openai/consistency_models

Issues

• the parameter "ts" and "step" in EDM Sample and CD sample

  #65 opened 20 days ago by szh404
  0

• Multistep sampling in Algorithm 1

  #7 opened 2 years ago by liyy201912
  10

• Why does the training not stop？

  #51 opened a year ago by ZhaohanWang0217
  1

• Inconsistent loss term with paper

  #12 opened 2 years ago by quantumiracle
  7

• Why use teacher model in Consistency training

  #53 opened 10 months ago by CellNw
  6

• from flash_attn.flash_attention import FlashAttention . Where is FlashAttention？Why can't I import it?

  #64 opened 5 months ago by yaoyaoleY
  5

• the libfabric EFA provider is operating in a condition that could result in memory corruption or other system errors.

  #63 opened 6 months ago by kuailexiaohunzi
  11

• Code for Zero-Shot Image Editing

  #5 opened 2 years ago by SDNAFIO
  8

• After running, no .npz file appeared

  #62 opened 8 months ago by chuanchengh
  2

• Super-resolution conditional training

  #29 opened 2 years ago by CallShaul
  3

• Is there a non-stochastic sampling method?

  #55 opened a year ago by krNeko9t
  1

• How is Consistency for Any Point on the ODE Path Achieved in the Consistency Model?

  #61 opened 8 months ago by 2019211753
  0

• pip install -e . error

  #31 opened 2 years ago by KwangryeolPark
  2

• Issue with `use_fp16=True` Leading to Type Conversion Error in `unet.py`

  #56 opened a year ago by DuoLi1999
  1

• cannot import name '_get_cpp_backtrace' from 'torch._C'请问这种问题该怎样解决呢

  #58 opened 10 months ago by liuyisu
  2

• ModuleNotFoundError: No module named 'cm'

  #52 opened a year ago by michaeltan53
  3

• The difference between Loss( f(x0+t_{n+1}*z) , f(x0+t_{n}*z) ) and Loss( f(x0+t_{n+1}*z), x0 )?

  #11 opened 2 years ago by ZGCTroy
  4

• An module error when running cm_train.py

  #57 opened a year ago by RICKand-MORTY
  1

• Is the sampling reversible? (Image to Noise)

  #15 opened 2 years ago by discordance
  3

• Can I use the checkpoint provided by the original EDM paper to train the consistency model?

  #54 opened a year ago by SYZhang0805
  0

• Training settings and pretrained models for CIFAR-10

  #39 opened a year ago by kayson-tihoi
  4

• GPU Training Time for CD and CT Models on ImageNet64X64?

  #50 opened a year ago by ZSHsh98
  0

• Request for unconditional model on ImageNet 64*64

  #49 opened a year ago by TRISKEL10N
  0

• How to implement on stable diffusion?

  #8 opened 2 years ago by ZGCTroy
  3

• [inpainting] wrong shape?

  #48 opened a year ago by jS5t3r
  0

• where can we get the ref_batches/imagenet64.npz file for FID evaluation

  #47 opened a year ago by lonestar1990
  1

• “multi-gpu error” dist.all_gather(gathered_samples, sample) # gather not supported with NCCL

  #46 opened a year ago by fikry102
  1

• RuntimeError: FlashAttention is only supported on CUDA 11 and above

  #23 opened a year ago by zhouzk5
  8

• Different batches size while training

  #45 opened a year ago by DannyXYZa
  0

• RuntimeError in the sample `diffusers` code in README.md

  #44 opened a year ago by take2rohit
  2

• Sampling with Diffusers

  #42 opened a year ago by RossoneriZhao
  1

• The parameter sigma_data for KarrasDenoiser.

  #41 opened a year ago by moonbow721
  0

• Suffering problem when installing dependence using pip install -e . on google colab

  #36 opened a year ago by MichaelPei2003
  2

• torch version

  #9 opened 2 years ago by XipengY
  5

• Not able to obtain results like the checkpoint models (CT, imagenet64) when trying to train from scratch

  #27 opened 2 years ago by aarontan-git
  1

• Is there any consistency model on face dataset?

  #30 opened 2 years ago by njuaplusplus
  1

• Description The sampling command blocked

  #34 opened 2 years ago by zs993381843
  1

• I use the pre-trained model to generate pictures, the effect is too bad, I don't know what's going on.

  #25 opened 2 years ago by stonecropa
  9

• implement a non-Cuda flash attention module

  #33 opened 2 years ago by by321
  1

• I wonder how to train my own datasets based on your models

  #4 opened 2 years ago by cg-light
  10

• An error was reported while running the project：

  #32 opened 2 years ago by zs993381843
  1

• Use one gpu to generate images using a pretrained model without the communication protocol nccl.

  #21 opened 2 years ago by stonecropa
  25

• CUDA out of memory

  #28 opened 2 years ago by xhsongstanford
  1

• question about running scripts: time out initializing process group

  #14 opened 2 years ago by boxwayne
  3

• why lg_loss_scale？

  #26 opened 2 years ago by 1999kevin
  0

• Specifying a specific image as supervisory information for the generated image？

  #22 opened 2 years ago by YLBGHXTX
  1

• Evaluation: Reference Batches for CIFAR10

  #10 opened 2 years ago by nfrumkin
  1

• How to visualize samples stored by np?

  #24 opened 2 years ago by ShyFoo
  4

• How to run on a single linux server with multiple GPUs

  #20 opened 2 years ago by 1999kevin
  12

• Differences between code and paper

  #18 opened 2 years ago by thorinf
  2