candle-flash-attn 0.8.0

Flash attention layer for the candle ML framework.
1
# candle-flash-attn