Faster image pipelines · FluxML/FastAI.jl

(0 comments) (0 reactions) (0 assignees)Julia (49 forks)batch import

enhancementhelp wanted

Repository metrics

Some low-hanging fruit that FastAI.jl can pick to become faster here:

builtin presizing support
- prototype at https://github.com/lorenzoh/DataBlockStorage.jl/
- using JpegTurbo.jl's downsized loading so no storing to disk is needed (https://github.com/JuliaIO/JpegTurbo.jl/issues/15#issuecomment-1031548344)
use inplace operations
- DataAugmentation.jl already supports inplace operations, FastAI.jl just needs to add a buffered encode! interface and use those paths
- JpegTurbo.jl image loading could also be done inplace (https://github.com/JuliaIO/JpegTurbo.jl/issues/23)
- the whole pipeline could become non-allocating
use MLUtils.jl (#196)
- new DataLoader seems more stable and possibly a little faster
run projective transformations on GPU (https://github.com/lorenzoh/DataAugmentation.jl/issues/48)

Research direction: Focus on implementing presizing support using JpegTurbo.jl's downsized loading, following the prototype in DataBlockStorage.jl. Test the integration and benchmark the speed improvements.
Tech stack: julia
Domain: machine learning
Issue type: Performance
Difficulty: 2
Estimated time: Half day
Activity status: Fresh
Clarity: Clear
Prerequisites: JuliaGitFastAI.jl
Newbie friendliness: 70