modelscope / modelscope/DiffSynth-Studio
FLUX.2-4B DMD2 训练需要多大显存？
Open
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 13.1k
- Forks
- 1.3k
- Avg merge
- 13h 12m
- Merged PRs (30d)
- 45
Description
尝试用1*8 80G H100 运行最近更新的 https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/flux2/model_training/special/dmd2/FLUX.2-klein-base-4B-DMD2.sh 代码，训练一个epoch就oom。想请教下大概需要多少计算资源呢？
Contributor guide
No contributing guide indexed for this repository
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start with examples/flux2/model_training/special/dmd2/FLUX.2-klein-base-4B-DMD2.sh and reproduce the one-epoch run on the reported 8×80 GB H100 setup. Inspect the training configuration and memory behavior to determine the required compute resources; done means providing a supported resource estimate or identifying a reproducible out-of-memory problem.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- machine-learning
- Issue type
- Bug
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Quiet
- Clarity
- Needs clarification
- Newbie friendliness
- 20/100