We read every piece of feedback, and take your input very seriously.
To see all available qualifiers, see our documentation.
PACK/UNPACK should support int16_t for int16x8 quantized models