face_tracking/3DMM files) and use Runscript as a custom learning video(already prepare custom trainning video).

When experimenting with the example Obama video, the above error occurred in the pre-processing part, so I used the data downloaded from the drive. And, another error occurs if you use a pt that has already been downloaded through the Obama video drive.

No such file or directory: 'data/obama/transforms_train.json'

I faced the problem of not being able to find the file even though I followed the data pre-processing part

How do I get json and pt files?? Thank you for your quick response.

ashawkey commented 1 year ago

It seems the face tracking failed. There should be another error before this?

merif commented 1 year ago

I read the comments and installed the environment again .I think the installation was missing in the middle of changing the conda version! It works fine now. Thank you!!

Oyiyi commented 8 months ago

@merif which installation you missed? Do you remember the package for example? I had the same issue but I'm sure I installed every step. Thank you in advance!