GPU acceleration for palletizattion
Open
feature request
- Dominant language
- Python
- Stars
- 5.4k
- Forks
- 850
- Avg merge
- 4d 5h
- Merged PRs (30d)
- 10
Description
would help a lot to speed up compression. my current plan is to use nvidia gpu since they have more compute, but support for apple silicon gpus would probably be useful for more people.
also i know it is currently possible to compress weights externally, but i think it is more user friendly to integrate in the cto api
Contributor guide
Research direction
The issue names no files, tests, or specific entry points; start by locating the model compression or weight palletization APIs in coremltools and the CTO API mentioned in the request. Define and validate GPU-backed compression for NVIDIA GPUs and Apple Silicon GPUs, with tests showing the supported workflows and expected results.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- machine-learning, performance
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 25/100