v0.15.2

🚀

v0.15.1

🚀

v0.15.0

Highlights

• Fast Metal GPU FFTs
  • On average ~30x faster than CPU
  • More benchmarks
• mx.distributed with all_sum and all_gather

Core

• Added dlpack device __dlpack_device__
• Fast GPU FFTs benchmarks
• Add docs for the mx.distributed
• Add mx.view op

NN

• softmin, hardshrink, and hardtanh activations

Bugfixes

• Fix broadcast bug in bitwise ops
• Allow more buffers for JIT compilation
• Fix matvec vector stride bug
• Fix multi-block sort stride management
• Stable cumprod grad at 0
• Buf fix with race condition in scan

v0.14.1

🚀

v0.14.0

Highlights

• Small-size build that JIT compiles kernels and omits the CPU backend which results in a binary <4MB
  • Series of PRs 1, 2, 3, 4, 5
• mx.gather_qmm quantized equivalent for mx.gather_mm which speeds up MoE inference by ~2x
  • Some numbers
• Grouped 2D convolutions
  • Some numbers

Core

• mx.conjugate
• mx.conv3d and nn.Conv3d
• List based indexing
• Started mx.distributed which uses MPI (if installed) for communication across machines
  • mx.distributed.init
  • mx.distributed.all_gather
  • mx.distributed.all_reduce_sum
• Support conversion to and from dlpack
• mx.linalg.cholesky on CPU
• mx.quantized_matmul sped up for vector-matrix products
• mx.trace
• mx.block_masked_mm now supports floating point masks!

Fixes

• Error messaging in eval
• Add some missing docs
• Scatter index bug
• The extensions example now compiles and runs
• CPU copy bug with many dimensions

v0.13.1

🚀

v0.13.0

Highlights

• Block sparse matrix multiply speeds up MoEs by >2x
  • some numbers
• Improved quantization algorithm should work well for all networks
  • see evaluations
• Improved gpu command submission speeds up training and inference
  • some numbers

Core

• Bitwise ops added:
  • mx.bitwise_[or|and|xor], mx.[left|right]_shift, operator overloads
• Groups added to Conv1d
• Added mx.metal.device_info to get better informed memory limits
• Added resettable memory stats
• mlx.optimizers.clip_grad_norm and mlx.utils.tree_reduce added
• Add mx.arctan2
• Unary ops now accept array-like inputs ie one can do mx.sqrt(2)

Bugfixes

• Fixed shape for slice update
• Bugfix in quantize that used slightly wrong scales/biases
• Fixed memory leak for multi-output primitives encountered with gradient checkpointing
• Fixed conversion from other frameworks for all datatypes
• Fixed index overflow for matmul with large batch size
• Fixed initialization ordering that occasionally caused segfaults

v0.12.2

Patch bump (#1067)

* version

* use 0.12.2

v0.12.0

Highlights

• Faster quantized matmul
  • Up to 40% faster QLoRA or prompt processing, some numbers

Core

• mx.synchronize to wait for computation dispatched with mx.async_eval
• mx.radians and mx.degrees
• mx.metal.clear_cache to return to the OS the memory held by MLX as a cache for future allocations
• Change quantization to always represent 0 exactly (relevant issue)

Bugfixes

• Fixed quantization of a block with all 0s that produced NaNs
• Fixed the len field in the buffer protocol implementation

v0.11.0

Core

• mx.block_masked_mm for block-level sparse matrix multiplication
• Shared events for synchronization and asynchronous evaluation

NN

• nn.QuantizedEmbedding layer
• nn.quantize for quantizing modules
• gelu_approx uses tanh for consistency with PyTorch