TensorFlow Lite and TensorFlow operator compatibility

TensorFlow Lite supports a number of TensorFlow operations used in common inference models. As they are processed by the TensorFlow Lite Optimizing Converter, those operations may be elided or fused, before the supported operations are mapped to their TensorFlow Lite counterparts.

Since the set of TensorFlow Lite operations is smaller than TensorFlow's, not every model is convertible. Even for supported operations, very specific usage patterns are sometimes expected, for performance reasons. We expect to expand the set of supported operations in future TensorFlow Lite releases. Additional ops can be included by using select TensorFlow ops, at the cost of binary size.

The best way to understand how to build a TensorFlow model that can be used with TensorFlow Lite is to carefully consider how operations are converted and optimized, along with the limitations imposed by this process.

Supported types

Most TensorFlow Lite operations target both floating-point (float32) and quantized (uint8, int8) inference, but many ops do not yet for other types like tf.float16 and strings.

Apart from using different version of the operations, the other difference between floating-point and quantized models lies in the way they are converted. Quantized conversion requires dynamic range information for tensors. This requires "fake-quantization" during model training, getting range information via a calibration data set, or doing "on-the-fly" range estimation. See quantization.

Data format and broadcasting

At the moment TensorFlow Lite supports only TensorFlow's "NHWC" format, and broadcasting is only support in a limited number of ops (tf.add, tf.mul, tf.sub, and tf.div).

Compatible operations

The following TensorFlow operations are usually mapped to their TensorFlow Lite counterparts:

tf.batch_to_space_nd —As long as the input tensor is 4D (1 batch + 2 spatial + 1 other) and the crops attribute is not used.
tf.exp
tf.fake_quant
tf.matmul —As the second argument is constant and transposition is not used*
tf.nn.avg_pool
tf.nn.conv2d —As long as the filter is constant.
tf.nn.depthwise_conv2d —As long as the filter is constant and rate is [1, 1].
tf.nn.l2_normalize —As long as normalization is done along the last dimension.
tf.nn.local_response_normalization
tf.nn.log_softmax —As long as axis is not provided.
tf.nn.max_pool
tf.nn.softmax —As long as tensors are 2D and axis is the last dimension.
tf.nn.top_k
tf.one_hot
tf.pad —As long as mode and constant_values are not used.
tf.reduce_mean —As long as the reduction_indices attribute is not used.
tf.reshape
tf.sigmoid
tf.space_to_batch_nd —As long as the input tensor is 4D (1 batch + 2 spatial + 1 other).
tf.space_to_depth
tf.split —As long as num is not provided and num_or_size_split contains number of splits as a 0D tensor.
tf.squeeze —As long as axis is not provided.
tf.squared_difference
tf.strided_slice —As long as ellipsis_mask and new_axis_mask are not used.
tf.transpose —As long as conjugate is not used.

Straight-forward conversions, constant-folding and fusing

A number of TensorFlow operations can be processed by TensorFlow Lite even though they have no direct equivalent. This is the case for operations that can be simply removed from the graph (tf.identity), replaced by tensors (tf.placeholder), or fused into more complex operations (tf.nn.bias_add). Even some supported operations may sometimes be removed through one of these processes.

Here is a non-exhaustive list of TensorFlow operations that are usually removed from the graph:

tf.add
tf.check_numerics
tf.constant
tf.div
tf.divide
tf.fake_quant_with_min_max_args
tf.fake_quant_with_min_max_vars
tf.identity
tf.maximum
tf.minimum
tf.multiply
tf.no_op
tf.placeholder
tf.placeholder_with_default
tf.realdiv
tf.reduce_max
tf.reduce_min
tf.reduce_sum
tf.rsqrt
tf.shape
tf.sqrt
tf.square
tf.subtract
tf.tile
tf.nn.batch_norm_with_global_normalization
tf.nn.bias_add
tf.nn.fused_batch_norm
tf.nn.relu
tf.nn.relu6

Unsupported operations

TensorFlow operation not listed above are likely unsupported. Notably, the following common ops are not supported at the moment:

tf.depth_to_space

TensorFlow Lite operations

The following TensorFlow Lite operations are fully supported and used in place of the TensorFlow operations listed above:

