This file was automatically generated by Vela using the `--supported-ops-report`

parameter.

Vela version: `3.11.0`

This file complies with **Gitiles Markdown syntax**

Summary table of constraints for:

The table below contains TFLite operators that can be placed on the Ethos-U NPU.

If the constraints are not met, then that operator will be scheduled on the CPU instead.

For any other TFLite operator not listed, will be left untouched and scheduled on the CPU.

Please check the supported operator list for your chosen runtime for further information.

This is a list of constraints most NPU operators must satisfy in order to be scheduled on the NPU. (Operators excluded from certain constraints are shown in brackets [ ] )

- All required operator attributes must be specified
- Input(s) and Output tensors must not be dynamic - [QUANTIZE]
- Input(s) and Output tensors must have a defined shape
- Output tensors cannot be scalar - [QUANTIZE]
- Scalar Input tensors are only valid for op type: ADD, ARG_MAX, EXPAND_DIMS, MAXIMUM, MEAN, MINIMUM, MUL, QUANTIZE, SPLIT, SPLIT_V, SUB
- Input(s) and Output tensors must not be greater than 4D
- Input(s), Output and Weight tensors must have quantization parameters - [ARG_MAX, MIRROR_PAD, SHAPE, TRANSPOSE]
- Input(s), Output and Weight tensors with quantization scales must be finite
- Input and Output tensors must have quantization scales that fit within float32 precision
- Constant tensors should not have NoneType-values
- Tensors must be of type: int16, int32, int8, uint8 - [ARG_MAX]
- Tensors which are int32 are only valid when op type is: ADD, ARG_MAX, MIRROR_PAD, MUL, SHAPE, SUB, TRANSPOSE
- Tensor dimensions must be in the range [1, 65535]
- Per-axis quantization is only supported for the following op types: CONV_2D, DEPTHWISE_CONV_2D, TRANSPOSE_CONV
- IFM Tensor batch size must be 1 - [FULLY_CONNECTED, RESHAPE, SHAPE, SLICE, SOFTMAX, SPLIT, SPLIT_V, SQUEEZE, STRIDED_SLICE, UNPACK]
- The fused activation function (if present) must be one of type: LOGISTIC, RELU, RELU6, RELU_N1_TO_1, TANH
- If a fused activation function is present, the Output tensor must be one of type: int16, int8, uint8

This is a list of constraints that the ABS operator must satisfy in order to be scheduled on the NPU.

- At least one Input's shape must match the OFM's shape
- IFM and OFM data types must match

This is a list of constraints that the ADD operator must satisfy in order to be scheduled on the NPU.

- At least one Input's shape must match the OFM's shape
- Both Input data types must match
- For IFM that are signed, OFM must also be signed
- For IFM that are unsigned, OFM must either be the same type or int32
- Broadcasting is only allowed for rank indices with dimension 1, from either IFM1 or IFM2

This is a list of constraints that the ARG_MAX operator must satisfy in order to be scheduled on the NPU.

- IFM must be int8 or uint8
- OFM must be int32 or int64
- Operation must be performed along the depth axis
- IFM depth must be no greater than 127

This is a list of constraints that the AVERAGE_POOL_2D operator must satisfy in order to be scheduled on the NPU.

- Stride values for both width and height must be integer types
- IFM and OFM data types must match
- Kernel filter values for both width and height must be integer types
- Strides must fulfil the following criteria:

- Stride h must be between 1 and 3 when ofm height is greater than 1

- Stride w must be between 1 and 3 when ofm height is greater than 1 or

stride w must be divisible by 2 or 3 and ifm width must be divisible

by stride_w/2 or stride_w/3 - Stride width must be greater than or equal to 1.

For stride width greater than 3, valid padding needs to be used. - Kernel filter values for both width and height must be in the range [1, 8]
- VALID padding: Kernel filter height must be in the range [1, 256]
- VALID padding: Product of kernel filter width and height must be in the range [1, 65536]

This is a list of constraints that the CONCATENATION operator must satisfy in order to be scheduled on the NPU.

