issues
search
xdit-project
/
xDiT
xDiT: A Scalable Inference Engine for Diffusion Transformers (DiTs) on multi-GPU Clusters
Apache License 2.0
441
stars
36
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
[docker] service dockerfile
#258
feifeibear
closed
1 day ago
0
add a http service demo.
#257
feifeibear
closed
1 day ago
0
变量is_dp_last_group未定义
#256
WeYeYeYeYe
opened
1 day ago
0
latte模型的pos_embed没有被正确初始化
#255
WeYeYeYeYe
opened
1 day ago
1
ValueError: Attention Processor class CogVideoXAttnProcessor2_0 is not supported by xFuser
#254
WeYeYeYeYe
closed
1 day ago
4
运行CogVideoX-2b模型显示缺少bin文件
#253
WeYeYeYeYe
closed
1 day ago
4
Flux sp=2 errors
#252
feifeibear
closed
1 day ago
0
[Fix]: image saving bugs
#251
Eigensystem
closed
4 days ago
0
dockerfile
#250
feifeibear
closed
4 days ago
0
[Fix] dp issue
#249
Eigensystem
closed
4 days ago
0
Enable SP for CogVideoX (Low Performance Version)
#248
dannyxiaocn
opened
5 days ago
0
SD3 example failed when batch size is greater than 1 .
#247
wushixong
closed
4 days ago
1
polish the run.sh in example
#246
feifeibear
closed
5 days ago
0
[Feat] support xdit_parallel interface for huggingface
#245
Eigensystem
closed
5 days ago
0
Cogvideox SP Dev Minor Fix
#244
dannyxiaocn
closed
5 days ago
0
[doc] add latte, sd3 perf results
#243
feifeibear
closed
6 days ago
0
benchmark output_type default pil
#242
feifeibear
closed
1 week ago
0
v100 disable torch.compile for attention processor
#241
feifeibear
closed
1 week ago
0
attention processor unitest
#240
feifeibear
closed
1 week ago
0
add unitest for hunyuandit attn and refactor cache manager import
#239
feifeibear
closed
1 week ago
0
feat: support simple xdit comfyui-demo
#238
Lay2000
closed
4 days ago
1
[refactor] move cache_manager out of DiTRuntimeState
#237
feifeibear
closed
1 week ago
0
[doc] update flux performance results with onediff and torch.compile
#236
feifeibear
closed
1 week ago
0
[doc] add hunyuandit performance
#235
feifeibear
closed
1 week ago
0
upgrade yunchang version to 0.3.0 and flash_attn to 2.6.0
#234
feifeibear
closed
1 week ago
0
Revert "[AMDGPU] add amd gpu support"
#233
yiakwy-xpu-ml-framework-team
closed
1 week ago
0
polish readme
#232
feifeibear
closed
1 week ago
0
[doc] update readme with compilation accelerations
#231
feifeibear
closed
1 week ago
0
Flux Speed Up with PipeFusion
#230
mali-afridi
opened
1 week ago
2
[AMDGPU] add amd gpu support
#229
yiakwy-xpu-ml-framework-team
closed
1 week ago
1
rename one_diff to onediff
#228
strint
closed
1 week ago
1
apply onediff
#227
feifeibear
closed
1 week ago
0
RuntimeError: expected scalar type Float but found Half
#226
ptkang
closed
1 week ago
2
Why Latte Model can‘t support PipeFusion
#225
philipwan
opened
1 week ago
3
flux model torch compile works for sp on multiple GPUs
#224
feifeibear
closed
1 week ago
0
flux.1 warmup for torch.compile using the same same shape
#223
feifeibear
closed
2 weeks ago
0
[Docs] add discord link to README
#222
Eigensystem
closed
2 weeks ago
0
enable_sequential_cpu_offload on multi-gpu
#221
feifeibear
closed
2 weeks ago
0
Running with torch.compile on Flux with ulysses_degree=2 results in incorrect outcomes.
#220
feifeibear
closed
1 week ago
0
add enable_sequential_cpu_offload
#219
feifeibear
closed
2 weeks ago
0
RTX 4090, flux model, out of memory; approach is not compatible with quantization
#218
csdY123
closed
1 week ago
4
TypeError: _flash_attn_forward() missing 1 required positional argument: 'softcap'
#217
Pumbaa-peng
closed
1 week ago
2
[doc] fix typos in readme
#216
feifeibear
closed
2 weeks ago
0
[NFC] add flux performance
#215
feifeibear
closed
2 weeks ago
0
fix typo in license
#214
kuangdao
closed
2 weeks ago
0
RoadMap and Looking for Contributions
#213
feifeibear
opened
2 weeks ago
0
fix vae readme
#212
kuangdao
closed
2 weeks ago
0
Adding CogVideoX into xDiT
#211
dannyxiaocn
closed
1 week ago
0
[Feat] support parallel vae
#210
Eigensystem
closed
2 weeks ago
0
[WIP] Adding CogVideoX SP version into xDiT
#209
dannyxiaocn
closed
2 weeks ago
0
Next