ABS

Inputs {
  0: a tensor
}
Outputs {
  0: elementwise abs of the input
}

ADD

Inputs {
  0: a tensor
  1: a tensor
}
Outputs {
  0: elementwise sum of the input tensors
}
Options {
  fused_activation_function:  NONE|RELU|RELU6
}

ADD_N

Inputs {
  0-N: any number of tensors (must have same size and shape)
}
Outputs {
  0: elementwise sum of the input tensors
}

ARG_MAX

Inputs {
  0: a tensor
  1: a tensor
}
Outputs {
  0: A tensor of indices of maximum values.
}

ARG_MIN

Inputs {
  0: a tensor
  1: a tensor
}
Outputs {
  0: A tensor of indices of minimum values.
}

AVERAGE_POOL_2D

Inputs {
  0: a tensor
}
Outputs {
  0: a tensor where each entry is the mean of the input values in the
     corresponding window.
}
Options {
  fused_activation_function:  NONE|RELU|RELU6
  padding: SAME|VALID
  stride_w,stride_h: stride of the sliding window
  filter_width,filter_height: size of the sliding window
}

BATCH_TO_SPACE_ND

Inputs {
  0: 4D tensor
  1: 1D tensor
  2: 2D tensor
}
Outputs {
  0: tensor rearranged using block_shape. See tf.batch_to_space_nd for
     details.
}

CONCATENATION

Inputs {
  0-N: any number of tensors
}
Outputs {
  0: concatenation of the input tensors along the given axis.
}
Options {
  fused_activation_function:  NONE|RELU|RELU6
  axis: dimension along which the concatenation is performed
}

CONV_2D

Inputs {
  0: 4D tensor
  1: filter
  2: bias (optional)
}
Outputs {
  0: result of 2D convolution of the input tensor
}
Options {
  fused_activation_function:  NONE|RELU|RELU6
  padding: SAME|VALID
  stride_w,stride_h: stride of the filter window
}

TRANSPOSE_CONV

Inputs {
  0: output_shape
  1: filter
  2: 4D tensor
}
Outputs {
  0: the transpose (gradient) of conv2d
}
Options {
  padding: SAME|VALID
  stride_w,stride_h: stride of the filter window
}

DEPTHWISE_CONV_2D

Inputs {
  0: 4D tensor
  1: filter
  2: bias (optional)
}
Outputs {
  0: result of a depthwise-2D convolution of the input tensor
}
Options {
  fused_activation_function:  NONE|RELU|RELU6
  padding: SAME|VALID
  stride_w,stride_h: stride of the filter window
  depth_multiplier: relation between the last dimension of the input and output
    tensors
}

ELU

Inputs {
  0: a tensor
}
Outputs {
  0: a tensor equivalent to exp(features) - 1 if < 0, features otherwise.
}

EQUAL

Inputs {
  0: a tensor
  1: a tensor
}
Outputs {
  0: a tensor of type bool, true whenever an element of the first tensor is
  equal to the corresponding element of the second tensor.
}

EXP

Inputs {
  0: tensor
}
Outputs {
  0: result of computing element-wise exponential of the input tensor
}

FILL

Inputs {
  0: a 1D tensor
  1: a 0D (scalar) tensor
}
Outputs {
  0: A tensor of shape `tensor 0` filled with the value in `tensor 1`.
}

FLOOR

inputs {
  0: tensor
}
outputs: {
  0: result of computing element-wise floor of the input tensor
}

FLOOR_DIV

Inputs {
  0: a tensor
  1: a tensor
}
Outputs {
  0: result of computing element-wise floor of `tensor 0` divided by `tensor 1`.
}

FLOOR_MOD

Inputs {
  0: a tensor
  1: a tensor
}
Outputs {
  0: result of computing element-wise floor of `tensor 0` modulo `tensor 1`.
}

CEIL

Inputs {
  0: a tensor
}
Outputs {
  0: result of computing element-wise ceil of the input tensor
}

