flashinfer.fused_moe.CuTileMxfp4Config¶ class flashinfer.fused_moe.CuTileMxfp4Config¶ cuTile MXFP4-weight x MXFP4-activation backend. Expert parallelism and fused shared experts are not supported. __init__() → None¶ Methods __init__() prepare_weights(w1_fp4, w1_block_scale, ...) Build the shared cutile_mxfp4 weight view. supported(arch)