File size: 17,861 Bytes
9c6594c |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 |
# Copyright The Lightning team.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
from typing import List, Optional, Tuple, Union
import torch
from torch import Tensor
from typing_extensions import Literal
from torchmetrics.functional.classification.roc import binary_roc, multiclass_roc, multilabel_roc
from torchmetrics.utilities import rank_zero_warn
from torchmetrics.utilities.compute import _auc_compute_without_check, _safe_divide
from torchmetrics.utilities.data import interp
from torchmetrics.utilities.enums import ClassificationTask
def _validate_fpr_range(fpr_range: Tuple[float, float]) -> None:
"""Validate the `fpr_range` argument for the logauc metric."""
if not isinstance(fpr_range, tuple) and not len(fpr_range) == 2:
raise ValueError(f"The `fpr_range` should be a tuple of two floats, but got {type(fpr_range)}.")
if not (0 <= fpr_range[0] < fpr_range[1] <= 1):
raise ValueError(f"The `fpr_range` should be a tuple of two floats in the range [0, 1], but got {fpr_range}.")
def _binary_logauc_compute(
fpr: Tensor,
tpr: Tensor,
fpr_range: Tuple[float, float] = (0.001, 0.1),
) -> Tensor:
"""Compute the logauc score for binary classification tasks."""
fpr_range = torch.tensor(fpr_range).to(fpr.device)
if fpr.numel() < 2 or tpr.numel() < 2:
rank_zero_warn(
"At least two values on for the fpr and tpr are required to compute the log AUC. Returns 0 score."
)
return torch.tensor(0.0, device=fpr.device)
tpr = torch.cat([tpr, interp(fpr_range, fpr, tpr)]).sort().values
fpr = torch.cat([fpr, fpr_range]).sort().values
log_fpr = torch.log10(fpr)
bounds = torch.log10(torch.tensor(fpr_range))
lower_bound_idx = torch.where(log_fpr == bounds[0])[0][-1]
upper_bound_idx = torch.where(log_fpr == bounds[1])[0][-1]
trimmed_log_fpr = log_fpr[lower_bound_idx : upper_bound_idx + 1]
trimmed_tpr = tpr[lower_bound_idx : upper_bound_idx + 1]
# compute area and rescale it to the range of fpr
return _auc_compute_without_check(trimmed_log_fpr, trimmed_tpr, 1.0) / (bounds[1] - bounds[0])
def _reduce_logauc(
fpr: Union[Tensor, List[Tensor]],
tpr: Union[Tensor, List[Tensor]],
fpr_range: Tuple[float, float] = (0.001, 0.1),
average: Optional[Literal["macro", "weighted", "none"]] = "macro",
weights: Optional[Tensor] = None,
) -> Tensor:
"""Reduce the logauc score to a single value for multiclass and multilabel classification tasks."""
scores = []
for fpr_i, tpr_i in zip(fpr, tpr):
scores.append(_binary_logauc_compute(fpr_i, tpr_i, fpr_range))
scores = torch.stack(scores)
if torch.isnan(scores).any():
rank_zero_warn(
"LogAUC score for one or more classes/labels was `nan`. Ignoring these classes in {average}-average."
)
idx = ~torch.isnan(scores)
if average is None or average == "none":
return scores
if average == "macro":
return scores[idx].mean()
if average == "weighted" and weights is not None:
weights = _safe_divide(weights[idx], weights[idx].sum())
return (scores[idx] * weights).sum()
raise ValueError(f"Got unknown average parameter: {average}. Please choose one of ['macro', 'weighted', 'none'].")
def binary_logauc(
preds: Tensor,
target: Tensor,
fpr_range: Tuple[float, float] = (0.001, 0.1),
thresholds: Optional[Union[int, List[float], Tensor]] = None,
ignore_index: Optional[int] = None,
validate_args: bool = True,
) -> Tensor:
r"""Compute the `Log AUC`_ score for binary classification tasks.
