Bug Fixes & Improvement

• Fix book guide training changes (#4340) @laggui
• Fix dequantize native debug statement (https://github.com/tracel-ai/cubek/pull/69) @laggui
• Do not point to pinned exact versions to allow pulling patch releases @laggui

Summary

This release marks a major turning point for the ecosystem with the introduction of CubeK. Our goal was to solve a classic challenge in deep learning: achieving peak performance on diverse hardware without maintaining fragmented codebases.

By unifying CPU and GPU kernels through CubeCL, we've managed...

What's Changed

• doc warning fix by @crutcher in https://github.com/tracel-ai/burn/pull/4130
• Fix tch bf16 into_data by @laggui in https://github.com/tracel-ai/burn/pull/4142
• Update raspberry-pi-pico example to use the Pico 2, and burnpack by @BjornTheProgrammer in https://github.com/tracel-ai/burn/pull/4132
• Unify all_reduce LocalCollectiveClient operation handling. by @crutcher in h...

What's Changed

• Bump version by @nathanielsimard in https://github.com/tracel-ai/burn/pull/4102
• Handle empty tensors in cat and slice_assign ops by @antimora in https://github.com/tracel-ai/burn/pull/4095
• Add network utilities to burn-std by @laggui in https://github.com/tracel-ai/burn/pull/4104
• Remove RefCell from onnx-ir Arguments by @antimora in https://github.com/tracel-ai/burn...

What's Changed

• Make TransformerEncoderLayer fields public by @Mnwa in https://github.com/tracel-ai/burn/pull/4053
• Feature muon by @NewBornRustacean in https://github.com/tracel-ai/burn/pull/3925
• Implement FromStr for Slice with parsing and error handling by @crutcher in https://github.com/tracel-ai/burn/pull/3983
• chore: Update to cubecl scalar refactor by @wingertge in https:/...