File size: 12,658 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 |
# 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 collections.abc import Sequence
from typing import Any, List, Optional, Union
import torch
from torch import Tensor
from torch.nn import Module
from torchmetrics.image.fid import NoTrainInceptionV3, _compute_fid
from torchmetrics.metric import Metric
from torchmetrics.utilities.data import dim_zero_cat
from torchmetrics.utilities.imports import _MATPLOTLIB_AVAILABLE, _TORCH_FIDELITY_AVAILABLE
from torchmetrics.utilities.plot import _AX_TYPE, _PLOT_OUT_TYPE
__doctest_requires__ = {
("MemorizationInformedFrechetInceptionDistance", "MemorizationInformedFrechetInceptionDistance.plot"): [
"torch_fidelity"
]
}
if not _MATPLOTLIB_AVAILABLE:
__doctest_skip__ = ["MemorizationInformedFrechetInceptionDistance.plot"]
def _compute_cosine_distance(features1: Tensor, features2: Tensor, cosine_distance_eps: float = 0.1) -> Tensor:
"""Compute the cosine distance between two sets of features."""
features1_nozero = features1[torch.sum(features1, dim=1) != 0]
features2_nozero = features2[torch.sum(features2, dim=1) != 0]
# normalize
norm_f1 = features1_nozero / torch.norm(features1_nozero, dim=1, keepdim=True)
norm_f2 = features2_nozero / torch.norm(features2_nozero, dim=1, keepdim=True)
d = 1.0 - torch.abs(torch.matmul(norm_f1, norm_f2.t()))
mean_min_d = torch.mean(d.min(dim=1).values)
return mean_min_d if mean_min_d < cosine_distance_eps else torch.ones_like(mean_min_d)
def _mifid_compute(
mu1: Tensor,
sigma1: Tensor,
features1: Tensor,
mu2: Tensor,
sigma2: Tensor,
features2: Tensor,
cosine_distance_eps: float = 0.1,
) -> Tensor:
"""Compute MIFID score given two sets of features and their statistics."""
fid_value = _compute_fid(mu1, sigma1, mu2, sigma2)
distance = _compute_cosine_distance(features1, features2, cosine_distance_eps)
# secure that very small fid values does not explode the mifid
return fid_value / (distance + 10e-15) if fid_value > 1e-8 else torch.zeros_like(fid_value)
class MemorizationInformedFrechetInceptionDistance(Metric):
r"""Calculate Memorization-Informed Frechet Inception Distance (MIFID_).
MIFID is a improved variation of the Frechet Inception Distance (FID_) that penalizes memorization of the training
set by the generator. It is calculated as
.. math::
MIFID = \frac{FID(F_{real}, F_{fake})}{M(F_{real}, F_{fake})}
where :math:`FID` is the normal FID score and :math:`M` is the memorization penalty. The memorization penalty
essentially corresponds to the average minimum cosine distance between the features of the real and fake
distribution.
Using the default feature extraction (Inception v3 using the original weights from `fid ref2`_), the input is
expected to be mini-batches of 3-channel RGB images of shape ``(3 x H x W)``. If argument ``normalize``
is ``True`` images are expected to be dtype ``float`` and have values in the ``[0, 1]`` range, else if
``normalize`` is set to ``False`` images are expected to have dtype ``uint8`` and take values in the ``[0, 255]``
range. All images will be resized to 299 x 299 which is the size of the original training data. The boolian
flag ``real`` determines if the images should update the statistics of the real distribution or the
fake distribution.
.. hint::
Using this metrics requires you to have ``scipy`` install. Either install as ``pip install
torchmetrics[image]`` or ``pip install scipy``
.. hint::
Using this metric with the default feature extractor requires that ``torch-fidelity``
is installed. Either install as ``pip install torchmetrics[image]`` or
``pip install torch-fidelity``
As input to ``forward`` and ``update`` the metric accepts the following input
- ``imgs`` (:class:`~torch.Tensor`): tensor with images feed to the feature extractor with
- ``real`` (:class:`~bool`): bool indicating if ``imgs`` belong to the real or the fake distribution
As output of `forward` and `compute` the metric returns the following output
- ``mifid`` (:class:`~torch.Tensor`): float scalar tensor with mean MIFID value over samples
Args:
feature:
Either an integer or ``nn.Module``:
- an integer will indicate the inceptionv3 feature layer to choose. Can be one of the following:
64, 192, 768, 2048
- an ``nn.Module`` for using a custom feature extractor. Expects that its forward method returns
an ``(N,d)`` matrix where ``N`` is the batch size and ``d`` is the feature size.
reset_real_features: Whether to also reset the real features. Since in many cases the real dataset does not
change, the features can be cached them to avoid recomputing them which is costly. Set this to ``False`` if
your dataset does not change.
cosine_distance_eps: Epsilon value for the cosine distance. If the cosine distance is larger than this value
it is set to 1 and thus ignored in the MIFID calculation.
kwargs: Additional keyword arguments, see :ref:`Metric kwargs` for more info.
