tracel-ai
burn
Burn is a next generation tensor library and Deep Learning Framework that doesn't compromise on flexibility, efficiency and portability.
Recommended issues
Issues stay on GitHub. Read the thread before starting work.
-
#2924 autodiff: Gradients::get panics on type mismatch instead of returning an error
buggood first issue
-
#544 train: add GPU utilization and memory metrics for non-NVIDIA GPUs
good first issuefeature
-
#282 docs: provide a cargo-generate starter template
help wantedgood first issueexampleaccessibility
-
#810 ci: add WebGPU integration tests (e.g. with Deno)
help wantedwgpuinfrastructure
-
#3571 train: hot reloading for faster prototyping
enhancementhelp wanted
Topics
autodiffdeep-learningmachine-learningrustscientific-computingndarraytensorneural-networkpytorchcross-platformkernel-fusiononnxwasmwebgpucudametalrocmvulkan