| CMakeLists.txt | |
| LICENSE | |
| MANIFEST.in | |
| README.md | |
| mlx.pc.in | |
| pyproject.toml | |
| setup.py | |
| cmake/FindNCCL.cmake | |
| cmake/extension.cmake | |
| mlx/CMakeLists.txt | |
| mlx/allocator.cpp | |
| mlx/allocator.h | |
| mlx/array.cpp | |
| mlx/array.h | |
| mlx/compile.cpp | |
| mlx/compile.h | |
| mlx/compile_impl.h | |
| mlx/device.cpp | |
| mlx/device.h | |
| mlx/dtype.cpp | |
| mlx/dtype.h | |
| mlx/dtype_utils.cpp | |
| mlx/dtype_utils.h | |
| mlx/einsum.cpp | |
| mlx/einsum.h | |
| mlx/event.h | |
| mlx/export.cpp | |
| mlx/export.h | |
| mlx/export_impl.h | |
| mlx/fast.cpp | |
| mlx/fast.h | |
| mlx/fast_primitives.h | |
| mlx/fence.h | |
| mlx/fft.cpp | |
| mlx/fft.h | |
| mlx/graph_utils.cpp | |
| mlx/graph_utils.h | |
| mlx/io.h | |
| mlx/linalg.cpp | |
| mlx/linalg.h | |
| mlx/memory.h | |
| mlx/mlx.h | |
| mlx/ops.cpp | |
| mlx/ops.h | |
| mlx/primitives.cpp | |
| mlx/primitives.h | |
| mlx/random.cpp | |
| mlx/random.h | |
| mlx/scheduler.cpp | |
| mlx/scheduler.h | |
| mlx/small_vector.h | |
| mlx/stream.h | |
| mlx/threadpool.h | |
| mlx/transforms.cpp | |
| mlx/transforms.h | |
| mlx/transforms_impl.h | |
| mlx/utils.cpp | |
| mlx/utils.h | |
| mlx/version.cpp | |
| mlx/version.h | |
| mlx/3rdparty/.clang-format | |
| mlx/3rdparty/pocketfft.h | |
| mlx/backend/common/CMakeLists.txt | |
| mlx/backend/common/binary.h | |
| mlx/backend/common/broadcasting.cpp | |
| mlx/backend/common/broadcasting.h | |
| mlx/backend/common/buffer_cache.h | |
| mlx/backend/common/common.cpp | |
| mlx/backend/common/compiled.cpp | |
| mlx/backend/common/compiled.h | |
| mlx/backend/common/copy.h | |
| mlx/backend/common/hadamard.h | |
| mlx/backend/common/load.cpp | |
| mlx/backend/common/matmul.h | |
| mlx/backend/common/reduce.cpp | |
| mlx/backend/common/reduce.h | |
| mlx/backend/common/slicing.cpp | |
| mlx/backend/common/slicing.h | |
| mlx/backend/common/ternary.h | |
| mlx/backend/common/unary.h | |
| mlx/backend/common/utils.cpp | |
| mlx/backend/common/utils.h | |
| mlx/backend/cpu/CMakeLists.txt | |
| mlx/backend/cpu/arange.h | |
| mlx/backend/cpu/arg_reduce.cpp | |
| mlx/backend/cpu/available.cpp | |
| mlx/backend/cpu/available.h | |
| mlx/backend/cpu/binary.cpp | |
| mlx/backend/cpu/binary.h | |
| mlx/backend/cpu/binary_ops.h | |
| mlx/backend/cpu/binary_two.h | |
| mlx/backend/cpu/cholesky.cpp | |
| mlx/backend/cpu/compiled.cpp | |
| mlx/backend/cpu/compiled_preamble.h | |
| mlx/backend/cpu/conv.cpp | |
| mlx/backend/cpu/copy.cpp | |
| mlx/backend/cpu/copy.h | |
| mlx/backend/cpu/distributed.cpp | |
| mlx/backend/cpu/eig.cpp | |
| mlx/backend/cpu/eigh.cpp | |
| mlx/backend/cpu/encoder.cpp | |
| mlx/backend/cpu/encoder.h | |
