o .wi-3@sddlmZddlmZmZmZmZddlZddlmZddl m Z ddl m Z m Z ddlmZddlmZdd lmZmZdd lmZmZd d giZesPd gZddedededefddZ ddedededededededefddZGdddeZdS))Sequence)AnyListOptionalUnionN)Tensor)Module)NoTrainInceptionV3 _compute_fid)Metric) dim_zero_cat)_MATPLOTLIB_AVAILABLE_TORCH_FIDELITY_AVAILABLE)_AX_TYPE_PLOT_OUT_TYPE),MemorizationInformedFrechetInceptionDistance1MemorizationInformedFrechetInceptionDistance.plottorch_fidelityr皙? features1 features2cosine_distance_epsreturnc Cs|tj|dddk}|tj|dddk}|tj|ddd}|tj|ddd}dtt||}t|jddj}||krG|St |S)z9Compute the cosine distance between two sets of features.dimrT)rkeepdimg?) torchsumnormabsmatmultmeanminvalues ones_like) rrrfeatures1_nozerofeatures2_nozeronorm_f1norm_f2d mean_min_dr-U/home/ubuntu/sommelier/.venv/lib/python3.10/site-packages/torchmetrics/image/mifid.py_compute_cosine_distance%sr/mu1sigma1mu2sigma2c Cs8t||||}t|||}|dkr||dSt|S)zDCompute MIFID score given two sets of features and their statistics.g:0yE>g+=)r r/r zeros_like) r0r1rr2r3rr fid_valuedistancer-r-r._mifid_compute3s r7c seZdZUdZdZeed<dZeed<dZeed<e e ed<e e ed<e ed<dZ e ed <  d"d eee fdededededdf fdd Zde deddfddZde fddZd#fdd Z d$deee ee fdeedefd d!ZZS)%raCalculate 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. normalize: Whether to normalize the input images. If ``True`` the input is expected to be in the range [0, 1] and converted to ``uint8``. If ``False`` the input is expected to already be in the range [0, 255] and of type ``uint8``. If a custom feature extractor is used, this argument is ignored. 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) Fhigher_is_betteris_differentiablefull_state_update real_features fake_features inceptionfeature_networkTrfeaturereset_real_features normalizerkwargsrNc stjdi|d|_t|tr5tstdgd}||vr*td|d|dtdt |gd|_ nt|t rA||_ d |_nt d t|t sNtd ||_t|t sZtd ||_t|trpd |krodksttdtd||_|jdgdd|jdgdddS)NFzMemorizationInformedFrechetInceptionDistance metric requires that `Torch-fidelity` is installed. Either install as `pip install torchmetrics[image]` or `pip install torch-fidelity`.)@ir?z3Integer input to argument `feature` must be one of z , but got .zinception-v3-compat)name features_listTz'Got unknown input to argument `feature`z4Argument `reset_real_features` expected to be a boolz*Argument `normalize` expected to be a boolrrzTArgument `cosine_distance_eps` expected to be a float greater than 0 and less than 1r;)dist_reduce_fxr<r-)super__init__used_custom_model isinstanceintrModuleNotFoundError ValueErrorr strr=r TypeErrorboolrArBfloatr add_state)selfr@rArBrrCvalid_int_input __class__r-r.rKs<    z5MemorizationInformedFrechetInceptionDistance.__init__imgsrealcCsZ|jr |js |dn|}||}|j|_|}|r%|j|dS|j |dS)z)Update the state with extracted features.N) rBrLbyter=dtype orig_dtypedoubler;appendr<)rVrZr[featuresr-r-r.updates z3MemorizationInformedFrechetInceptionDistance.updatec Cspt|j}t|j}tj|ddtj|dd}}t|t|}}t|||||||jd |j S)zWCalculate FID score based on accumulated extracted features from the two distributions.rr)r) r r;r<rr#covr"r7rtor_)rVr;r< mean_real mean_fakecov_realcov_faker-r-r.computes  z4MemorizationInformedFrechetInceptionDistance.computecs8|js|jd}t||jd<dStdS)zReset metric states.r;N)rA _defaultspoprJreset)rVvaluerXr-r.rms   z2MemorizationInformedFrechetInceptionDistance.resetvalaxcCs |||S)aPlot 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) )_plot)rVrorpr-r-r.plots 0r)r?TFr)rN)NN)__name__ __module__ __qualname____doc__r8rS__annotations__r9r:rrrr>rQrrNrTrrKrcrjrmrrrrrr __classcell__r-r-rXr.rCsJ  S      0  r)r)collections.abcrtypingrrrrrrtorch.nnrtorchmetrics.image.fidr r torchmetrics.metricr torchmetrics.utilities.datar torchmetrics.utilities.importsr rtorchmetrics.utilities.plotrr__doctest_requires____doctest_skip__rTr/r7rr-r-r-r.sH