search

initialneil / SplattingAvatar

[CVPR2024] Official implementation of SplattingAvatar.
Other
336 stars 32 forks source link

Question about Evaluation code #21

Open zhoutianyang2002 opened 3 months ago

zhoutianyang2002 commented 3 months ago

Hi! I am trying to run the given demo. I use the dataset person_0004. I have already finished the training. However, when I trying to evaluate the model I encounter an error. what's wrong with it? I am a newbee of computer graphics so it is difficult for me to understand the source code now. Thanks a lot!

image image
DongChengdongHangZhou commented 2 months ago

Dear Tianyang,

I can help you resolve this problem.

The authors of this repository made a mistake, when you run the eval_splatting_avatar.py, you should organize the checkpoints as follows:

results/output-splatting/last_checkpoint |---eval_30000 |---image |---point_cloud |---config.yaml |---flame_params.json

the folder last_checkpoint/image should be the same as person004/person004/images, and you should move person004/person004/flame_params.json to results/output-splatting/last_checkpoint/flame_params.json

Finally you should run the command as follows: python eval_splatting_avatar.py --config configs/splatting_avatar.yaml --dat_dir results/output-splatting/last_checkpoint --pc_dir results/output-splatting/last_checkpoint/point_cloud/iteration_30000

In addition, the authors didn't provide the landmark_embedding.npy, and I finally find it on https://github.com/yfeng95/DECA/tree/master/data you can download it from this weblink

zhoutianyang2002 commented 2 months ago

Dear Tianyang,

I can help you resolve this problem.

The authors of this repository made a mistake, when you run the eval_splatting_avatar.py, you should organize the checkpoints as follows:

results/output-splatting/last_checkpoint |---eval_30000 |---image |---point_cloud |---config.yaml |---flame_params.json

the folder last_checkpoint/image should be the same as person004/person004/images, and you should move person004/person004/flame_params.json to results/output-splatting/last_checkpoint/flame_params.json

Finally you should run the command as follows: python eval_splatting_avatar.py --config configs/splatting_avatar.yaml --dat_dir results/output-splatting/last_checkpoint --pc_dir results/output-splatting/last_checkpoint/point_cloud/iteration_30000

In addition, the authors didn't provide the landmark_embedding.npy, and I finally find it on https://github.com/yfeng95/DECA/tree/master/data you can download it from this weblink

Thank you for your clear and helpful reply! Best wishes!