| mlx/backend/cpu/eval.cpp | |
| mlx/backend/cpu/eval.h | |
| mlx/backend/cpu/fft.cpp | |
| mlx/backend/cpu/gemm.h | |
| mlx/backend/cpu/hadamard.cpp | |
| mlx/backend/cpu/indexing.cpp | |
| mlx/backend/cpu/inverse.cpp | |
| mlx/backend/cpu/jit_compiler.cpp | |
| mlx/backend/cpu/jit_compiler.h | |
| mlx/backend/cpu/lapack.h | |
| mlx/backend/cpu/logsumexp.cpp | |
| mlx/backend/cpu/luf.cpp | |
| mlx/backend/cpu/make_compiled_preamble.ps1 | |
| mlx/backend/cpu/make_compiled_preamble.sh | |
| mlx/backend/cpu/masked_mm.cpp | |
| mlx/backend/cpu/matmul.cpp | |
| mlx/backend/cpu/primitives.cpp | |
| mlx/backend/cpu/qrf.cpp | |
| mlx/backend/cpu/quantized.cpp | |
| mlx/backend/cpu/reduce.cpp | |
| mlx/backend/cpu/scan.cpp | |
| mlx/backend/cpu/select.cpp | |
| mlx/backend/cpu/slicing.h | |
| mlx/backend/cpu/softmax.cpp | |
| mlx/backend/cpu/sort.cpp | |
| mlx/backend/cpu/svd.cpp | |
| mlx/backend/cpu/ternary.h | |
| mlx/backend/cpu/threefry.cpp | |
| mlx/backend/cpu/threefry.h | |
| mlx/backend/cpu/unary.cpp | |
| mlx/backend/cpu/unary.h | |
| mlx/backend/cpu/unary_ops.h | |
| mlx/backend/cpu/gemms/bnns.cpp | |
| mlx/backend/cpu/gemms/cblas.cpp | |
| mlx/backend/cpu/gemms/simd_bf16.cpp | |
| mlx/backend/cpu/gemms/simd_fp16.cpp | |
| mlx/backend/cpu/gemms/simd_gemm.h | |
| mlx/backend/cpu/simd/accelerate_fp16_simd.h | |
| mlx/backend/cpu/simd/accelerate_simd.h | |
| mlx/backend/cpu/simd/base_simd.h | |
| mlx/backend/cpu/simd/math.h | |
| mlx/backend/cpu/simd/neon_fp16_simd.h | |
| mlx/backend/cpu/simd/simd.h | |
| mlx/backend/cpu/simd/type.h | |
| mlx/backend/cuda/CMakeLists.txt | |
| mlx/backend/cuda/allocator.cpp | |
| mlx/backend/cuda/allocator.h | |
| mlx/backend/cuda/arange.cu | |
| mlx/backend/cuda/arg_reduce.cu | |
| mlx/backend/cuda/bin2h.cmake | |
| mlx/backend/cuda/binary_two.cu | |
| mlx/backend/cuda/compiled.cpp | |
| mlx/backend/cuda/conv.cpp | |
| mlx/backend/cuda/copy.cu | |
| mlx/backend/cuda/cuda.cpp | |
| mlx/backend/cuda/cuda.h | |
| mlx/backend/cuda/cudnn_utils.cpp | |
| mlx/backend/cuda/cudnn_utils.h | |
| mlx/backend/cuda/custom_kernel.cpp | |
| mlx/backend/cuda/device.cpp | |
| mlx/backend/cuda/device.h | |
| mlx/backend/cuda/distributed.cu | |
| mlx/backend/cuda/eval.cpp | |
| mlx/backend/cuda/event.cu | |
| mlx/backend/cuda/event.h | |
| mlx/backend/cuda/fence.cpp | |
| mlx/backend/cuda/indexing.cpp | |
| mlx/backend/cuda/jit_module.cpp | |
| mlx/backend/cuda/jit_module.h | |
| mlx/backend/cuda/kernel_utils.cu | |
| mlx/backend/cuda/kernel_utils.cuh | |
| mlx/backend/cuda/layer_norm.cu | |
| mlx/backend/cuda/logsumexp.cu | |
| mlx/backend/cuda/lru_cache.h | |
| mlx/backend/cuda/matmul.cpp | |
| mlx/backend/cuda/no_cuda.cpp | |
| mlx/backend/cuda/primitives.cpp | |
