ios – Massive app measurement for utilizing ‘TensorFlowLiteSelectTfOps’, ‘~> 0.0.1-nightly’

I’m utilizing –
‘TensorFlowLiteSwift’
‘TensorFlowLiteSelectTfOps’, ‘~> 0.0.1-nightly’
utilizing pod for working my Tflite mannequin within the iOS app. Nevertheless, the app measurement is extraordinarily massive, virtually 202 MB. If I take away pod – ‘TensorFlowLiteSelectTfOps’, ‘~> 0.0.1-nightly’ then the app measurement is diminished to 32 MB.

So now I’m making an attempt to construct the framework – TensorFlowLiteSelectTfOps.framework

Utilizing a Python script I acquired that my mannequin makes use of the under ops –

{'MUL', 'FlexRealDiv', 'FLOOR_DIV', 'FILL', 'SPLIT_V', 'DIV', 'TRANSPOSE', 'MAXIMUM', 'EXPAND_DIMS', 'FlexIRFFT', 'MIRROR_PAD', 'DEQUANTIZE', 'CONV_2D', 'SUB', 'FlexStridedSlice', 'TRANSPOSE_CONV', 'GATHER', 'PACK', 'ADD', 'SQRT', 'STRIDED_SLICE', 'CONCATENATION', 'IMAG', 'SPLIT', 'RFFT2D', 'RSQRT', 'BATCH_MATMUL', 'SQUARED_DIFFERENCE', 'SUM', 'SHAPE', 'DELEGATE', 'GELU', 'REAL', 'NEG', 'FULLY_CONNECTED', 'BATCH_TO_SPACE_ND', 'RANGE', 'FlexPad', 'FlexComplex', 'SQUEEZE', 'LOGISTIC', 'FlexTranspose', 'MEAN', 'SPACE_TO_BATCH_ND', 'PAD', 'RESHAPE', 'SOFTMAX'}

I did construct the framework utilizing the under command however the measurement will not be diminished –

bazel construct -c choose --config=ios_arm64                                  
--define=TF_LITE_STATIC_SELECT_TF_OPS=1 
--define=tflite_with_xnnpack=false 
--copt=-DTF_LITE_STATIC_SELECT_OPS_LIST="MUL,FlexRealDiv,FLOOR_DIV,FILL,SPLIT_V,DIV,TRANSPOSE,MAXIMUM,EXPAND_DIMS,FlexIRFFT,MIRROR_PAD,DEQUANTIZE,CONV_2D,SUB,FlexStridedSlice,TRANSPOSE_CONV,GATHER,PACK,ADD,SQRT,STRIDED_SLICE,CONCATENATION,IMAG,SPLIT,RFFT2D,RSQRT,BATCH_MATMUL,SQUARED_DIFFERENCE,SUM,SHAPE,DELEGATE,GELU,REAL,NEG,FULLY_CONNECTED,BATCH_TO_SPACE_ND,RANGE,FlexPad,FlexComplex,SQUEEZE,LOGISTIC,FlexTranspose,MEAN,SPACE_TO_BATCH_ND,PAD,RESHAPE,SOFTMAX" 
//tensorflow/lite/ios:TensorFlowLiteSelectTfOps_framework