*       stateless, and therefore thread-safe.
 *   <li><b>collision-averse:</b> while it can't be helped that a hash function will sometimes
 *       produce the same hash code for distinct inputs (a "collision"), every hash function strives
 *       to <i>some</i> degree to make this unlikely. (Without this condition, a function that
 *       always returns zero could be called a hash function. It is not.)
 * </ul>
 *

    SEQUENTIAL {
      @Override
      final <T extends @Nullable Object, A extends @Nullable Object, R extends @Nullable Object>
          A result(Collector<T, A, R> collector, Iterable<T> inputs) {
        A accum = collector.supplier().get();
        for (T input : inputs) {
          collector.accumulator().accept(accum, input);
        }
        return accum;
      }
    },

        String[] inputs = { "x", "y" };
        CharMappingItem item = new CharMappingItem(0L, inputs, "result");

        assertEquals(0L, item.getId());
        assertNotNull(item.getNewInputs());
        assertEquals("result", item.getNewOutput());
        // NewInputs should have same content as inputs for id=0 (defensive copy is returned)
        assertArrayEquals(inputs, item.getNewInputs());

     * @param id the unique identifier for this mapping item
     * @param inputs array of input character sequences that will be mapped to the output
     * @param output the output character sequence that inputs will be mapped to
     */
    public CharMappingItem(final long id, final String[] inputs, final String output) {
        this.id = id;
        this.inputs = inputs;
        this.output = output == null ? null : output.replace("\n", " ");

  ASSERT_EQ(TF_OK, TF_GetCode(status.get())) << TF_Message(status.get());

  // Build inputs and outputs.
  TF_AbstractTensor* inputs[2] = {at, at};
  TF_OutputList* o = TF_NewOutputList();
  TF_OutputListSetNumOutputs(o, 1, status.get());
  ASSERT_EQ(TF_OK, TF_GetCode(status.get())) << TF_Message(status.get());

  // Execute.
  TF_ExecuteOperation(op, 2, inputs, o, status.get());

jobs:
  cherrypick:
    name: Cherrypick to ${{ github.event.inputs.release_branch}} - ${{ github.event.inputs.git_commit }}
    runs-on: ubuntu-latest
    if: github.repository == 'tensorflow/tensorflow' # Don't do this in forks
    steps:
    - name: Checkout code
      uses: actions/checkout@8e8c483db84b4bee98b60c0593521ed34d9990e8 # v6.0.1
      with:
        ref: ${{ github.event.inputs.release_branch }}

    SEQUENTIAL {
      @Override
      final <T extends @Nullable Object, A extends @Nullable Object, R extends @Nullable Object>
          A result(Collector<T, A, R> collector, Iterable<T> inputs) {
        A accum = collector.supplier().get();
        for (T input : inputs) {
          collector.accumulator().accept(accum, input);
        }
        return accum;
      }
    },

import com.google.common.collect.Lists;
import com.google.common.primitives.Doubles;
import com.google.common.primitives.Ints;
import java.math.BigInteger;
import java.util.List;
import java.util.stream.DoubleStream;
import org.jspecify.annotations.NullUnmarked;

/**
 * Inputs, expected outputs, and helper methods for tests of {@link StatsAccumulator}, {@link
 * Stats}, {@link PairedStatsAccumulator}, and {@link PairedStats}.

    const TF_Output* inputs, std::vector<OutputTensor>* input_tensors,
    std::unordered_map<const Node*, std::vector<int>>* input_nodes)
    TF_EXCLUSIVE_LOCKS_REQUIRED(fn_body->mu) {
  input_tensors->reserve(ninputs);
  for (int i = 0; i < ninputs; ++i) {
    Node* node = inputs[i].oper ? &inputs[i].oper->node : nullptr;
    int idx = inputs[i].index;

    TF_RETURN_WITH_CONTEXT_IF_ERROR(

                         absl::Span<AbstractTensorHandle*> outputs) {
  return ops::MatMul(ctx, inputs[0], inputs[1], &outputs[0],
                     /*transpose_a=*/false,
                     /*transpose_b=*/false, "MatMul");
}

absl::Status MulModel(AbstractContext* ctx,
                      absl::Span<AbstractTensorHandle* const> inputs,