### Number of Processes on the Official Docker Image

The **number of processes** on this image is **computed automatically** from the CPU **cores** available.

This means that it will try to **squeeze** as much **performance** from the CPU as possible.

You can also adjust it with the configurations using **environment variables**, etc.

  %0 = "tfl.squeeze"(%arg0) {squeeze_dims = [0]}: (tensor<*xf32>) -> tensor<*xf32>
  func.return %0: tensor<*xf32>

// CHECK-LABEL: DontConvertSqueezeToReshape
// CHECK: %[[RESULT:.*]] = "tfl.squeeze"(%arg0)
// CHECK:  return %[[RESULT]]
}

func.func @ConvertSqueezeToReshapeOnMultiDynamicDims(%arg0: tensor<?x?xf32>) -> tensor<?x?xf32> {

// CHECK-DAG:         %[[CST_1:.*]] = "tf.Const"() <{value = dense<[1, 1504]> : tensor<2xi32>}> : () -> tensor<2xi32>
// CHECK:             %[[VAL_2:.*]] = "tf.Squeeze"(%[[VAL_1]]) <{squeeze_dims = [0]}> : (tensor<1x2xi32>) -> tensor<2xi32>
// CHECK:             %[[VAL_3:.*]] = "tf.Slice"(%[[ARG_0]], %[[VAL_2]], %[[CST_1]]) : (tensor<1x2944xi32>, tensor<2xi32>, tensor<2xi32>) -> tensor<1x1504xi32>

  if (!input_type) return success();  // Can't verify squeeze dims.

  int64_t input_rank = input_type.getRank();
  for (const auto &squeeze_dim_apint :
       op.getSqueezeDims().getAsValueRange<IntegerAttr>()) {
    int64_t squeeze_dim = squeeze_dim_apint.getSExtValue();
    if (squeeze_dim < -input_rank || squeeze_dim >= input_rank) {
      return op.emitOpError()
             << "squeeze dimension " << squeeze_dim << " not in ["

          {"MISC",
           // clang-format off
     {"ApproxTopK", "BroadcastTo", "ExpandDims", "Fill", "NoOp",
      "Range", "Rank", "Reshape", "Shape", "ShapeN", "Size", "Squeeze",
      "Transpose", "ZerosLike", "OnesLike", "BiasAdd" /*PW + Broadcast*/,
      "BroadcastArgs", "BroadcastGradientArgs", "OneHot", "Concat", "ConcatV2",

	for len(out) > 0 {
		n := copy(out, d.storage[d.i:d.n])
		d.i += n
		out = out[n:]

		// Apply the permutation if we've squeezed the sponge dry.
		if d.i == d.rate {
			d.permute()
		}
	}

	return
}

// Sum applies padding to the hash state and then squeezes out the desired
// number of output bytes. It panics if any output has already been read.
func (d *state) Sum(in []byte) []byte {

  func.return
}

// -----

func.func @testSqueezeOutOfBounds(%arg0: tensor<?x?x10xf32>) -> tensor<?x10xf32> {
  // expected-error @+1 {{squeeze dimension -4 not in [-3, 3)}}
  %0 = "tf.Squeeze"(%arg0) { squeeze_dims = [-4] }: (tensor<?x?x10xf32>) -> tensor<?x10xf32>
  func.return %0 : tensor<?x10xf32>
}

// -----

func.func @testTernaryEinsum(%arg0: tensor<2x3xf32>){

// TFL_TransposeOp when the tensor has some dimensions with value==1
// Example- "tfl.transpose"(tensor<56x8x56x1x1x1x7xf32>, [4, 5, 1, 2, 0, 6, 3])
// Permutation before squeese is [4, 5, 1, 2, 0, 6, 3] becomes [1, 2, 0, 3]
// after squeeze is perfomed to retain the relative ordering of the non-1 dims.
DenseElementsAttr GetSqueezedPermutation(Value input_value,
                                         Value input_permutation) {

    SmallVector<int64_t, 4> begin_indices(value_rank, 0);
    auto end_indices = llvm::to_vector<4>(value_type.getShape());
    SmallVector<int64_t, 4> strides(value_rank, 1);

    // All HLO slice+squeeze results used to replace the original tf.Unpack op.
    SmallVector<Value, 4> results;
    results.reserve(op.getNumResults());

    for (int i = 0, end = op.getNumResults(); i < end; ++i) {

include "mlir/IR/PatternBase.td"
include "mlir/Dialect/Arith/IR/ArithOps.td"
include "tensorflow/compiler/mlir/lite/ir/tfl_ops.td"
include "tensorflow/compiler/mlir/lite/utils/utils.td"

// Returns Squeezed shape of a ranked-tensor.
// Squeezed, here, means eliminating any 1s' in the
// dimensions of the tensor
def GetSqueezedShape: NativeCodeCall<"GetSqueezedShape($0)">;