modelscope / modelscope/DiffSynth-Studio
Hope to improve the dataloader of Wan
Open
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 13.1k
- Forks
- 1.3k
- Avg merge
- 13h 12m
- Merged PRs (30d)
- 45
Description
For every step, VideoDataset returns videos on the CPU and then transfers them to the GPU. This process is a bit slow and leads to fluctuating GPU utilization, as shown below:
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
Locate the VideoDataset implementation and trace where each batch moves from CPU to GPU. First reproduce the fluctuating GPU utilization described in the issue, then investigate the data-loading path and benchmark any proposed improvement. Done means the dataloader reduces the transfer bottleneck and produces more consistent GPU utilization without breaking video loading.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- data, machine-learning
- Issue type
- Feature
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 35/100