FluxML
Website / doc improvements to improve user experience
- [x] Update Features and Ecosystems - Remove some of the older ones
- [x] Update all the GPU links
- [x] Separate section: Showcase community projects that have the common models - GeometricFlux, Flux3D, MetalHead, ObjectDetector.jl, Transformers.jl, SciML (DiffEqFlux), TextAnalysis
- [x] Separate section: Showcase some more research-y things that are end applications - Yao, RayTracer, ...
- [ ] Deployment - with PackageCompiler. We know this won't work - but figure out where it breaks and file relevant issues.
- [x] Blog: Can we link to other blogs that are not written by us? Aggregation example: JuliaGPU.
- [ ] Have a table of all available models. Could be model zoo (where you cut and paste) or complete models that you can import (MetalHead). Should this be a table, where models are listed as rows, and then level of availability is checkboxes in columns?
- [ ] Observe some first time users, see where they fall, and make sure we have a quickstart or an FAQ to address common issues
- [x] Update partners. Add CMU, maybe others.
- [ ] Performance comparisons with PyTorch and TensorFlow. This should be front and center, and we should not be shy about admitting where work needs to be done. This will help address the perception that Flux is slow.
- [ ] How can people contribute - improvements, new models, new models with weights. Document that.
- [ ] Docs probably need to be organized for end users and for model builders. Model builders may need to have more information on GPU dev. Common gotchas (like scalar indexing), etc.
- [x] Link all the users of Flux packages: https://juliahub.com/ui/Packages/Flux/QdkVy/0.10.4?t=2 There's the laundry list, but perhaps some can also be classified into domains.
- [ ] What should users install? Torch.jl is optional - so publish blog and mention in docs, etc.
- [ ] Papers
- [ ] Add JuliaAcademy link
https://github.com/FluxML/Flux.jl/pull/1204 for updates to CUDA references in the docs.
Some progress on listing models and the ecosystem here (not yet linked, we just need to add the trained models table).
I just wanted to bring https://github.com/JuliaCommunity/ML-Coordination-Tracker to attention. We should reference this somewhere (e.g. a contributing section). There's also an MLH fellow helping us out with one of the community projects which should address at least two of the checkboxes above.
- Available models: the MLH fellow is working on FluxModels.jl which is an effort to provide a single point of access for standard models across various domains
- Performance comparison: once we have some models up and running in FluxModels.jl, we are planning on creating a repo similar to DiffEqBenchmarks. This will have performance comparisons against PyTorch and TF. And the Julia implementations can serve as regression tests if we integrate with @dhairyagandhi96 's bot.
We're planning on discussing the above plus other efforts in the next weekly call. But I thought I'd mention it here briefly since it is relevant to the website changes.
Great stuff. I imagine it could be useful to flesh out the 'community' section of the docs with resources like this, and of course specific projects like FluxModels can go on the list I posted above for any of those things. Would be great to have PRs to the right places.