pub fn shard_fp8_block_scaled(
src: &Fp8Weight,
kind: TpShardKind,
tp_rank: usize,
tp_size: usize,
block_size: usize,
gpu: &dyn GpuBackend,
) -> Result<Fp8Weight>Expand description
Shard an FP8 block-scaled weight. weight is [N, K] FP8 bytes;
row_scale is [N/block_size, K/block_size] FP32 (widened at load).
Both slice on the same axis at block granularity.