| mlx/backend/cuda/random.cu | |
| mlx/backend/cuda/reduce.cu | |
| mlx/backend/cuda/rms_norm.cu | |
| mlx/backend/cuda/rope.cu | |
| mlx/backend/cuda/scaled_dot_product_attention.cu | |
| mlx/backend/cuda/scan.cu | |
| mlx/backend/cuda/slicing.cpp | |
| mlx/backend/cuda/softmax.cu | |
| mlx/backend/cuda/sort.cu | |
| mlx/backend/cuda/ternary.cu | |
| mlx/backend/cuda/unary.cu | |
| mlx/backend/cuda/utils.cpp | |
| mlx/backend/cuda/utils.h | |
| mlx/backend/cuda/worker.cpp | |
| mlx/backend/cuda/worker.h | |
| mlx/backend/cuda/binary/CMakeLists.txt | |
| mlx/backend/cuda/binary/add.cu | |
| mlx/backend/cuda/binary/arctan2.cu | |
| mlx/backend/cuda/binary/binary.cuh | |
| mlx/backend/cuda/binary/bitwise_binary.cu | |
| mlx/backend/cuda/binary/divide.cu | |
| mlx/backend/cuda/binary/equal.cu | |
| mlx/backend/cuda/binary/greater.cu | |
| mlx/backend/cuda/binary/greater_equal.cu | |
| mlx/backend/cuda/binary/less.cu | |
| mlx/backend/cuda/binary/less_equal.cu | |
| mlx/backend/cuda/binary/log_add_exp.cu | |
| mlx/backend/cuda/binary/logical_and.cu | |
| mlx/backend/cuda/binary/logical_or.cu | |
| mlx/backend/cuda/binary/maximum.cu | |
| mlx/backend/cuda/binary/minimum.cu | |
| mlx/backend/cuda/binary/multiply.cu | |
| mlx/backend/cuda/binary/not_equal.cu | |
| mlx/backend/cuda/binary/power.cu | |
| mlx/backend/cuda/binary/remainder.cu | |
| mlx/backend/cuda/binary/subtract.cu | |
| mlx/backend/cuda/conv/conv.h | |
| mlx/backend/cuda/conv/gemm_conv.cu | |
| mlx/backend/cuda/conv/gemm_grouped_conv.cu | |
| mlx/backend/cuda/copy/copy.cuh | |
| mlx/backend/cuda/copy/copy_contiguous.cu | |
| mlx/backend/cuda/copy/copy_general.cu | |
| mlx/backend/cuda/copy/copy_general_dynamic.cu | |
| mlx/backend/cuda/copy/copy_general_input.cu | |
| mlx/backend/cuda/device/atomic_ops.cuh | |
| mlx/backend/cuda/device/binary_ops.cuh | |
| mlx/backend/cuda/device/cast_op.cuh | |
| mlx/backend/cuda/device/complex.cuh | |
| mlx/backend/cuda/device/config.h | |
| mlx/backend/cuda/device/fp16_math.cuh | |
| mlx/backend/cuda/device/gather.cuh | |
| mlx/backend/cuda/device/gather_axis.cuh | |
| mlx/backend/cuda/device/indexing.cuh | |
| mlx/backend/cuda/device/scatter.cuh | |
| mlx/backend/cuda/device/scatter_axis.cuh | |
| mlx/backend/cuda/device/scatter_ops.cuh | |
| mlx/backend/cuda/device/ternary_ops.cuh | |
| mlx/backend/cuda/device/unary_ops.cuh | |
| mlx/backend/cuda/device/utils.cuh | |
| mlx/backend/cuda/gemms/cublas_gemm.cpp | |
| mlx/backend/cuda/gemms/cublas_gemm.h | |
| mlx/backend/cuda/gemms/cublas_gemm_batched_12_0.cpp | |
| mlx/backend/cuda/gemms/cublas_gemm_batched_12_9.cu | |
| mlx/backend/cuda/gemms/gemv.cu | |
| mlx/backend/cuda/gemms/gemv.h | |
