modelscope/DiffSynth-Studio

View on GitHub

Enjoy the magic of Diffusion models!

Stars: 13.1k
Forks: 1.3k
Open beginner issues: 1
Indexed issues: 542
Avg merge: 13h 12m
Merged PRs (30d): 45

Dominant language: Python
License: Apache-2.0
Last GitHub push: Sep 20, 2026
Latest indexed: Sep 20, 2026
Contributing guide: No contributing guide
Code of conduct: No code of conduct
Beginner labels: No beginner labels indexed

Beginner friendly Everything

Open All

542 open issues indexed

Details on direct distillation Open

Difficulty 4/5 3-5 days Newbie friendliness 25/100

modelscope/DiffSynth-Studio#1181 · Opened Jan 8, 2026
LTX-2的支持 Open

Difficulty 5/5 Over a week Newbie friendliness 15/100

modelscope/DiffSynth-Studio#1180 · 4 reactions · Opened Jan 8, 2026
Is the bug of Skip download feature? Open

Difficulty 3/5 1-2 days Newbie friendliness 48/100

modelscope/DiffSynth-Studio#1179 · 1 comment · Opened Jan 8, 2026
flux2的lora权重无法在comfy上使用 Open

Difficulty 5/5 Over a week Newbie friendliness 25/100

modelscope/DiffSynth-Studio#1178 · 1 comment · Opened Jan 7, 2026
[BUG] Wan2.2 5B training demo failed in DiffSynth2.0 Open

Difficulty 4/5 3-5 days Newbie friendliness 35/100

modelscope/DiffSynth-Studio#1175 · 5 comments · Opened Jan 6, 2026
how to use acceleration lora provided by lightx2v in this framework? Open

Difficulty 5/5 Over a week Newbie friendliness 20/100

modelscope/DiffSynth-Studio#1174 · 1 comment · Opened Jan 6, 2026
Qwen-Image-Edit 多机多卡训练 Open

Difficulty 5/5 Over a week Newbie friendliness 20/100

modelscope/DiffSynth-Studio#1172 · 3 comments · Opened Jan 5, 2026
关于Qwen-image-edit训练时的图像缩放 Open

Difficulty 1/5 Under an hour Newbie friendliness 35/100

modelscope/DiffSynth-Studio#1168 · 9 comments · Opened Dec 30, 2025
ImportError: cannot import name 'DINOv3ViTModel' from 'transformers' Open

Difficulty 3/5 1-2 days Newbie friendliness 25/100

modelscope/DiffSynth-Studio#1163 · 1 comment · Opened Dec 29, 2025
Qwen edit 2511 training speed Open

Difficulty 5/5 Over a week Newbie friendliness 25/100

modelscope/DiffSynth-Studio#1162 · 17 comments · 2 reactions · Opened Dec 25, 2025
Question about generated image is blurry compared to the first frame Open

Difficulty 4/5 3-5 days Newbie friendliness 20/100

modelscope/DiffSynth-Studio#1161 · 1 comment · Opened Dec 25, 2025
如何实现/支持wan-animate 局部损失 Open

Difficulty 5/5 Over a week Newbie friendliness 20/100

modelscope/DiffSynth-Studio#1160 · Opened Dec 24, 2025
是否支持lightx2v的4步蒸馏模型推理呢？Wan2.2-I2V-4step-Distill的，直接加载的话出来的图像噪声很大。 Open

Difficulty 4/5 3-5 days Newbie friendliness 35/100

modelscope/DiffSynth-Studio#1159 · 1 comment · 4 reactions · Opened Dec 24, 2025
hello，能否支持对Infinitetalk的训练？ Open

Difficulty 5/5 Over a week Newbie friendliness 15/100

modelscope/DiffSynth-Studio#1158 · 1 comment · Opened Dec 24, 2025
accelerate_config_zero3_14B num_frames=81 Open

Difficulty 4/5 3-5 days Newbie friendliness 25/100

modelscope/DiffSynth-Studio#1157 · 1 comment · Opened Dec 23, 2025
No metadata in trained lora? no images validation? Open

