mirror of
https://github.com/ml-explore/mlx.git
synced 2025-06-24 17:31:16 +08:00
02bec0bb6d
* Rough INIT * [WIP]: Loading and Matmuls added * [WIP]: Reductions and min working aligned kernel at headdim = 64 * [WIP] Added headdim 80 for testing * [WIP] Update dispatch params for testing * [WIP] Add support for unaligned seq lengths - still looks messy * Update sdpa_benchmarks * Update sdpa_benchmarks * Update sdpa_benchmarks * Enable gqa support * Update benchmark and switch off 128 headdim * Update headdim 128 tuning * Remove older fast attention code. Write out O strided * Disable hd=128 until further optimizations * Enable bf16 * Fix data size bug * Enable attn build outside of jit |
||
|---|---|---|
| .. | ||
| blas | ||
| comparative | ||
| batch_matmul_bench.py | ||
| compile_bench.py | ||
| conv1d_bench.py | ||
| conv2d_bench_cpu.py | ||
| conv2d_train_bench_cpu.py | ||
| conv2d_transpose_bench_cpu.py | ||
| conv3d_bench_cpu.py | ||
| conv3d_train_bench_cpu.py | ||
| conv3d_transpose_bench_cpu.py | ||
| conv_bench.py | ||
| conv_transpose_bench.py | ||
| distributed_bench.py | ||
| einsum_bench.py | ||
| fft_bench.py | ||
| gather_bench.py | ||
| hadamard_bench.py | ||
| layer_norm_bench.py | ||
| rms_norm_bench.py | ||
| rope_bench.py | ||
| scatter_bench.py | ||
| sdpa_bench.py | ||
| sdpa_vector_bench.py | ||
| single_ops.py | ||
| time_utils.py | ||