xdit-project/xDiT

Issues

• .

  #444 opened 2 months ago by shaun-ba
  1

• [Support] Request for supporting Sd3.5 Medium

  #372 opened 4 months ago by pprp
  4

• Do you have plan to support some Text to Speech Model？

  #402 opened 3 months ago by ZhikangNiu
  6

• Does xDiT have plans to support teacache?

  #435 opened 2 months ago by DefTruth
  1

• question about training pipeline

  #440 opened 2 months ago by bendanzzc
  1

• Http server crashes on "paralleling scheduler."

  #441 opened 2 months ago by Aktsvigun
  1

• Tensor Mismatch Error for CogVideoX1.5-5B with Official Script

  #439 opened 2 months ago by liming-ai
  1

• No Improvement with DiTFastattn for opensora or video models

  #428 opened 2 months ago by nighting0le01
  2

• huggingface-cli models

  #433 opened 2 months ago by kazakovaanastasia
  1

• quantization support

  #436 opened 2 months ago by Andy0422
  1

• !bash examples/run_consisid_usp.sh error:cannot import name 'ConsisIDPipeline' from 'diffusers' (python version env maybe?)

  #432 opened 2 months ago by kazakovaanastasia
  1

• Any plans to support AMD GPUs?

  #437 opened 2 months ago by rtlinux
  1

• How to use xDiT speed cogvideox-5B-v1.5-I2V？

  #385 opened 3 months ago by wen020
  3

• How to configure or replace sample

  #425 opened 2 months ago by lcax200000
  2

• PixArt model torchrun error

  #430 opened 2 months ago by kazakovaanastasia
  1

• runs always stucks on INFO 01-12 20:17:39 [base_pipeline.py:343] Scheduler found, paralleling scheduler...

  #431 opened 2 months ago by kazakovaanastasia
  0

• Cogvideo run

  #429 opened 2 months ago by kazakovaanastasia
  3

• How to Add DiTFastAttn in our custom model? do i need to write a pipeline or will it work if i can just swap out some attention layers/ processors

  #420 opened 3 months ago by asahni04
  4

• When height and width change, the inference speed will significantly slow down.

  #423 opened 3 months ago by serend1p1ty
  9

• Doubt about repeated joint_tensor_query in all_to_all

  #424 opened 3 months ago by Edenzzzz
  3

• Add Support for LTXV

  #421 opened 3 months ago by eranlevinlt
  0

• ComfyUI支持多卡并行的计划及合作方式

  #419 opened 3 months ago by joeshow79
  1

• flux-dev oom with 2gpus(each gpu is 24576MiB)

  #345 opened 4 months ago by algorithmconquer
  11

• CogVideo import error

  #417 opened 3 months ago by ZichengDuan
  1

• USP latency test

  #416 opened 3 months ago by eppane
  3

• How to use quantization methods to load models in a multi-machine environment

  #415 opened 3 months ago by etersin
  1

• FLUX.1-dev run in MultiNodes error

  #409 opened 3 months ago by etersin
  1

• Do you have plan to support SANA?

  #393 opened 3 months ago by xieenze
  2

• xDiT supports EasyAnimate

  #406 opened 3 months ago by paradiseHIT
  1

• Failed to load flux.1-dev with enable_sequential_cpu_offload and use_fp8_t5_encoder (4090)

  #407 opened 3 months ago by WeiboXu
  3

• Error occured when setting the environment

  #388 opened 3 months ago by TTTanger
  5

• Error when running Flux1.0 dev and HunyuanDiT-v1.2-Diffusers with multiple prompts

  #398 opened 3 months ago by henryhe4004
  3

• Flux.1 Hopper Performance

  #378 opened 4 months ago by thomasbtnfr
  3

• [feature] apply USP+FSDP to reduce memory overhead

  #400 opened 3 months ago by feifeibear
  0

• Reasons about inference speed difference with diffusers on single GPU without parrallelism?

  #392 opened 3 months ago by xyyan0123
  2

• Add support for ConsisID

  #389 opened 3 months ago by SHYuanBest
  1

• SD3 model error on onediff+pipefusion with official example

  #346 opened 4 months ago by wxsms
  2

• max_sequence_length will lead to error

  #380 opened 4 months ago by fy1214
  2

• Can comfyui-xdit run in multiple servers?

  #350 opened 4 months ago by VincentXWD
  4

• host.py get some error

  #373 opened 4 months ago by fy1214
  2

• Time compute

  #370 opened 4 months ago by thomasbtnfr
  4

• RuntimeError: CUDA error

  #374 opened 4 months ago by algorithmconquer
  3

• flux_example.py get stuck

  #363 opened 4 months ago by fy1214
  12

• Some questions about PipeFusion

  #360 opened 4 months ago by ictzyqq
  5

• flux-dev run error

  #354 opened 4 months ago by algorithmconquer
  4

• why cogvideo only support 4gpus dit parallel?

  #358 opened 4 months ago by Bensong0506
  3

• TypeError: FluxPipeline.__call__() got an unexpected keyword argument 'negative_prompt'

  #347 opened 4 months ago by ictzyqq
  2

• SD3-medium OOM on xDiT with 2x4090 but fine on diffusers with 1x4090

  #343 opened 4 months ago by wxsms
  4

• How about quantized models?

  #344 opened 4 months ago by wxsms
  0

• stable diffusion 3 在推理的时候，启用并行化会在卡0上启动多个进程，这个正常吗

  #340 opened 4 months ago by westnight
  1