- Axis attribute must exist
- Axis attribute must be in the range [0, <ofm_dimensions>)
- All Input dimensionalities must match OFM dimensionality
- All Input dimensions must match OFM dimension in all axes except the one defined by the axis attribute
- The size of the OFM axis must match the sum of all IFM axis defined by the axis attribute

This is a list of constraints that the CONV_2D operator must satisfy in order to be scheduled on the NPU.

- Stride values for both width and height must be integer types
- IFM depth must be a whole multiple of the filter kernel depth
- Number of filter kernels must be equally divisible by the number of convolution groups
- Dilation factor values for both width and height must be integer types
- Strides must fulfil the following criteria:

- Stride h must be between 1 and 3 when ofm height is greater than 1

- Stride w must be between 1 and 3 when ofm height is greater than 1 or

stride w must be divisible by 2 or 3 and ifm width must be divisible

by stride_w/2 or stride_w/3 - Dilated kernel height must be in the range [1, 64]
- Product of dilated kernel width and height must be in the range [1, 4096]
- Weight tensor must be 8-bit
- Weight tensor must be constant
- The sum of the weights cannot exceed 8323072
- Optional Bias tensor must be of shape: 1D
- Optional Bias tensor must be of type: int32, int64
- Optional Bias tensor values must fit within 40-bits

This is a list of constraints that the DEPTHWISE_CONV_2D operator must satisfy in order to be scheduled on the NPU.

- Stride values for both width and height must be integer types
- Dilation factor values for both width and height must be integer types
- Dilated kernel height must be in the range [1, 64]
- Product of dilated kernel width and height must be in the range [1, 4096]
- Weight tensor must be 8-bit
- Weight tensor must be constant
- The sum of the weights cannot exceed 8323072
- Optional Bias tensor must be of shape: 1D
- Optional Bias tensor must be of type: int32, int64
- Optional Bias tensor values must fit within 40-bits
- Stride values for both width and height must be between 1 and 3
- For depth multipliers > 1, IFM channels must be 1 and OFM channels must be equal to the depth multiplier

This is a list of constraints that the EXP operator must satisfy in order to be scheduled on the NPU.

- At least one Input's shape must match the OFM's shape
- IFM and OFM data types must match
- IFM must be int8 or int16

This is a list of constraints that the EXPAND_DIMS operator must satisfy in order to be scheduled on the NPU.

- Input and output quantisation must match.
- Input and output number of elements must match.

This is a list of constraints that the FULLY_CONNECTED operator must satisfy in order to be scheduled on the NPU.

- The output tensor(s) must have 2D shape
- The IFM and OFM must have the same number of dimensions if keep_num_dims is set to true
- Weight tensor must be 8-bit
- Weight tensor must be constant
- Optional Bias tensor must be of shape: 1D
- Optional Bias tensor must be of type: int32, int64
- Optional Bias tensor values must fit within 40-bits

This is a list of constraints that the HARD_SWISH operator must satisfy in order to be scheduled on the NPU.

- IFM must be int8 or uint8
- IFM and OFM data types must match

This is a list of constraints that the LEAKY_RELU operator must satisfy in order to be scheduled on the NPU.

- At least one Input's shape must match the OFM's shape
- IFM and OFM data types must match

This is a list of constraints that the MAXIMUM operator must satisfy in order to be scheduled on the NPU.

- At least one Input's shape must match the OFM's shape
- IFM and OFM data types must match
- Both Input quantization parameters must match OFM quantization parameters
- Broadcasting is only allowed for rank indices with dimension 1, from either IFM1 or IFM2

This is a list of constraints that the MAX_POOL_2D operator must satisfy in order to be scheduled on the NPU.

