ageitgey / ageitgey/face_recognition
Face detect using GPU parallelization
Open
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 56.8k
- Forks
- 13.7k
- PR merge metrics
- No merged PRs in 30d
Description
- face_recognition version: Latest
- Python version: 2.7
- Operating System: Ubuntu 16.04
Description
Hi.
When I extract faces on a video file, It takes a lot of time.
I use dlib and cuda, I think my system was not perfectly using GPU.
It only use 100% of CPUs and GPU-util is 1%.
How can I improve inference speed using GPU?
What I Did
Writing frame 2655 / 4820
Writing frame 2656 / 4820
Writing frame 2657 / 4820
Writing frame 2658 / 4820
Writing frame 2659 / 4820
Writing frame 2660 / 4820
CUDA version is 8.0, CUDNN version is 5.1.
Contributor guide
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 the reported face extraction workload and the dlib, CUDA 8.0, and cuDNN 5.1 environment details. Verify whether this video-processing path supports GPU inference and identify why CPU usage is high while GPU utilization remains at 1%; done means documenting or resolving the cause with a reproducible speed improvement.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- computer-vision, performance
- Issue type
- Bug
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 25/100