Mael-zys/T2M-GPT

Issues

• t2m extractor files are no longer available

  #71 opened 3 months ago by PPierzc
  2

• How to convert the 3D data into 2D form？

  #72 opened 3 months ago by Pride-Huang
  0

• A long time during evaluation

  #67 opened 4 months ago by shuowang666
  1

• About text_mot_match

  #68 opened 4 months ago by Haifeng-Lu
  1

• About Sampling and Calculating FID Score

  #69 opened 4 months ago by RohollahHS
  4

• SMPL mesh rendering sigle image

  #70 opened 4 months ago by ChangeNext
  4

• Render Multiple Frames in Single Image as Presented in the Paper

  #55 opened 5 months ago by exitudio
  1

• Can this really run with just 32GB vram?

  #45 opened a year ago by YOUNG-WLI
  1

• Can not reproduce the reported results.

  #61 opened 9 months ago by blue-blue272
  3

• Unable to reproduce the results using the pretrained weights

  #28 opened 2 years ago by kentfuji
  6

• Motion in-betweening

  #63 opened 8 months ago by julkaztwittera
  1

• VQVAE -- translation initialization of training sequences

  #64 opened 8 months ago by purvaten
  2

• Why is the length of the generated action sequences always different?

  #65 opened 8 months ago by DINGYANB
  3

• One question about VQ-VAE

  #59 opened 9 months ago by smandava98
  1

• Can you provide the pretrained model for HumanML3D?

  #66 opened 9 months ago by feifeitu1
  2

• Dataset difference may exist between your work and official HumanML3D

  #49 opened a year ago by qrzou
  1

• The issue on the evaluation of VQVAE

  #62 opened a year ago by LinghaoChan
  2

• ValueError: Imaginary component

  #22 opened 2 years ago by npapargyr
  1

• 2D motion prediction and keypoint format

  #60 opened a year ago by julkaztwittera
  1

• Error in running demo code

  #27 opened a year ago by ZombieLei
  1

• Use Hugging Face

  #31 opened a year ago by ymx-fat
  0

• How to calculate the trans between each frame?

  #50 opened a year ago by laugher07
  1

• Constraining starting and ending pose

  #56 opened a year ago by LeyangWen
  1

• where is training code for models used in evaluations?

  #58 opened a year ago by kwon852456
  1

• render smpl gif not the same as joints

  #25 opened 2 years ago by ywdong
  4

• KIT VQ-VAE Training details

  #52 opened a year ago by EricGuo5513
  1

• How many motion indexes do you have?

  #53 opened a year ago by gwyong
  5

• small question about the vq-vae paper

  #57 opened a year ago by kwon852456
  0

• Maximum length of the code index sequence

  #42 opened a year ago by npapargyr95
  4

• Generate SMPL Mesh for rotation data in colab

  #54 opened a year ago by Mighty303
  0

• A Collab result inconsistent with the paper

  #23 opened 2 years ago by kingchurch
  2

• About the model's height and body shape

  #51 opened a year ago by shingo1216
  0

• Difference between your work and TM2T.

  #39 opened a year ago by phj128
  1

• T2M-GPT generated the Fatal Error joints points

  #46 opened a year ago by Thunderltx
  3

• question about vq-vae reconstruction loss

  #48 opened a year ago by pangwong
  0

• Number of motions computation

  #47 opened a year ago by rd20karim
  1

• Controlling seed

  #44 opened a year ago by lassemt
  2

• export onnx model

  #43 opened a year ago by songjiahao-wq
  1

• [Feature] How to use `QuantizeEMAReset` with DDP elegantly and efficiently?

  #34 opened 2 years ago by Dai-Wenxun
  1

• Why the first stage dataset is only 64 frames?

  #40 opened a year ago by exitudio
  2

• Environment file

  #35 opened 2 years ago by tommyshelby4
  2

• question about motion representations

  #38 opened 2 years ago by leisurehippo
  0

• Generate SMPL mesh

  #37 opened 2 years ago by laugher07
  0

• Getting issue while evaluating model

  #36 opened 2 years ago by LokeshSaipureddi
  3

• Can you provide the pretrained model for KIT?

  #32 opened 2 years ago by qiqiApink
  1

• How were the meta data for the mean and standard deviation generated ?

  #33 opened 2 years ago by kingchurch
  1

• The result of real motion

  #30 opened 2 years ago by Ssstirm
  1

• training from scratch

  #29 opened 2 years ago by deeptimhe
  2

• There is a mistake in `dataset/prepare/download_smpl.sh`

  #26 opened 2 years ago by spzhuang
  4

• 95% confidence interval

  #24 opened 2 years ago by exitudio
  5