motion-latent-diffusion

motion-latent-diffusion copied to clipboard

Training configuration on KIT dataset

8

Hi, I would like to train MLD on KIT dataset. What should I tune in this configuration file [https://github.com/ChenFengYe/motion-latent-diffusion/blob/main/configs/config_vae_humanml3d.yaml](url) ? Its JOINT_TYPE may be configure as 'mmm'? Many Thanks

Dean-UQ

SkipTransformer and None Skip Transformer

1

Hi, I tried to modify the MLD-VAE from skipTransformer to the normal transformer, but I found that the loss doesn't drop at around total loss = 0.17. But I found...

ZenGengChin

Hello，I want to know How many epochs do you train on your VAE？Thank you for your reply！！！

feifeitu1

Hello, I must say amazing work. I was wondering if you could clarify how did you get the word embedding for each word as 300 in the our_vab_data.npy??

kaustesseract

Why not using VQVAE like stable diffusion?

10

Thank you for bringing interesting work. I'm curious if you have tried VQVAE?

daidaiershidi

Error with "name_list" and "length_list" in datasets

7

Congrats on your great paper! I am running the "**python demo.py --cfg ./configs/config_mld_humanml3d.yaml --cfg_assets ./configs/assets.yaml --example ./demo/example.txt**" to get demo results on my colab. First of all, data should be...

alsokh

Result of novae model

6

Hi, authors! I used your pre-trained novae model and no VAE config to evaluate on HumanML3D dataset but got very low performance(FID 38, R-precision 0.01). Are there any bugs in...

hoyeYang

Visualize SMPL

3

Thanks for nice woel! I have tried to visualize smpl by appling SMPLify to 3D coordinates to SMPL parameters, and 6890 vertices' coordinares according to your instuctions. However, the final...

wang-zm18

Render SMPL meshes

2

Thank you for your great work. I am trying to render SMLP using blender follow your instruction in "Visualization/ 3. Render SMPL meshes" section. I also change **FACES_PATH** in *configs/render_mld.yaml*...

exitudio

Results Stage1 KIT Dataset

3

Hi, thank you for your work, I find it brilliant and really useful for future developments in this field! Regarding the HumanML3D dataset, the results are easily replicable. However when...

AlessioSam