mit-han-lab/distrifuser
[CVPR 2024 Highlight] DistriFusion: Distributed Parallel Inference for High-Resolution Diffusion Models
PythonMIT
Issues
- 4
How cuda_graph to acclerate the denosing?
#22 opened by LHQUer - 2
There is an error, when I want to generate a 768*768 image with 8 GPU using sd2.0
#17 opened by dchiu0525 - 2
Will the whole mixture activation require OP/Fl(linear, convolution, or attention)?
#20 opened by LHQUer - 1
What is the difference between ‘full_sync’ mode and ‘sync_gn’ and ‘corrected_async_gn’?
#23 opened by ZijianZhouPro - 0
Support for Flux models
#21 opened by mali-afridi - 4
- 1
- 4
I have a question about running code. There is an error when running the command torchrun --nproc_per_node=2 scripts/sdxl_example.py. My torch version is 2.2.1, cuda version is 11.8, and python version is 3.10.
#13 opened by CharvinMei - 2
Chaos Picture
#16 opened by Qlalq - 9
DistriConv2dPP and other models
#14 opened by alivisionrd - 3
Odd number of GPUs
#15 opened by FlaviuRadulescu - 1
What's the best DUAL GPU for distrifuser ?
#12 opened by Glorias7 - 1
Failed to initialize process group: Error initializing torch.distributed using env:// rendezvous: environment variable RANK expected, but not set, falling back to single GPU
#11 opened by ibrahim737701 - 3
vscode debug launch.json file
#9 opened by Kyuseok-nam - 3
Measuring Communication Amount
#10 opened by tonyluozn - 3
- 3
- 2
- 1
Can distrifuser support SD v2-1 or v1-5?
#5 opened by czg1225 - 1
- 1
- 1
Does DistriFusion support ControlNet?
#2 opened by KeyaoZhao - 1
Some questions for Fig5.
#1 opened by yinfan98