The score is computed by first computing the ROC curve, which then is interpolated to the specified range of false
positive rates (FPR) and then the log is taken of the FPR before the area under the curve (AUC) is computed. The
score is commonly used in applications where the positive and negative are imbalanced and a low false positive rate
is of high importance.
Accepts the following input tensors:
- ``preds`` (float tensor): ``(N, ...)``. Preds should be a tensor containing probabilities or logits for each
observation. If preds has values outside [0,1] range we consider the input to be logits and will auto apply
sigmoid per element.
- ``target`` (int tensor): ``(N, ...)``. Target should be a tensor containing ground truth labels, and therefore
only contain {0,1} values (except if `ignore_index` is specified). The value 1 always encodes the positive class.
Additional dimension ``...`` will be flattened into the batch dimension.
The implementation both supports calculating the metric in a non-binned but accurate version and a binned version
that is less accurate but more memory efficient. Setting the `thresholds` argument to `None` will activate the
non-binned version that uses memory of size :math:`\mathcal{O}(n_{samples})` whereas setting the `thresholds`
argument to either an integer, list or a 1d tensor will use a binned version that uses memory of
size :math:`\mathcal{O}(n_{thresholds})` (constant memory).
Args:
preds: Tensor with predictions
target: Tensor with ground truth labels
fpr_range: 2-element tuple with the lower and upper bound of the false positive rate range to compute the log
AUC score.
thresholds:
Can be one of:
- If set to `None`, will use a non-binned approach where thresholds are dynamically calculated from
all the data. Most accurate but also most memory consuming approach.
- If set to an `int` (larger than 1), will use that number of thresholds linearly spaced from
0 to 1 as bins for the calculation.
- If set to an `list` of floats, will use the indicated thresholds in the list as bins for the calculation
- If set to an 1d `tensor` of floats, will use the indicated thresholds in the tensor as
bins for the calculation.
ignore_index:
Specifies a target value that is ignored and does not contribute to the metric calculation
validate_args: bool indicating if input arguments and tensors should be validated for correctness.
Set to ``False`` for faster computations.
Returns:
A single scalar with the log auc score
Example:
>>> from torchmetrics.functional.classification import binary_logauc
>>> from torch import tensor
>>> preds = tensor([0.75, 0.05, 0.05, 0.05, 0.05])
>>> target = tensor([1, 0, 0, 0, 0])
>>> binary_logauc(preds, target)
tensor(1.)
"""
_validate_fpr_range(fpr_range)
fpr, tpr, _ = binary_roc(preds, target, thresholds, ignore_index, validate_args)
return _binary_logauc_compute(fpr, tpr, fpr_range)
def multiclass_logauc(
preds: Tensor,
target: Tensor,
num_classes: int,
fpr_range: Tuple[float, float] = (0.001, 0.1),
average: Optional[Literal["macro", "none"]] = "macro",
thresholds: Optional[Union[int, List[float], Tensor]] = None,
ignore_index: Optional[int] = None,
validate_args: bool = True,
) -> Tensor:
r"""Compute the `Log AUC`_ score for multiclass classification tasks.
The score is computed by first computing the ROC curve, which then is interpolated to the specified range of false
positive rates (FPR) and then the log is taken of the FPR before the area under the curve (AUC) is computed. The
score is commonly used in applications where the positive and negative are imbalanced and a low false positive rate
is of high importance.
Accepts the following input tensors:
- ``preds`` (float tensor): ``(N, C, ...)``. Preds should be a tensor containing probabilities or logits for each
observation. If preds has values outside [0,1] range we consider the input to be logits and will auto apply
softmax per sample.
- ``target`` (int tensor): ``(N, ...)``. Target should be a tensor containing ground truth labels, and therefore
only contain values in the [0, n_classes-1] range (except if `ignore_index` is specified).
Additional dimension ``...`` will be flattened into the batch dimension.