| mlx/backend/cuda/quantized/affine_quantize.cu | |
| mlx/backend/cuda/quantized/quantized.cpp | |
| mlx/backend/cuda/quantized/quantized.h | |
| mlx/backend/cuda/quantized/quantized_utils.cuh | |
| mlx/backend/cuda/reduce/all_reduce.cu | |
| mlx/backend/cuda/reduce/col_reduce.cu | |
| mlx/backend/cuda/reduce/init_reduce.cu | |
| mlx/backend/cuda/reduce/reduce.cuh | |
| mlx/backend/cuda/reduce/reduce_ops.cuh | |
| mlx/backend/cuda/reduce/reduce_utils.cuh | |
| mlx/backend/cuda/reduce/row_reduce.cu | |
| mlx/backend/cuda/steel/defines.cuh | |
| mlx/backend/cuda/steel/gemm.cuh | |
| mlx/backend/cuda/steel/mma.cuh | |
| mlx/backend/cuda/steel/tiles.cuh | |
| mlx/backend/cuda/steel/utils.cuh | |
| mlx/backend/cuda/unary/CMakeLists.txt | |
| mlx/backend/cuda/unary/abs.cu | |
| mlx/backend/cuda/unary/arccos.cu | |
| mlx/backend/cuda/unary/arccosh.cu | |
| mlx/backend/cuda/unary/arcsin.cu | |
| mlx/backend/cuda/unary/arcsinh.cu | |
| mlx/backend/cuda/unary/arctan.cu | |
| mlx/backend/cuda/unary/arctanh.cu | |
| mlx/backend/cuda/unary/bitwise_invert.cu | |
| mlx/backend/cuda/unary/ceil.cu | |
| mlx/backend/cuda/unary/conjugate.cu | |
| mlx/backend/cuda/unary/cos.cu | |
| mlx/backend/cuda/unary/cosh.cu | |
| mlx/backend/cuda/unary/erf.cu | |
| mlx/backend/cuda/unary/erf_inv.cu | |
| mlx/backend/cuda/unary/exp.cu | |
| mlx/backend/cuda/unary/expm1.cu | |
| mlx/backend/cuda/unary/floor.cu | |
| mlx/backend/cuda/unary/imag.cu | |
| mlx/backend/cuda/unary/log.cu | |
| mlx/backend/cuda/unary/log1p.cu | |
| mlx/backend/cuda/unary/logical_not.cu | |
| mlx/backend/cuda/unary/negative.cu | |
| mlx/backend/cuda/unary/real.cu | |
| mlx/backend/cuda/unary/round.cu | |
| mlx/backend/cuda/unary/sigmoid.cu | |
| mlx/backend/cuda/unary/sign.cu | |
| mlx/backend/cuda/unary/sin.cu | |
| mlx/backend/cuda/unary/sinh.cu | |
| mlx/backend/cuda/unary/sqrt.cu | |
| mlx/backend/cuda/unary/square.cu | |
| mlx/backend/cuda/unary/tan.cu | |
| mlx/backend/cuda/unary/tanh.cu | |
| mlx/backend/cuda/unary/unary.cuh | |
| mlx/backend/gpu/CMakeLists.txt | |
| mlx/backend/gpu/available.h | |
| mlx/backend/gpu/copy.cpp | |
| mlx/backend/gpu/copy.h | |
| mlx/backend/gpu/eval.h | |
| mlx/backend/gpu/primitives.cpp | |
| mlx/backend/gpu/slicing.cpp | |
| mlx/backend/gpu/slicing.h | |
| mlx/backend/metal/CMakeLists.txt | |
| mlx/backend/metal/allocator.cpp | |
| mlx/backend/metal/allocator.h | |
| mlx/backend/metal/binary.cpp | |
| mlx/backend/metal/binary.h | |
| mlx/backend/metal/compiled.cpp | |
| mlx/backend/metal/conv.cpp | |
| mlx/backend/metal/copy.cpp | |
| mlx/backend/metal/custom_kernel.cpp | |
| mlx/backend/metal/device.cpp | |
| mlx/backend/metal/device.h | |
| mlx/backend/metal/distributed.cpp | |
