In Python, use the generic va.array constructor: arr = va.array(data, dtype=va.int32, axis=va.axis()) arr = va.array(data, dtype=va.fi(1, 24, 12), axis=va.axis(w_user=8)) In MATLAB, use type-specific constructors with the axis name-value argument: arr = varray.int32(data, axis=varray.axis()); arr = varray.fi(data, 1, 24, 12, axis=varray.axis(w_user=8)); Constructor HLS Stream Type varray.int8(data, axis=cfg) hls::axis<ap_int<8>> varray.int16(data, axis=cfg) hls::axis<ap_int<16>> varray.int32(data, axis=cfg) hls::axis<ap_int<32>> varray.int64(data, axis=cfg) hls::axis<ap_int<64>> varray.uint8(data, axis=cfg) hls::axis<ap_uint<8>> varray.uint16(data, axis=cfg) hls::axis<ap_uint<16>> varray.uint32(data, axis=cfg) hls::axis<ap_uint<32>> varray.uint64(data, axis=cfg) hls::axis<ap_uint<64>> varray.float(data, axis=cfg) hls::axis<float> varray.fi(data, S, WL, FL, axis=cfg) hls::axis<ap_fixed<WL, WL-FL>> varray.cint16(data, axis=cfg) hls::axis<std::complex<short>> varray.cint32(data, axis=cfg) hls::axis<std::complex<int>> varray.cfloat(data, axis=cfg) hls::axis<std::complex<float>> varray.cfi(data, S, WL, FL, axis=cfg) hls::axis<std::complex<ap_int<WL>>>