The implementation both supports calculating the metric in a non-binned but accurate version and a binned version
that is less accurate but more memory efficient. Setting the `thresholds` argument to `None` will activate the
non-binned version that uses memory of size :math:`\mathcal{O}(n_{samples})` whereas setting the `thresholds`
argument to either an integer, list or a 1d tensor will use a binned version that uses memory of
size :math:`\mathcal{O}(n_{thresholds} \times n_{classes})` (constant memory).
Args:
preds: Tensor with predictions
target: Tensor with true labels
num_classes: Integer specifying the number of classes
fpr_range: 2-element tuple with the lower and upper bound of the false positive rate range to compute the log
AUC score.
thresholds:
Can be one of:
- If set to `None`, will use a non-binned approach where thresholds are dynamically calculated from
all the data. Most accurate but also most memory consuming approach.
- If set to an `int` (larger than 1), will use that number of thresholds linearly spaced from
0 to 1 as bins for the calculation.
- If set to an `list` of floats, will use the indicated thresholds in the list as bins for the calculation
- If set to an 1d `tensor` of floats, will use the indicated thresholds in the tensor as
bins for the calculation.
average:
Defines the reduction that is applied over classes. Should be one of the following:
- ``macro``: Calculate score for each class and average them
- ``"none"`` or ``None``: calculates score for each class and applies no reduction
ignore_index:
Specifies a target value that is ignored and does not contribute to the metric calculation
validate_args: bool indicating if input arguments and tensors should be validated for correctness.
Set to ``False`` for faster computations.
Example:
>>> from torchmetrics.functional.classification import multiclass_logauc
>>> preds = torch.tensor([[0.75, 0.05, 0.05, 0.05, 0.05],
... [0.05, 0.75, 0.05, 0.05, 0.05],
... [0.05, 0.05, 0.75, 0.05, 0.05],
... [0.05, 0.05, 0.05, 0.75, 0.05]])
>>> target = torch.tensor([0, 1, 3, 2])
>>> multiclass_logauc(preds, target, num_classes=5, average="macro", thresholds=None)
tensor(0.4000)
>>> multiclass_logauc(preds, target, num_classes=5, average=None, thresholds=None)
tensor([1., 1., 0., 0., 0.])
"""
if validate_args:
_validate_fpr_range(fpr_range)
fpr, tpr, _ = multiclass_roc(
preds, target, num_classes, thresholds, average=None, ignore_index=ignore_index, validate_args=validate_args
)
return _reduce_logauc(fpr, tpr, fpr_range, average)
def multilabel_logauc(
preds: Tensor,
target: Tensor,
num_labels: int,
fpr_range: Tuple[float, float] = (0.001, 0.1),
average: Optional[Literal["macro", "none"]] = "macro",
thresholds: Optional[Union[int, List[float], Tensor]] = None,
ignore_index: Optional[int] = None,
validate_args: bool = True,
) -> Tensor:
r"""Compute the `Log AUC`_ score for multilabel classification tasks.
The score is computed by first computing the ROC curve, which then is interpolated to the specified range of false
positive rates (FPR) and then the log is taken of the FPR before the area under the curve (AUC) is computed. The
score is commonly used in applications where the positive and negative are imbalanced and a low false positive rate
is of high importance.
Accepts the following input tensors:
- ``preds`` (float tensor): ``(N, C, ...)``. Preds should be a tensor containing probabilities or logits for each
observation. If preds has values outside [0,1] range we consider the input to be logits and will auto apply
sigmoid per element.
- ``target`` (int tensor): ``(N, C, ...)``. Target should be a tensor containing ground truth labels, and therefore
only contain {0,1} values (except if `ignore_index` is specified).
Additional dimension ``...`` will be flattened into the batch dimension.