| mlx/backend/metal/eval.cpp | |
| mlx/backend/metal/event.cpp | |
| mlx/backend/metal/fence.cpp | |
| mlx/backend/metal/fft.cpp | |
| mlx/backend/metal/hadamard.cpp | |
| mlx/backend/metal/indexing.cpp | |
| mlx/backend/metal/jit_kernels.cpp | |
| mlx/backend/metal/kernels.h | |
| mlx/backend/metal/logsumexp.cpp | |
| mlx/backend/metal/make_compiled_preamble.sh | |
| mlx/backend/metal/matmul.cpp | |
| mlx/backend/metal/matmul.h | |
| mlx/backend/metal/metal.cpp | |
| mlx/backend/metal/metal.h | |
| mlx/backend/metal/no_metal.cpp | |
| mlx/backend/metal/nojit_kernels.cpp | |
| mlx/backend/metal/normalization.cpp | |
| mlx/backend/metal/primitives.cpp | |
| mlx/backend/metal/quantized.cpp | |
| mlx/backend/metal/reduce.cpp | |
| mlx/backend/metal/reduce.h | |
| mlx/backend/metal/resident.cpp | |
| mlx/backend/metal/resident.h | |
| mlx/backend/metal/rope.cpp | |
| mlx/backend/metal/scaled_dot_product_attention.cpp | |
| mlx/backend/metal/scan.cpp | |
| mlx/backend/metal/slicing.cpp | |
| mlx/backend/metal/softmax.cpp | |
| mlx/backend/metal/sort.cpp | |
| mlx/backend/metal/ternary.cpp | |
| mlx/backend/metal/ternary.h | |
| mlx/backend/metal/unary.cpp | |
| mlx/backend/metal/unary.h | |
| mlx/backend/metal/utils.cpp | |
| mlx/backend/metal/utils.h | |
| mlx/backend/metal/jit/includes.h | |
| mlx/backend/metal/jit/indexing.h | |
| mlx/backend/metal/kernels/CMakeLists.txt | |
| mlx/backend/metal/kernels/arange.h | |
| mlx/backend/metal/kernels/arange.metal | |
| mlx/backend/metal/kernels/arg_reduce.metal | |
| mlx/backend/metal/kernels/atomic.h | |
| mlx/backend/metal/kernels/bf16_math.h | |
| mlx/backend/metal/kernels/binary.h | |
| mlx/backend/metal/kernels/binary.metal | |
| mlx/backend/metal/kernels/binary_ops.h | |
| mlx/backend/metal/kernels/binary_two.h | |
| mlx/backend/metal/kernels/binary_two.metal | |
| mlx/backend/metal/kernels/cexpf.h | |
| mlx/backend/metal/kernels/complex.h | |
| mlx/backend/metal/kernels/conv.metal | |
| mlx/backend/metal/kernels/copy.h | |
| mlx/backend/metal/kernels/copy.metal | |
| mlx/backend/metal/kernels/defines.h | |
| mlx/backend/metal/kernels/erf.h | |
| mlx/backend/metal/kernels/expm1f.h | |
| mlx/backend/metal/kernels/fence.metal | |
| mlx/backend/metal/kernels/fft.h | |
| mlx/backend/metal/kernels/fft.metal | |
| mlx/backend/metal/kernels/fp4_quantized.h | |
| mlx/backend/metal/kernels/fp4_quantized.metal | |
| mlx/backend/metal/kernels/gemv.metal | |
| mlx/backend/metal/kernels/gemv_masked.h | |
| mlx/backend/metal/kernels/gemv_masked.metal | |
| mlx/backend/metal/kernels/hadamard.h | |
| mlx/backend/metal/kernels/layer_norm.metal | |
| mlx/backend/metal/kernels/logsumexp.h | |
| mlx/backend/metal/kernels/logsumexp.metal | |
