willisma/SiT

Issues

• Some questions about the open-source plan

  #28 opened 16 days ago by stein-666
  0

• Some Questions about Training

  #27 opened 16 days ago by stein-666
  0

• sample_eps maybe None

  #26 opened 22 days ago by DQSSSSS
  0

• Regarding timestep embeddings

  #24 opened a month ago by realfolkcode
  3

• Some questions about forward_with_cfg

  #23 opened 2 months ago by maxin-cn
  1

• Has anyone encounter that using SDE will produce nan vector while the results by using ODE is correct?

  #22 opened 3 months ago by xmhGit
  3

• What is the function of the learn_sigma in SiT model forward?

  #20 opened 3 months ago by xmhGit
  3

• Wandb issue

  #7 opened 3 months ago by stefenmax
  1

• Releasing base (B) and large (L) model weights

  #13 opened 3 months ago by treasan
  1

• A bug while training with multi-nodes

  #8 opened 3 months ago by JingfengYao
  1

• Parameters for reproducing the FID of the released XL/2

  #18 opened 3 months ago by HanshuYAN
  1

• Current resume training does not record and continue the epoch and step number correctly

  #19 opened 3 months ago by xmhGit
  1

• The number of function evaluation (NFE) in ODE mode

  #21 opened 4 months ago by LucasYFL
  6

• The reproduction related

  #14 opened 7 months ago by WANGSSSSSSS
  5

• Question about proof in the paper

  #15 opened 6 months ago by chinmay5
  2

• Maybe inconsistent behavior of the method compute_drift() in VPCPlan and ICPlan ?

  #17 opened 6 months ago by zhengyunqq
  0

• Any plan for developing a diffusers version?

  #16 opened 7 months ago by JincanDeng
  0

• OOM error

  #11 opened 7 months ago by zen-d
  5

• Question about sample process

  #12 opened 7 months ago by wl3181
  1

• size mismatch for final_layer.linear.weight and final_layer.linear.bias

  #10 opened 8 months ago by GreameLee
  1

• AttributeError: 'Namespace' object has no attribute 'model'

  #4 opened 8 months ago by stefenmax
  1

• can not download the model

  #3 opened 8 months ago by silvercherry
  1

• Error when loading trained checkpoint

  #9 opened 8 months ago by pacocp
  0

• Why do ema models not need gradients only when there are pre-trained weights?

  #6 opened 9 months ago by LinB203
  1

• What is the version of timm?

  #5 opened 9 months ago by YuyangYin
  0

• A Kind Inquiry for Releasing SiT-XL/2 Models

  #2 opened 9 months ago by yifanpu001
  4

• sample_eps/train_eps are not passed to Transport in create_transport

  #1 opened 10 months ago by Daiver
  1