atc转换模型失败

平台:
ss928 svp nnn架构
转换命令：

atc  --input_shape="images:1,3,640,640" \
--input_format=NCHW --output="./model/cat" \
--soc_version=SS928V100 --framework=5 \
--model="amctmodel/quantmodel_deploy_model.onnx" \
--insert_op_conf=./model/insert_op.cfg

报错：

begin net parsing....
[ERROR][LoadLibrary][46] ERROR: dlopen libpath[libsvp_custom.so] failed
[ERROR][ParseLayer][179] Unknown Op[AscendQuant]. Known Ops[Abs, AbsVal, Acos, Acosh, Add, And, ArgMax, ArgMin, Asin, Asinh, Atan, Atanh, AveragePool, BILSTM, BIRNN, BNLL, BatchNorm, BatchNormalization, Bias, BinaryMath, BitShift, CReLU, Ceil, Celu, Clip, Cmp, Compress, Concat, CondEnd, ConstantOfShape, ConvTranspose, Convolution, Cos, Cosh, Crop, CumSum, Custom, DecBBox, Deconvolution, DepthToSpace, DepthwiseConv, Det, DetectionOutput, Div, Dropout, ELU, Einsum, ElseBegin, Eltwise, Equal, Exp, Expand, Extract, ExtractSlice, EyeLike, Filter, Flatten, Floor, GRU, Gather, GatherElements, Gemm, GlobalAveragePool, GlobalLpPool, GlobalMaxPool, Greater, GreaterOrEqual, HardSigmoid, Hardmax, Hswish, Identity, If, InnerProduct, Input, InstanceNormalization, Interp, LRN, LSTM, LeakyRelu, Less, LessOrEqual, Log, Logical, Loop, LpNorm, LpPool, MVN, MatMul, Max, MaxPool, MaxRoiPool, MaxUnpool, Mean, MeanVarianceNormalization, Min, Mish, Mod, Mul, NMS, Neg, NonMaxSuppression, NonZero, Nop, Normalize, Not, OneHot, Or, PReLU, PRelu, PSROI, PSROIPooling, Pad, Parameter, PassThrough, Passthrough, Permute, Pooling, PoolingMask, Power, QLinearMatMul, RNN, ROIPooling, RReLU, Range, ReLU, ReLU6, Reciprocal, ReduceL1, ReduceL2, ReduceLogSum, ReduceLogSumExp, ReduceMax, ReduceMean, ReduceMin, ReduceProd, ReduceSum, ReduceSumSquare, Reduction, Reorg, Reshape, Resize, Reverse, Round, SPP, Scale, Scan, Scatter, ScatterElements, Selu, Shape, Shrink, ShuffleChannel, Sigmoid, Sign, Silence, Sin, Sinh, Size, Slice, Softmax, Softplus, Softsign, Sort, SpaceToDepth, Split, Sqrt, Squeeze, Sub, Sum, Tan, TanH, Tanh, ThenEnd, Threshold, ThresholdedRelu, Tile, TopK, Transpose, Unique, Unsqueeze, Upsample, Where, Xor], ExtendedOps[].
[ERROR][ParseNormlNode][326] Run failed, error code 0x00000001.
[ERROR][ParseModel][812] Run failed, error code 0x00000001.
[ERROR][Parse][50] Run failed, error code 0x00000001.
[ERROR][StartStageParsing][143] Run failed, error code 0x00000001.
[ERROR][Mapping][449] Run failed, error code 0x00000001.
[ERROR][PicoMapper][363] Run failed, error code 0x00000001.
[ERROR][StartMapper][281] Run failed, error code 0x00000001.
[ERROR][main][57] Run failed, error code 0x00000001.

quantmodel_deploy_model.onnx是经过amct_onnx量化后的模型，量化命令如下：

 amct_onnx calibration --model /mnt/hgfs/share/cat.onnx \
 --save_path ./quantmodel \
 --input_shape "images:1,3,640,640" \
 --data_dir "./calib_bin_data/" --data_types "float32"

ubuntu环境下做的开发，网上查的是缺少libsvp_custom.so这个库，这个库是安装cann工具的时候产生的；但是find / -name libsvp_custom.so整个系统都没搜到这个库；
请问怎么解决；麻烦帮忙看下