Difficulty 4/5 3-5 days Newbie friendliness 35/100

modelscope/DiffSynth-Studio#1155 · 1 comment · Opened Dec 22, 2025
qwen-image-inference-error Open

Difficulty 3/5 1-2 days Newbie friendliness 35/100

modelscope/DiffSynth-Studio#1154 · 1 reaction · Opened Dec 22, 2025
The weight of Qwen/Qwen-Image-Edit-2511 Open

Difficulty 3/5 1-2 days Newbie friendliness 25/100

modelscope/DiffSynth-Studio#1151 · 1 comment · 3 reactions · Opened Dec 20, 2025
Trying to create tensor with negative dimension Open

Difficulty 3/5 1-2 days Newbie friendliness 35/100

modelscope/DiffSynth-Studio#1149 · Opened Dec 19, 2025
I find a BUG on loss_weight. Open

Difficulty 4/5 3-5 days Newbie friendliness 25/100

modelscope/DiffSynth-Studio#1148 · Opened Dec 19, 2025
diffsynth2.0 bug ImportError: cannot import name 'ModelManager' from 'diffsynth' Open

Difficulty 2/5 1-3 hours Newbie friendliness 35/100

modelscope/DiffSynth-Studio#1146 · Opened Dec 18, 2025
加载本地的权重加载不到 Open

Difficulty 3/5 1-2 days Newbie friendliness 25/100

modelscope/DiffSynth-Studio#1145 · 2 comments · Opened Dec 18, 2025
Wan2.2-Fun-5B-Control-Camera 模型加载失败 Open

Difficulty 3/5 1-2 days Newbie friendliness 45/100

modelscope/DiffSynth-Studio#1144 · 1 comment · Opened Dec 18, 2025
DiffSynth-Studio低显存示例代码中的vram_config参数与ModelConfig不兼容导致报错 Open

Difficulty 4/5 3-5 days Newbie friendliness 35/100

modelscope/DiffSynth-Studio#1142 · 2 comments · Opened Dec 17, 2025
How to align Qwen-Image Edit 2509 results between this repo and diffusers? Open

Difficulty 5/5 Over a week Newbie friendliness 25/100

modelscope/DiffSynth-Studio#1140 · 1 comment · Opened Dec 17, 2025
Qwen-image的llora微调支持拆分+Fp8吗？ Open

Difficulty 5/5 Over a week Newbie friendliness 25/100

modelscope/DiffSynth-Studio#1138 · Opened Dec 16, 2025
hello，会支持one to all animate、steady dance、SCAIL这些基于wan2.1模型的驱动跳舞的Lora训练嘛？ Open

Difficulty 5/5 Over a week Newbie friendliness 15/100

modelscope/DiffSynth-Studio#1135 · Opened Dec 16, 2025
z-image训练技巧提供的标准sft训练是指lora吗？ Open

Difficulty 3/5 1-2 days Newbie friendliness 30/100

modelscope/DiffSynth-Studio#1133 · Opened Dec 16, 2025
是否可以1min的转绘 Open

Difficulty 5/5 Over a week Newbie friendliness 20/100

modelscope/DiffSynth-Studio#1132 · Opened Dec 15, 2025
在跑qwenimage的推理的时候遇到问题 Open

Difficulty 4/5 3-5 days Newbie friendliness 35/100

modelscope/DiffSynth-Studio#1131 · 1 comment · Opened Dec 15, 2025
Support ND-Parallel training？ Open

Difficulty 5/5 Over a week Newbie friendliness 20/100

modelscope/DiffSynth-Studio#1130 · Opened Dec 15, 2025
关于wan的camera control实现 Open

Difficulty 5/5 Over a week Newbie friendliness 20/100

modelscope/DiffSynth-Studio#1127 · 2 comments · Opened Dec 15, 2025
z-image cotrolnet train and infer Open