The implementation both supports calculating the metric in a non-binned but accurate version and a binned version
that is less accurate but more memory efficient. Setting the `thresholds` argument to `None` will activate the
non-binned version that uses memory of size :math:`\mathcal{O}(n_{samples})` whereas setting the `thresholds`
argument to either an integer, list or a 1d tensor will use a binned version that uses memory of
size :math:`\mathcal{O}(n_{thresholds} \times n_{labels})` (constant memory).
Args:
preds: Tensor with predictions
target: Tensor with true labels
num_labels: Integer specifying the number of labels
fpr_range: 2-element tuple with the lower and upper bound of the false positive rate range to compute the log
AUC score.
average:
Defines the reduction that is applied over labels. Should be one of the following:
- ``macro``: Calculate score for each label and average them
- ``"none"`` or ``None``: calculates score for each label and applies no reduction
thresholds:
Can be one of:
- If set to `None`, will use a non-binned approach where thresholds are dynamically calculated from
all the data. Most accurate but also most memory consuming approach.
- If set to an `int` (larger than 1), will use that number of thresholds linearly spaced from
0 to 1 as bins for the calculation.
- If set to an `list` of floats, will use the indicated thresholds in the list as bins for the calculation
- If set to an 1d `tensor` of floats, will use the indicated thresholds in the tensor as
bins for the calculation.
ignore_index:
Specifies a target value that is ignored and does not contribute to the metric calculation
validate_args: bool indicating if input arguments and tensors should be validated for correctness.
Set to ``False`` for faster computations.
Example:
>>> from torchmetrics.functional.classification import multilabel_logauc
>>> preds = torch.tensor([[0.75, 0.05, 0.35],
... [0.45, 0.75, 0.05],
... [0.05, 0.55, 0.75],
... [0.05, 0.65, 0.05]])
>>> target = torch.tensor([[1, 0, 1],
... [0, 0, 0],
... [0, 1, 1],
... [1, 1, 1]])
>>> multilabel_logauc(preds, target, num_labels=3, average="macro", thresholds=None)
tensor(0.3945)
>>> multilabel_logauc(preds, target, num_labels=3, average=None, thresholds=None)
tensor([0.5000, 0.0000, 0.6835])
"""
fpr, tpr, _ = multilabel_roc(preds, target, num_labels, thresholds, ignore_index, validate_args)
return _reduce_logauc(fpr, tpr, fpr_range, average=average)
def logauc(
preds: Tensor,
target: Tensor,
task: Literal["binary", "multiclass", "multilabel"],
thresholds: Optional[Union[int, List[float], Tensor]] = None,
num_classes: Optional[int] = None,
num_labels: Optional[int] = None,
fpr_range: Tuple[float, float] = (0.001, 0.1),
average: Optional[Literal["macro", "none"]] = None,
ignore_index: Optional[int] = None,
validate_args: bool = True,
) -> Optional[Tensor]:
r"""Compute the `Log AUC`_ score for classification tasks.
The score is computed by first computing the ROC curve, which then is interpolated to the specified range of false
positive rates (FPR) and then the log is taken of the FPR before the area under the curve (AUC) is computed. The
score is commonly used in applications where the positive and negative are imbalanced and a low false positive rate
is of high importance.
"""
task = ClassificationTask.from_str(task)
if task == ClassificationTask.BINARY:
return binary_logauc(preds, target, fpr_range, thresholds, ignore_index, validate_args)
if task == ClassificationTask.MULTICLASS:
if not isinstance(num_classes, int):
raise ValueError(f"`num_classes` is expected to be `int` but `{type(num_classes)} was passed.`")
return multiclass_logauc(
preds, target, num_classes, fpr_range, average, thresholds, ignore_index, validate_args
)
if task == ClassificationTask.MULTILABEL:
if not isinstance(num_labels, int):
raise ValueError(f"`num_labels` is expected to be `int` but `{type(num_labels)} was passed.`")
return multilabel_logauc(preds, target, num_labels, fpr_range, average, thresholds, ignore_index, validate_args)
return None
|