FULLY_CONNECTED

Inputs {
  0: 4D tensor
  1: filter
  2: bias (optional)
}
Outputs {
  0: output of a fully (densely) connected layer, which connects all
     elements in the input tensor with each element in this tensor.
}
Options {
  fused_activation_function:  NONE|RELU|RELU6
}

GATHER

Inputs {
  0: params tensor
  1: indices tensor
  2: axis tensor (optional)
}
Outputs {
  0: a tensor with same type as the params tensor.
}

GATHER_ND

Inputs {
  0: params tensor
  1: indices tensor
}
Outputs {
  0: a tensor with same type as the params tensor.
}

GREATER

Inputs {
  0: a tensor
  1: a tensor
}
Outputs {
  0: a tensor of type bool, true whenever an element of the first tensor is
  greater than the corresponding element of the second tensor.
}

GREATER_EQUAL

Inputs {
  0: a tensor
  1: a tensor
}
Outputs {
  0: a tensor of type bool, true whenever an element of the first tensor is
  greater than or equal to the corresponding element of the second tensor.
}

L2_NORMALIZATION

Inputs {
  0: input tensor
}
Outputs {
  0: normalized tensor (along the last dimension)
}
Options {
  fused_activation_function:  NONE|RELU|RELU6
}

L2_POOL_2D

Inputs {
  0: a tensor
}
Outputs {
  0: a tensor equivalent to tf.sqrt(tf.nn.ave_pool(tf.square(input))
}
Options {
  fused_activation_function:  NONE|RELU|RELU6
  padding: SAME|VALID
  stride_w,stride_h: stride of the sliding window
  filter_width,filter_height: size of the sliding window
}

LEAKY_RELU

Inputs {
  0: a tensor
}
Outputs {
  0: a tensor equivalent to max(input, input * alpha)
}
Options {
  alpha: slope of the activation at x < 0 (provided alpha <= 1)
}

LESS

Inputs {
  0: a tensor
  1: a tensor
}
Outputs {
  0: a tensor of type bool, true whenever an element of the first tensor is less
  than the corresponding element of the second tensor.
}

LESS_EQUAL

Inputs {
  0: a tensor
  1: a tensor
}
Outputs {
  0: a tensor of type bool, true whenever an element of the first tensor is less
  than or equal to the corresponding element of the second tensor.
}

LOCAL_RESPONSE_NORMALIZATION

Inputs {
  0: a tensor
}
Outputs {
  0: a tensor equivalent to tf.nn.local_response_normalization
}
Options {
  radius
  bias
  alpha
  beta
}

LOGICAL_OR

Inputs {
  0: a list of tensors.
  1: a list of tensors.
}
Outputs {
  0: A tensor of logical_or output tensors.
}

LOGISTIC

Inputs {
  0: a tensor
}
Outputs {
  0: a tensor equivalent to 1 / (1 + exp(-input))
}

LOG

Inputs {
  0: a tensor
}
Outputs {
  0: a tensor equivalent to log(input)
}

LOG_SOFTMAX

Inputs {
  0: tensor
}
Outputs {
  0: tensor equivalent to logits - log(reduce_sum(exp(logits), -1))
}

MAX_POOL_2D

Inputs {
  0: a tensor
}
Outputs {
  0: a tensor where each entry is the maximum of the input values in the
     corresponding window.
}
Options {
  fused_activation_function:  NONE|RELU|RELU6
  padding: SAME|VALID
  stride_w,stride_h: stride of the sliding window
  filter_width,filter_height: size of the sliding window
}

MUL

Inputs {
  0: a tensor
  1: a tensor
}
Outputs {
  0: elementwise multiplication of the input tensors
}
Options {
  fused_activation_function:  NONE|RELU|RELU6
}

NEG

Inputs {
  0: a tensor
}
Outputs {
  0: elementwise negation of the input tensor
}

NON_MAX_SUPPRESSION_V4

Inputs {
  0: boxes in format [y1, x1, y2, x2]
  1: scores
  2: max number of detections
  3: IOU threshold
  4: score threshold
}
Outputs {
  0: selected indices
  1: number of selected indices
}