Difficulty 5/5 Over a week Newbie friendliness 20/100

modelscope/DiffSynth-Studio#1125 · Opened Dec 12, 2025
wan多机多卡启动命令 Open

Difficulty 3/5 1-2 days Newbie friendliness 25/100

modelscope/DiffSynth-Studio#1123 · 1 comment · Opened Dec 12, 2025
服了连个路径都搞不明白 Open

Difficulty 3/5 1-2 days Newbie friendliness 35/100

modelscope/DiffSynth-Studio#1121 · Opened Dec 12, 2025
Z-Image训练的lora，comfy中无法正常推理 Open

Difficulty 4/5 3-5 days Newbie friendliness 25/100

modelscope/DiffSynth-Studio#1120 · 2 comments · Opened Dec 12, 2025
batch output Open

Difficulty 3/5 1-2 days Newbie friendliness 35/100

modelscope/DiffSynth-Studio#1119 · Opened Dec 12, 2025
请问为什么训练出来的flux2的Lora在comfyui里不生效？ Open

Difficulty 4/5 3-5 days Newbie friendliness 20/100

modelscope/DiffSynth-Studio#1117 · 11 comments · Opened Dec 10, 2025
请问，如何加载Z-image的Fp8模型？或者直接使用fp8推理？ Open

Difficulty 5/5 Over a week Newbie friendliness 25/100

modelscope/DiffSynth-Studio#1116 · 1 reaction · Opened Dec 10, 2025
Qwen-Image-i2L使用的时候建议使用几张图片？ Open

Difficulty 1/5 Under an hour Newbie friendliness 25/100

modelscope/DiffSynth-Studio#1115 · 2 reactions · Opened Dec 10, 2025
昇腾ComfyUI不适配了吗？ Open

Difficulty 5/5 Over a week Newbie friendliness 20/100

modelscope/DiffSynth-Studio#1114 · Opened Dec 10, 2025
wan训练时vae encode错误 Open

Difficulty 4/5 3-5 days Newbie friendliness 35/100

modelscope/DiffSynth-Studio#1113 · 2 comments · Opened Dec 10, 2025
Does Flux.2 support LORA training for single-image editing? Open

Difficulty 5/5 Over a week Newbie friendliness 25/100

modelscope/DiffSynth-Studio#1112 · Opened Dec 10, 2025
DiffSynth后续是否有针对文生图或者编辑模型加入强化学习训练思路的计划 Open

Difficulty 5/5 Over a week Newbie friendliness 20/100

modelscope/DiffSynth-Studio#1111 · 9 reactions · Opened Dec 10, 2025
DiffSynth 是否有计划为 Wan-Series 视频生成模型添加强化学习 (RL) 训练支持？ Open

Difficulty 5/5 Over a week Newbie friendliness 20/100

modelscope/DiffSynth-Studio#1110 · 1 comment · 4 reactions · Opened Dec 9, 2025
kontext全参数训练进行sft:data_process处理时报错 Open

Difficulty 4/5 3-5 days Newbie friendliness 35/100

modelscope/DiffSynth-Studio#1107 · 2 comments · Opened Dec 9, 2025
support of LongCat-Image-Edit？ Open

Difficulty 5/5 Over a week Newbie friendliness 20/100

modelscope/DiffSynth-Studio#1106 · 1 comment · 1 reaction · Opened Dec 9, 2025
How to train a 4-step or 8-step LoRA for the Qwen-Image-Edit model？ Open

Difficulty 5/5 Over a week Newbie friendliness 20/100

modelscope/DiffSynth-Studio#1103 · 3 comments · Opened Dec 8, 2025
分享，支持int8-quanto量化微调Qwen-Image-Edit-2509,48G显存即可微调🎉 Open

Difficulty 4/5 3-5 days Newbie friendliness 20/100

modelscope/DiffSynth-Studio#1102 · 2 comments · Opened Dec 7, 2025
Question about qwen-image-edit training (resize) Open

Difficulty 4/5 3-5 days Newbie friendliness 25/100

modelscope/DiffSynth-Studio#1099 · 1 reaction · Opened Dec 5, 2025

modelscope/DiffSynth-Studio

Get new issues in your inbox