| mlx/backend/metal/kernels/quantized.h | |
| mlx/backend/metal/kernels/quantized.metal | |
| mlx/backend/metal/kernels/quantized_utils.h | |
| mlx/backend/metal/kernels/random.metal | |
| mlx/backend/metal/kernels/reduce.h | |
| mlx/backend/metal/kernels/reduce.metal | |
| mlx/backend/metal/kernels/reduce_utils.h | |
| mlx/backend/metal/kernels/rms_norm.metal | |
| mlx/backend/metal/kernels/rope.metal | |
| mlx/backend/metal/kernels/scaled_dot_product_attention.metal | |
| mlx/backend/metal/kernels/scan.h | |
| mlx/backend/metal/kernels/scan.metal | |
| mlx/backend/metal/kernels/sdpa_vector.h | |
| mlx/backend/metal/kernels/softmax.h | |
| mlx/backend/metal/kernels/softmax.metal | |
| mlx/backend/metal/kernels/sort.h | |
| mlx/backend/metal/kernels/sort.metal | |
| mlx/backend/metal/kernels/ternary.h | |
| mlx/backend/metal/kernels/ternary.metal | |
| mlx/backend/metal/kernels/ternary_ops.h | |
| mlx/backend/metal/kernels/unary.h | |
| mlx/backend/metal/kernels/unary.metal | |
| mlx/backend/metal/kernels/unary_ops.h | |
| mlx/backend/metal/kernels/utils.h | |
| mlx/backend/metal/kernels/fft/radix.h | |
| mlx/backend/metal/kernels/fft/readwrite.h | |
| mlx/backend/metal/kernels/indexing/gather.h | |
| mlx/backend/metal/kernels/indexing/gather_axis.h | |
| mlx/backend/metal/kernels/indexing/gather_front.h | |
| mlx/backend/metal/kernels/indexing/indexing.h | |
| mlx/backend/metal/kernels/indexing/scatter.h | |
| mlx/backend/metal/kernels/indexing/scatter_axis.h | |
| mlx/backend/metal/kernels/jit/bf16.h | |
| mlx/backend/metal/kernels/metal_3_0/bf16.h | |
| mlx/backend/metal/kernels/metal_3_1/bf16.h | |
| mlx/backend/metal/kernels/reduction/ops.h | |
| mlx/backend/metal/kernels/reduction/reduce_all.h | |
| mlx/backend/metal/kernels/reduction/reduce_col.h | |
| mlx/backend/metal/kernels/reduction/reduce_init.h | |
| mlx/backend/metal/kernels/reduction/reduce_row.h | |
| mlx/backend/metal/kernels/steel/defines.h | |
| mlx/backend/metal/kernels/steel/utils.h | |
| mlx/backend/metal/kernels/steel/attn/attn.h | |
| mlx/backend/metal/kernels/steel/attn/loader.h | |
| mlx/backend/metal/kernels/steel/attn/mma.h | |
| mlx/backend/metal/kernels/steel/attn/params.h | |
| mlx/backend/metal/kernels/steel/attn/transforms.h | |
| mlx/backend/metal/kernels/steel/attn/kernels/steel_attention.h | |
| mlx/backend/metal/kernels/steel/attn/kernels/steel_attention.metal | |
| mlx/backend/metal/kernels/steel/conv/conv.h | |
| mlx/backend/metal/kernels/steel/conv/loader.h | |
| mlx/backend/metal/kernels/steel/conv/params.h | |
| mlx/backend/metal/kernels/steel/conv/kernels/steel_conv.h | |
| mlx/backend/metal/kernels/steel/conv/kernels/steel_conv.metal | |