NON_MAX_SUPPRESSION_V5

Inputs {
  0: boxes in format [y1, x1, y2, x2]
  1: scores
  2: max number of detections
  3: IOU threshold
  4: score threshold
  5: soft NMS sigma
}
Outputs {
  0: selected indices
  1: selected scores
  2: number of selected indices
}

PACK

Inputs {
  0: a list of tensors.
  1: an integer.
}
Outputs {
  0: A tensor of stacked tensors.
}

PAD

Inputs {
  0: tensor
  1: tensor
}
Outputs {
  0: tensor where additional values are added before and after the contents of
     each dimension
}

MEAN (tf.reduce_mean)

Inputs {
  0: tensor
  1: tensor
}
Outputs {
  0: tensor containing the mean of the elements
}
Options {
  keep_dims: whether to retain reduced dimensions
}

NOT_EQUAL

Inputs {
  0: a tensor
  1: a tensor
}
Outputs {
  0: a tensor of type bool, true whenever an element of the first tensor is not
  equal to the corresponding element of the second tensor.
}

POW

Inputs {
  0: a tensor
  1: a tensor
}
Outputs {
  0: elementwise pow of the input tensors
}

RANGE

Inputs {
  0: a 0D (scalar) tensor
  1: a 0D (scalar) tensor
  2: a 0D (scalar) tensor
}
Outputs {
  0: A 1D tensor of type `dtype` defined by a sequence where `tensor 0` is the
  start, `tensor 1` is the limit, and `tensor 2` is the delta.
}
Options {
  dtype
}

RANK

Inputs {
  0: a tensor
}
Outputs {
  0: a 0-D int32 Tensor representing the rank of input
}

RELU

Inputs {
  0: a tensor
}
Outputs {
  0: a tensor equivalent to max(0, input)
}

RELU_N1_TO_1

Inputs {
  0: a tensor
}
Outputs {
  0: a tensor equivalent to max(-1, min(input, 1)
}

RELU6

Inputs {
  0: a tensor
}
Outputs {
  0: a tensor equivalent to max(0, min(input, 6)
}

RESHAPE

Inputs {
  0: a tensor
  1: ignored
}
Outputs {
  0: a tensor with the same elements as the input but with the new shape
}
Options {
  new_shape
}

RESIZE_BILINEAR

Inputs {
  0: a 4D tensor
  1: a 1D tensor with 2 elements
}
Outputs {
  0: A tensor of type `tensor 0` resized according to `tensor 1` height/width values
  using bilinear interpolation.
}
Options {
  align_corners
}

RESIZE_NEAREST_NEIGHBOR

Inputs {
  0: a 4D tensor
  1: a 1D tensor with 2 elements
}
Outputs {
  0: A tensor of type `tensor 0` resized according to `tensor 1` height/width values
  using nearest neighbors interpolation.
}
Options {
  align_corners
}

RSQRT

Inputs {
  0: a tensor
}
Outputs {
  0: result of computing element-wise reciprocal square root of the input tensor
}

REVERSE_SEQUENCE

Inputs {
  0: a tensor
  1: a 1-D tensor which specifies the length of sequence to be reversed in each
  dim
}
Outputs {
  0: a tensor with the same shape as the input tensor
}
Options {
  seq_dim: a 0-D int tensor (scalar). The dimension which is partially
  reversed.
  batch_dim: a 0-D int tensor (scalar). Defaults to 0. The dimension along
  which reversal is performed.
}

SHAPE

Inputs {
  0: a tensor
}
Outputs {
  0: a 1D tensor representing the shape of the input tensor
}
Options {
  out_type: the output type of the op (int32 or int64). Defaults to int32.
}

ROUND

Inputs {
  0: a tensor
}
Outputs {
  0: result of computing element-wise round of the input tensor
}

SLICE

Inputs {
  0: tensor
  1: 1D tensor
  2: 1D tensor
}
Outputs {
  0: slice of the input tensor of the given size from the given begin index.
}