- Stride values for both width and height must be integer types
- IFM and OFM data types must match
- Kernel filter values for both width and height must be integer types
- Stride values for both width and height must be in the range [1, 3]
- Kernel filter height must be in the range [1, 256]
- Product of kernel filter width and height must be in the range [1, 65536]

This is a list of constraints that the MEAN operator must satisfy in order to be scheduled on the NPU.

- Input tensor must be at least 2D
- Requirements for axis parameter:

When IFM tensor is 2D:

- Reduction in both axes is supported.

When IFM tensor is 3D or 4D:

- Reduction in Batch axis is only supported if batch size is 1.

- Reduction in both Height and Width axes is supported.

- Reduction in Depth axis is supported if at least one of H,W,C are of size 1. - Product of reduced axes must be no greater than:

- 16777216 for signed 8-bit inputs.

- 8388608 for unsigned 8-bit inputs.

- 65536 for signed 16-bit inputs. - If Width axis is reduced its shape must be no greater than 4096.
- If Depth axis is reduced its shape must be no greater than 4096.

This is a list of constraints that the MINIMUM operator must satisfy in order to be scheduled on the NPU.

- At least one Input's shape must match the OFM's shape
- IFM and OFM data types must match
- Both Input quantization parameters must match OFM quantization parameters
- Broadcasting is only allowed for rank indices with dimension 1, from either IFM1 or IFM2

This is a list of constraints that the MIRROR_PAD operator must satisfy in order to be scheduled on the NPU.

- The padding tensor must have the shape [3,2] or [4,2]
- The pad tensor can only pad width and height
- Pad tensor must be of type: int32, int64
- The number of pad values for each direction must not be larger than the ifm size in that dimension

This is a list of constraints that the MUL operator must satisfy in order to be scheduled on the NPU.

- At least one Input's shape must match the OFM's shape
- Both Input data types must match
- For IFM that are signed, OFM must also be signed
- For IFM that are unsigned, OFM must either be the same type or int32
- Broadcasting is only allowed for rank indices with dimension 1, from either IFM1 or IFM2

This is a list of constraints that the PAD operator must satisfy in order to be scheduled on the NPU.

- Number of input tensors must be exactly 2
- The padding tensor must be constant
- Shape of output tensor must equal to size of input tensor plus padding
- The padding tensor must have the shape [3,2] or [4,2]
- The pad tensor can only pad width and height
- Pad tensor must be of type: int32, int64

This is a list of constraints that the RESHAPE operator must satisfy in order to be scheduled on the NPU.

- Input and output quantisation must match.
- Input and output number of elements must match.
- Shape must be constant

This is a list of constraints that the RESIZE_BILINEAR operator must satisfy in order to be scheduled on the NPU.

- The width and height of the IFM and OFM must match one of the following criteria:

IFM W and H must both be 1

IFM must match OFM

W and H scaling must be equal and OFM W-1 and H-1 must be 2x/4x/8x IFM W-1 and H-1, if align_corners is True

W and H scaling must be equal and OFM W and H must be 2x/4x/8x IFM W and H, if align_corners is False - The size tensor must match the output tensor shape
- Both align_corners and half_pixel_centers can't be True
- For half_pixel_centers the width and height of the IFM and OFM must match one of the following criteria:

IFM W and H are both 1

OFM W and H is 2x IFM W and H

This is a list of constraints that the RESIZE_NEAREST_NEIGHBOR operator must satisfy in order to be scheduled on the NPU.

- The width and height of the IFM and OFM must match one of the following criteria:

IFM W and H must both be 1

IFM must match OFM

W and H scaling must be equal and OFM W-1 and H-1 must be 2x/4x/8x IFM W-1 and H-1, if align_corners is True

W and H scaling must be equal and OFM W and H must be 2x/4x/8x IFM W and H, if align_corners is False - The size tensor must match the output tensor shape
- Both align_corners and half_pixel_centers can't be True

This is a list of constraints that the RSQRT operator must satisfy in order to be scheduled on the NPU.

