Lines Matching refs:num_thresholds
481 num_thresholds = len(thresholds)
500 array_ops.tile(array_ops.transpose(predictions_2d), [num_thresholds, 1]),
506 label_is_pos = array_ops.tile(labels_2d, [num_thresholds, 1])
514 array_ops.reshape(weights, [1, -1]), [num_thresholds, 1])
525 [num_thresholds], dtypes.float32, name='true_positives')
537 [num_thresholds], dtypes.float32, name='false_negatives')
549 [num_thresholds], dtypes.float32, name='true_negatives')
561 [num_thresholds], dtypes.float32, name='false_positives')
578 num_thresholds=200,
596 (computed using the aforementioned variables). The `num_thresholds` variable
599 dramatically depending on `num_thresholds`.
620 num_thresholds: The number of thresholds to use when discretizing the roc
657 (i + 1) * 1.0 / (num_thresholds - 1) for i in range(num_thresholds - 2)
680 math_ops.multiply(x[:num_thresholds - 1] - x[1:],
681 (y[:num_thresholds - 1] + y[1:]) / 2.),
685 math_ops.multiply(x[:num_thresholds - 1] - x[1:],
686 math_ops.minimum(y[:num_thresholds - 1], y[1:])),
690 math_ops.multiply(x[:num_thresholds - 1] - x[1:],
691 math_ops.maximum(y[:num_thresholds - 1], y[1:])),
2661 num_thresholds=200,
2693 num_thresholds: The number of thresholds to use for matching the given
2726 (i + 1) * 1.0 / (num_thresholds - 1) for i in range(num_thresholds - 2)
3427 num_thresholds=200,
3459 num_thresholds: The number of thresholds to use for matching the given
3492 (i + 1) * 1.0 / (num_thresholds - 1) for i in range(num_thresholds - 2)
