🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

Wondering if the tiny codebook (16 elements) opens any opportunities for GPU kernels (or if the scaling vectors negate it).

WWDC wishes (all long shots): - low-level ANE access (a la kernels) - actual quantized activations (for KV cache) - CoreML fast Hadamard transform - share weights between CoreML and MLX (or MLX ANE backend) - ANE HW metrics: GB/s, FLOPs

Wonder if we’re gonna get a new version of coremltools. Last year it dropped on Monday.

Incoming new coremltools looks like it has some nice bits: - 8 bit input/output tensors (previously all 8bit compute was kept internal) - >1 input can be enumerated shapes (👀ANE)