- At least one Input's shape must match the OFM's shape
- IFM and OFM data types must match
- IFM must be int8

This is a list of constraints that the SLICE operator must satisfy in order to be scheduled on the NPU.

- Begin and Size Input tensors must be constant

This is a list of constraints that the SOFTMAX operator must satisfy in order to be scheduled on the NPU.

- IFM and OFM shapes must match
- IFM and OFM data types must match
- Beta value needs to be positive

This is a list of constraints that the SPLIT operator must satisfy in order to be scheduled on the NPU.

- Axis value must be in the range [-RANK(IFM) to +RANK(IFM))
- Axis must be divisible by number of splits

This is a list of constraints that the SPLIT_V operator must satisfy in order to be scheduled on the NPU.

- Only one size is allowed to be inferred

This is a list of constraints that the SQUARED_DIFFERENCE operator must satisfy in order to be scheduled on the NPU.

- At least one Input's shape must match the OFM's shape

This is a list of constraints that the SQUEEZE operator must satisfy in order to be scheduled on the NPU.

- Input and output quantisation must match.
- Input and output number of elements must match.

This is a list of constraints that the STRIDED_SLICE operator must satisfy in order to be scheduled on the NPU.

- Exactly 4 Input tensors are required
- Begin, End and Stride Input tensors must be constant
- ellipsis_mask must be 0
- new_axis_mask and shrink_axis_mask cannot both be set
- Slice 'end' values must be greater than 'begin' values
- All Strides values must be 1
- Offset attribute must be False

This is a list of constraints that the SUB operator must satisfy in order to be scheduled on the NPU.

- At least one Input's shape must match the OFM's shape
- Both Input data types must match
- For IFM that are signed, OFM must also be signed
- For IFM that are unsigned, OFM must either be the same type or int32
- Broadcasting is only allowed for rank indices with dimension 1, from either IFM1 or IFM2

This is a list of constraints that the TRANSPOSE operator must satisfy in order to be scheduled on the NPU.

- Permutation array must be a 1D tensor with RANK(IFM) elements
- Permutation array must have constant values in the range [0, RANK(IFM))
- The following shape/permutations are supported for transpose:

When ifm rank is 2: WxC -> CxW

When ifm rank is 3: HxWxC -> WxHxC, 1xWxC -> 1xCxW, Hx1xC -> Cx1xH

When ifm rank is 4: 1xHxWxC -> 1xWxHxC, 1x1xWxC -> 1x1xCxW, 1xHx1xC -> 1xCx1xW

This is a list of constraints that the TRANSPOSE_CONV operator must satisfy in order to be scheduled on the NPU.

- Stride values for both width and height must be integer types
- Dilated kernel height must be in the range [1, 64]
- Product of dilated kernel width and height must be in the range [1, 4096]
- Weight tensor must be 8-bit
- Weight tensor must be constant
- The sum of the weights cannot exceed 8323072
- Optional Bias tensor must be of shape: 1D
- Optional Bias tensor must be of type: int32, int64
- Optional Bias tensor values must fit within 40-bits
- Stride values for width and height must match one of the following criteria:

Stride values WxH must be 1x1 or 2x2

Stride WxH 2x1 supported if ifm height and kernel height = 1 - SAME padding: OFM dimensions must equal IFM dimensions multiplied by stride
- VALID padding: OFM dimensions must equal IFM dimensions multiplied by stride,

minus difference between kernel size and stride

This is a list of constraints that the UNIDIRECTIONAL_SEQUENCE_LSTM operator must satisfy in order to be scheduled on the NPU.

- IFM must be int8 or int16
- IFM and OFM data types must match
- IFM and OFM must have 3D shape
- Must have 24 input tensors
- Must have 5 intermediate tensors
- State tensors must be variable
- Must not use CIFG
- Must not use Peephole
- Must not use Projection
- Must not use Normalisation
- All input and recurrent weights must be available
- All recurrent weights must be 2D