SOFTMAX

Inputs {
  0: a tensor
}
Outputs {
  0: a tensor equivalent to exp(input) / tf.reduce_sum(exp(input * beta), dim),
     where dim is always the last dimension of the input tensor.
}
Options {
  beta
}

SPACE_TO_DEPTH

Inputs {
  0: a 4D tensor
}
Outputs {
  0: a tensor rearranged using block_size. See tf.space_to_depth for details.
}
Options {
  block_size
}

SPACE_TO_BATCH_ND

Inputs {
  0: 4D tensor
  1: 1D tensor
  2: 2D tensor
}
Outputs {
  0: a tensor rearranged using block_shape. See tf.space_to_batch_nd for
     details.
}

SPARSE_TO_DENSE

Inputs {
  0: 0D or 1D or 2D tensor
  1: 1D tensor
  2: 0D or 1D tensor
  3: 0D tensor
  4: a boolean value
}
Outputs {
  0: Dense Tensor of shape output_shape. Has the same type as sparse_values.
}

SPLIT

Inputs {
  0: 0D tensor (axis)
  1: tensor (input)
}
Outputs {
  0-N: subtensors built from the input tensors
}
Options {
  num_splits: Specifies number of outputs
}

SPLIT_V

Inputs {
  0: tensor (input)
  1: 1-D tensor (size_splits)
  2: 0-D tensor (axis)
}
Outputs {
  0-N: subtensors built from the input tensors
}
Options {
  num_splits: Specifies number of outputs
}

SQRT

Inputs {
  0: a tensor
}
Outputs {
  0: result of computing element-wise square root of the input tensor
}

SQUEEZE

Inputs {
  0: tensor
}
Outputs {
  0: tensor without any dimensions of size 1
}
Options {
  squeeze_dims
}

STRIDED_SLICE

Inputs {
  0: tensor
  1: 1D tensor
  2: 1D tensor
  3: 1D tensor
}
Outputs {
  0: slice of the input tensor of the given size
}
Options {
  begin_mask: mask for begin indices
  end_mask: mask for end indices
  shrink_axis_mask: mask that indicates which dimensions to remove
}

TANH

Inputs {
  0: a tensor
}
Outputs {
  0: result of computing element-wise hyperbolic tangent of the input tensor
}

TOP_K

Inputs {
  0: tensor
  1: OD tensor
}
Outputs {
  0: k largest element along each last dimensional slice
  1: indices of values within the last dimension of the input tensor
}

TRANSPOSE

Inputs {
  0: tensor
  1: tensor
}
Outputs {
  0: tensor permuted according to perm
}

SELECT

Inputs {
  0: tensor
  1: tensor
  2: tensor
}
Outputs {
  0: tensor that contains the elementwise values of 'tensor 1' if the
  corresponding value of 'tensor 0' is true or the value of 'tensor 2' if false.
}

UNPACK

Inputs {
  0: a tensor.
  1: an integer.
  2: an integer.
}
Outputs {
  0-N: tensors of unpacked tensor.
}

WHERE

Inputs {
  0: A tensor of type bool.
  1: A tensor which may have the same shape as condition. If condition is rank
     1, x may have higher rank, but its first dimension must match the size of
     condition.
  2: A tensor with the same shape and type as x.
}
Outputs {
  0: A tensor with the same type and shape as x, y if they are non-None, or
     a tensor with shape (num_true, dim_size(condition)).
}

ZEROS_LIKE

Inputs {
  0: a tensor
}
Outputs {
  0: A tensor of the same shape and type as x but filled with zeros
}

FILL

Inputs {
  0: A Tensor. Must be one of the following types: int32, int64. 1-D. Represents the shape of the output tensor.
  1: A Tensor. 0-D (scalar). Value to fill the returned tensor.
}
Outputs {
  0: A tensor of the same type as value (input1).
}

And these are TensorFlow Lite operations that are present but not ready for custom models yet:

CALL
CONCAT_EMBEDDINGS
CUSTOM
EMBEDDING_LOOKUP
EMBEDDING_LOOKUP_SPARSE
HASHTABLE_LOOKUP
LSH_PROJECTION
LSTM
RNN
SKIP_GRAM
SVDF