| mlx/backend/metal/kernels/steel/conv/kernels/steel_conv_general.h | |
| mlx/backend/metal/kernels/steel/conv/kernels/steel_conv_general.metal | |
| mlx/backend/metal/kernels/steel/conv/loaders/loader_channel_l.h | |
| mlx/backend/metal/kernels/steel/conv/loaders/loader_channel_n.h | |
| mlx/backend/metal/kernels/steel/conv/loaders/loader_general.h | |
| mlx/backend/metal/kernels/steel/gemm/gemm.h | |
| mlx/backend/metal/kernels/steel/gemm/loader.h | |
| mlx/backend/metal/kernels/steel/gemm/mma.h | |
| mlx/backend/metal/kernels/steel/gemm/params.h | |
| mlx/backend/metal/kernels/steel/gemm/transforms.h | |
| mlx/backend/metal/kernels/steel/gemm/kernels/steel_gemm_fused.h | |
| mlx/backend/metal/kernels/steel/gemm/kernels/steel_gemm_fused.metal | |
| mlx/backend/metal/kernels/steel/gemm/kernels/steel_gemm_gather.h | |
| mlx/backend/metal/kernels/steel/gemm/kernels/steel_gemm_gather.metal | |
| mlx/backend/metal/kernels/steel/gemm/kernels/steel_gemm_masked.h | |
| mlx/backend/metal/kernels/steel/gemm/kernels/steel_gemm_masked.metal | |
| mlx/backend/metal/kernels/steel/gemm/kernels/steel_gemm_segmented.h | |
| mlx/backend/metal/kernels/steel/gemm/kernels/steel_gemm_segmented.metal | |
| mlx/backend/metal/kernels/steel/gemm/kernels/steel_gemm_splitk.h | |
| mlx/backend/metal/kernels/steel/gemm/kernels/steel_gemm_splitk.metal | |
| mlx/backend/metal/kernels/steel/utils/integral_constant.h | |
| mlx/backend/metal/kernels/steel/utils/type_traits.h | |
| mlx/backend/no_cpu/CMakeLists.txt | |
| mlx/backend/no_cpu/available.cpp | |
| mlx/backend/no_cpu/compiled.cpp | |
| mlx/backend/no_cpu/primitives.cpp | |
| mlx/backend/no_gpu/CMakeLists.txt | |
| mlx/backend/no_gpu/allocator.cpp | |
| mlx/backend/no_gpu/apple_memory.h | |
| mlx/backend/no_gpu/eval.cpp | |
| mlx/backend/no_gpu/event.cpp | |
| mlx/backend/no_gpu/fence.cpp | |
| mlx/backend/no_gpu/linux_memory.h | |
| mlx/backend/no_gpu/primitives.cpp | |
| mlx/distributed/CMakeLists.txt | |
| mlx/distributed/distributed.cpp | |
| mlx/distributed/distributed.h | |
| mlx/distributed/distributed_impl.h | |
| mlx/distributed/ops.cpp | |
| mlx/distributed/ops.h | |
| mlx/distributed/primitives.cpp | |
| mlx/distributed/primitives.h | |
| mlx/distributed/mpi/CMakeLists.txt | |
| mlx/distributed/mpi/mpi.cpp | |
| mlx/distributed/mpi/mpi.h | |
| mlx/distributed/mpi/mpi_declarations.h | |
| mlx/distributed/mpi/no_mpi.cpp | |
| mlx/distributed/nccl/CMakeLists.txt | |
| mlx/distributed/nccl/nccl.cpp | |
| mlx/distributed/nccl/nccl.h | |
| mlx/distributed/nccl/no_nccl.cpp | |
| mlx/distributed/nccl/nccl_stub/CMakeLists.txt | |
| mlx/distributed/nccl/nccl_stub/nccl_stubs.cpp | |
| mlx/distributed/ring/CMakeLists.txt | |
| mlx/distributed/ring/no_ring.cpp | |
