NNAPI Support Library

About the NNAPI Support Library

If you need help building the Arm NN NNAPI Support Library, please take a look at our build guide.

Support Library Options

The general list of runtime options are described in Runtime options

Supported Operators

The NNAPI Support Library currently supports the following operators:

• ABS
• ADD
• ARGMAX
• ARGMIN
• AVERAGE_POOL_2D
• BATCH_MATMUL
• BATCH_TO_SPACE_ND
• CAST
• CONCATENATION
• CONV_2D
• DEPTH_TO_SPACE
• DEPTHWISE_CONV_2D
• DEQUANTIZE
• DIV
• ELU
• EQUAL
• EXP
• EXPAND_DIMS
• FILL
• FLOOR
• FULLY_CONNECTED
• GATHER
• GREATER
• GREATER_EQUAL
• GROUPED_CONV_2D
• HARD_SWISH
• INSTANCE_NORMALIZATION
• L2_NORMALIZATION
• L2_POOL_2D
• LESS
• LESS_EQUAL
• LOCAL_RESPONSE_NORMALIZATION
• LOG
• LOGICAL_AND
• LOGICAL_NOT
• LOGICAL_OR
• LOGISTIC
• LOG_SOFTMAX
• LSTM
• MAX_POOL_2D
• MAXIMUM
• MEAN
• MINIMUM
• MUL
• NEG
• NOT_EQUAL
• PAD
• PADV2
• PRELU
• POW
• QUANTIZE
• QUANTIZED_LSTM
• QUANTIZED_16BIT_LSTM
• RANK
• REDUCE_MAX
• REDUCE_MIN
• REDUCE_PROD
• REDUCE_SUM
• RELU
• RELU1
• RELU6
• RESHAPE
• RESIZE_BILINEAR
• RESIZE_NEAREST_NEIGHBOR
• REVERSE
• RSQRT
• SIN
• SOFTMAX
• SPACE_TO_BATCH_ND
• SPACE_TO_DEPTH
• SQRT
• SQUEEZE
• STRIDED_SLICE
• SUB
• TILE
• TRANSPOSE
• TRANSPOSE_CONV_2D
• TANH