Raises:
RuntimeError:
If ``torch`` is version less than 1.10
ValueError:
If ``feature`` is set to an ``int`` and ``torch-fidelity`` is not installed
ValueError:
If ``feature`` is set to an ``int`` not in [64, 192, 768, 2048]
TypeError:
If ``feature`` is not an ``str``, ``int`` or ``torch.nn.Module``
ValueError:
If ``reset_real_features`` is not an ``bool``
Example::
>>> from torch import randint
>>> from torchmetrics.image.mifid import MemorizationInformedFrechetInceptionDistance
>>> mifid = MemorizationInformedFrechetInceptionDistance(feature=64)
>>> # generate two slightly overlapping image intensity distributions
>>> imgs_dist1 = randint(0, 200, (100, 3, 299, 299), dtype=torch.uint8)
>>> imgs_dist2 = randint(100, 255, (100, 3, 299, 299), dtype=torch.uint8)
>>> mifid.update(imgs_dist1, real=True)
>>> mifid.update(imgs_dist2, real=False)
>>> mifid.compute()
tensor(3003.3691)
"""
higher_is_better: bool = False
is_differentiable: bool = False
full_state_update: bool = False
real_features: List[Tensor]
fake_features: List[Tensor]
inception: Module
feature_network: str = "inception"
def __init__(
self,
feature: Union[int, Module] = 2048,
reset_real_features: bool = True,
normalize: bool = False,
cosine_distance_eps: float = 0.1,
**kwargs: Any,
) -> None:
super().__init__(**kwargs)
if isinstance(feature, int):
if not _TORCH_FIDELITY_AVAILABLE:
raise ModuleNotFoundError(
"MemorizationInformedFrechetInceptionDistance metric requires that `Torch-fidelity` is installed."
" Either install as `pip install torchmetrics[image]` or `pip install torch-fidelity`."
)
valid_int_input = [64, 192, 768, 2048]
if feature not in valid_int_input:
raise ValueError(
f"Integer input to argument `feature` must be one of {valid_int_input}, but got {feature}."
)
self.inception = NoTrainInceptionV3(name="inception-v3-compat", features_list=[str(feature)])
elif isinstance(feature, Module):
self.inception = feature
else:
raise TypeError("Got unknown input to argument `feature`")
if not isinstance(reset_real_features, bool):
raise ValueError("Argument `reset_real_features` expected to be a bool")
self.reset_real_features = reset_real_features
if not isinstance(normalize, bool):
raise ValueError("Argument `normalize` expected to be a bool")
self.normalize = normalize
if not (isinstance(cosine_distance_eps, float) and 1 >= cosine_distance_eps > 0):
raise ValueError("Argument `cosine_distance_eps` expected to be a float greater than 0 and less than 1")
self.cosine_distance_eps = cosine_distance_eps
# states for extracted features
self.add_state("real_features", [], dist_reduce_fx=None)
self.add_state("fake_features", [], dist_reduce_fx=None)
def update(self, imgs: Tensor, real: bool) -> None:
"""Update the state with extracted features."""
imgs = (imgs * 255).byte() if self.normalize else imgs
features = self.inception(imgs)
self.orig_dtype = features.dtype
features = features.double()
if real:
self.real_features.append(features)
else:
self.fake_features.append(features)
def compute(self) -> Tensor:
"""Calculate FID score based on accumulated extracted features from the two distributions."""
real_features = dim_zero_cat(self.real_features)
fake_features = dim_zero_cat(self.fake_features)
mean_real, mean_fake = torch.mean(real_features, dim=0), torch.mean(fake_features, dim=0)
cov_real, cov_fake = torch.cov(real_features.t()), torch.cov(fake_features.t())
return _mifid_compute(
mean_real,
cov_real,
real_features,
mean_fake,
cov_fake,
fake_features,
cosine_distance_eps=self.cosine_distance_eps,
).to(self.orig_dtype)
def reset(self) -> None:
"""Reset metric states."""
if not self.reset_real_features:
# remove temporarily to avoid resetting
value = self._defaults.pop("real_features")
super().reset()
self._defaults["real_features"] = value
else:
super().reset()
def plot(
self, val: Optional[Union[Tensor, Sequence[Tensor]]] = None, ax: Optional[_AX_TYPE] = None
) -> _PLOT_OUT_TYPE:
"""Plot a single or multiple values from the metric.
Args:
val: Either a single result from calling `metric.forward` or `metric.compute` or a list of these results.
If no value is provided, will automatically call `metric.compute` and plot that result.
ax: An matplotlib axis object. If provided will add plot to that axis
Returns:
Figure and Axes object
Raises:
ModuleNotFoundError:
If `matplotlib` is not installed
.. plot::
:scale: 75
>>> # Example plotting a single value
>>> import torch
>>> from torchmetrics.image.mifid import MemorizationInformedFrechetInceptionDistance
>>> imgs_dist1 = torch.randint(0, 200, (100, 3, 299, 299), dtype=torch.uint8)
>>> imgs_dist2 = torch.randint(100, 255, (100, 3, 299, 299), dtype=torch.uint8)
>>> metric = MemorizationInformedFrechetInceptionDistance(feature=64)
>>> metric.update(imgs_dist1, real=True)
>>> metric.update(imgs_dist2, real=False)
>>> fig_, ax_ = metric.plot()
.. plot::
:scale: 75
>>> # Example plotting multiple values
>>> import torch
>>> from torchmetrics.image.mifid import MemorizationInformedFrechetInceptionDistance
>>> imgs_dist1 = lambda: torch.randint(0, 200, (100, 3, 299, 299), dtype=torch.uint8)
>>> imgs_dist2 = lambda: torch.randint(100, 255, (100, 3, 299, 299), dtype=torch.uint8)
>>> metric = MemorizationInformedFrechetInceptionDistance(feature=64)
>>> values = [ ]
>>> for _ in range(3):
... metric.update(imgs_dist1(), real=True)
... metric.update(imgs_dist2(), real=False)
... values.append(metric.compute())
... metric.reset()
>>> fig_, ax_ = metric.plot(values)
"""
return self._plot(val, ax)
|