| mlx/distributed/ring/ring.cpp | |
| mlx/distributed/ring/ring.h | |
| mlx/io/CMakeLists.txt | |
| mlx/io/gguf.cpp | |
| mlx/io/gguf.h | |
| mlx/io/gguf_quants.cpp | |
| mlx/io/load.cpp | |
| mlx/io/load.h | |
| mlx/io/no_gguf.cpp | |
| mlx/io/no_safetensors.cpp | |
| mlx/io/safetensors.cpp | |
| mlx/types/bf16.h | |
| mlx/types/complex.h | |
| mlx/types/fp16.h | |
| mlx/types/half_types.h | |
| mlx/types/limits.h | |
| python/mlx/__main__.py | |
| python/mlx/_os_warning.py | |
| python/mlx/_reprlib_fix.py | |
| python/mlx/distributed_run.py | |
| python/mlx/extension.py | |
| python/mlx/py.typed | |
| python/mlx/utils.py | |
| python/mlx.egg-info/PKG-INFO | |
| python/mlx.egg-info/SOURCES.txt | |
| python/mlx.egg-info/dependency_links.txt | |
| python/mlx.egg-info/entry_points.txt | |
| python/mlx.egg-info/not-zip-safe | |
| python/mlx.egg-info/requires.txt | |
| python/mlx.egg-info/top_level.txt | |
| python/mlx/nn/__init__.py | |
| python/mlx/nn/init.py | |
| python/mlx/nn/losses.py | |
| python/mlx/nn/utils.py | |
| python/mlx/nn/layers/__init__.py | |
| python/mlx/nn/layers/activations.py | |
| python/mlx/nn/layers/base.py | |
| python/mlx/nn/layers/containers.py | |
| python/mlx/nn/layers/convolution.py | |
| python/mlx/nn/layers/convolution_transpose.py | |
| python/mlx/nn/layers/distributed.py | |
| python/mlx/nn/layers/dropout.py | |
| python/mlx/nn/layers/embedding.py | |
| python/mlx/nn/layers/linear.py | |
| python/mlx/nn/layers/normalization.py | |
| python/mlx/nn/layers/pooling.py | |
| python/mlx/nn/layers/positional_encoding.py | |
| python/mlx/nn/layers/quantized.py | |
| python/mlx/nn/layers/recurrent.py | |
| python/mlx/nn/layers/transformer.py | |
| python/mlx/nn/layers/upsample.py | |
| python/mlx/optimizers/__init__.py | |
| python/mlx/optimizers/optimizers.py | |
| python/mlx/optimizers/schedulers.py | |
| python/src/CMakeLists.txt | |
| python/src/array.cpp | |
| python/src/buffer.h | |
| python/src/constants.cpp | |
| python/src/convert.cpp | |
| python/src/convert.h | |
| python/src/cuda.cpp | |
| python/src/device.cpp | |
| python/src/distributed.cpp | |
| python/src/export.cpp | |
| python/src/fast.cpp | |
| python/src/fft.cpp | |
| python/src/indexing.cpp | |
| python/src/indexing.h | |
| python/src/linalg.cpp | |
| python/src/load.cpp | |
| python/src/load.h | |
| python/src/memory.cpp | |
| python/src/metal.cpp | |
| python/src/mlx.cpp | |
| python/src/mlx_func.cpp | |
| python/src/mlx_func.h | |
| python/src/ops.cpp | |
| python/src/random.cpp | |
| python/src/small_vector.h | |
| python/src/stream.cpp | |
| python/src/transforms.cpp | |
| python/src/trees.cpp | |
| python/src/trees.h | |
| python/src/utils.cpp | |
| python/src/utils.h |