Mlx

- some modules are deprecated and folded into MLX (main module)
- MLXFast, MLXFFT, MLXRandom and MLXLinalg
- previous imports work, previous names work -- backward compatible
- see 205

What's Changed
* Remove package.resolved by louen in https://github.com/ml-explore/mlx-swift/pull/194
* add missing at functions by davidkoski in https://github.com/ml-explore/mlx-swift/pull/193
* add missing roll function by davidkoski in https://github.com/ml-explore/mlx-swift/pull/192
* Fix https://github.com/ml-explore/mlx-swift-examples/issues/218 by davidkoski in https://github.com/ml-explore/mlx-swift/pull/200
* address issues that prevent using composition for layers like LoRA by davidkoski in https://github.com/ml-explore/mlx-swift/pull/177
* Update mlx-c to 0.1.2 by barronalex in https://github.com/ml-explore/mlx-swift/pull/204
* Reorganize MLXFast, MLXFFT, MLXRandom and MLXLinalg by barronalex in https://github.com/ml-explore/mlx-swift/pull/205

New Contributors
* barronalex made their first contribution in https://github.com/ml-explore/mlx-swift/pull/204

**Full Changelog**: https://github.com/ml-explore/mlx-swift/compare/0.21.3...0.23.1

What's Changed

* Update check all set by LaurentMazare in https://github.com/ml-explore/mlx-swift/pull/174
* remove gguf -- fix 111 by davidkoski in https://github.com/ml-explore/mlx-swift/pull/176
* Add setErrorHandler and fatalErrorHandler by finnvoor in https://github.com/ml-explore/mlx-swift/pull/179
* fix: Correct wrong defatul value on docs of `SinusoidalPositionalEncoding` by minghuaw in https://github.com/ml-explore/mlx-swift/pull/182
* add dilation parameter to convolution layers (match mlx.nn) by davidkoski in https://github.com/ml-explore/mlx-swift/pull/183
* remove symlinks by davidkoski in https://github.com/ml-explore/mlx-swift/pull/191

New Contributors
* LaurentMazare made their first contribution in https://github.com/ml-explore/mlx-swift/pull/174
* finnvoor made their first contribution in https://github.com/ml-explore/mlx-swift/pull/179

**Full Changelog**: https://github.com/ml-explore/mlx-swift/compare/0.21.2...0.21.3

What's Changed
* fix 172 -- missing steel_attenion.metal by davidkoski in https://github.com/ml-explore/mlx-swift/pull/173


**Full Changelog**: https://github.com/ml-explore/mlx-swift/compare/0.21.1...0.21.2

What's Changed
* MLXArray (Data, Dtype) constructor , DType size. by louen in https://github.com/ml-explore/mlx-swift/pull/158

**Full Changelog**: https://github.com/ml-explore/mlx-swift/compare/0.21.0...0.21.1

What's Changed

* Add a shell script to replace the PrepareMetalHeaders plugin by louen in https://github.com/ml-explore/mlx-swift/pull/163
* make update method open by davidkoski in https://github.com/ml-explore/mlx-swift/pull/167
* fix: Fix wrong Wh shape in GRU and LSTM by minghuaw in https://github.com/ml-explore/mlx-swift/pull/168
* fix update of module tuple (seen for Qwen2VL) by davidkoski in https://github.com/ml-explore/mlx-swift/pull/164
* fix pow with scalar input by davidkoski in https://github.com/ml-explore/mlx-swift/pull/170
* fix: Fix wrong layer norm used in `TransformerDecoderLayer` by minghuaw in https://github.com/ml-explore/mlx-swift/pull/171
* adopt new mlx-c api, move to mlx v0.21.0 by davidkoski in https://github.com/ml-explore/mlx-swift/pull/150

New Contributors
* louen made their first contribution in https://github.com/ml-explore/mlx-swift/pull/163

**Full Changelog**: https://github.com/ml-explore/mlx-swift/compare/0.18.1...0.21.0

Highlights
- Even faster GEMMs
- Peaking at 23.89 TFlops on M2 Ultra [benchmarks](https://github.com/ml-explore/mlx/pull/1518)
- BFS graph optimizations
- Over 120tks with Mistral 7B!
- Fast batched QMV/QVM for KV quantized attention [benchmarks](https://github.com/ml-explore/mlx/pull/1564)

Core

- New Features
- `mx.linalg.eigh` and `mx.linalg.eigvalsh`
- `mx.nn.init.sparse`
- 64bit type support for `mx.cumprod`, `mx.cumsum`
- Performance
- Faster long column reductions
- Wired buffer support for large models
- Better Winograd dispatch condition for convs
- Faster scatter/gather
- Faster `mx.random.uniform` and `mx.random.bernoulli`
- Better threadgroup sizes for large arrays
- Misc
- Added Python 3.13 to CI
- C++20 compatibility

Bugfixes
- Fix command encoder synchronization
- Fix `mx.vmap` with gather and constant outputs
- Fix fused sdpa with differing key and value strides
- Support `mx.array.__format__` with spec
- Fix multi output array leak
- Fix RMSNorm weight mismatch error