Merge pull request #2 from FabricioArendTorres/rename

renamed package to flowcon

.github/workflows/python-publish.yml

@@ -44,7 +44,7 @@ jobs:
     # Specifying a GitHub environment is optional, but strongly encouraged
     environment: 
       name: release
-      url: https://pypi.org/p/enflows
+      url: https://pypi.org/p/flowcon
     permissions:
       # IMPORTANT: this permission is mandatory for trusted publishing
       id-token: write

.gitignore

@@ -1,5 +1,5 @@
 # data location
-enflows/datasets/uci/data/
+flowcon/datasets/uci/data/
 
 
 # Development files and python cache

README.md

@@ -72,18 +72,17 @@ $ python examples/toy_2d.py
 As the core is based on `nflows`, its usage is similar. To define a flow:
 
 ```python
-from enflows import transforms, distributions, flows
+from flowcon import transforms, distributions, flows
 
 # Define an invertible transformation.
 transform = transforms.CompositeTransform([
-    transforms.MaskedAffineAutoregressiveTransform(features=2, hidden_features=4),
-    transforms.RandomPermutation(features=2)
+  transforms.MaskedAffineAutoregressiveTransform(features=2, hidden_features=4),
+  transforms.RandomPermutation(features=2)
 ])
 
 # Define a base distribution.
 base_distribution = distributions.StandardNormal(shape=[2])
 
-
 # Combine into a flow.
 flow = flows.Flow(transform=transform, distribution=base_distribution)
 ```

examples/conditional_toy_2d.py

@@ -5,15 +5,15 @@
 
 from torch import optim
 
-from enflows.flows import Flow
-from enflows.distributions.normal import DiagonalNormal
-from enflows.transforms import (ActNorm,
+from flowcon.flows import Flow
+from flowcon.distributions.normal import DiagonalNormal
+from flowcon.transforms import (ActNorm,
                                 iResBlock,
                                 CompositeTransform,
                                 )
-from enflows.nn.nets import *
-from enflows.utils.torchutils import *
-from enflows.datasets.base import load_plane_dataset, InfiniteLoader, PlaneDataset
+from flowcon.nn.nets import *
+from flowcon.utils.torchutils import *
+from flowcon.datasets.base import load_plane_dataset, InfiniteLoader, PlaneDataset
 import logging
 
 logging.basicConfig(

examples/toy_2d.py

@@ -4,11 +4,11 @@
 import torch
 from torch import optim
 
-from enflows.flows import Flow
-from enflows.distributions import StandardNormal
-from enflows.datasets import load_plane_dataset, InfiniteLoader
-from enflows.transforms import *
-from enflows.nn.nets import *
+from flowcon.flows import Flow
+from flowcon.distributions import StandardNormal
+from flowcon.datasets import load_plane_dataset, InfiniteLoader
+from flowcon.transforms import *
+from flowcon.nn.nets import *
 
 if torch.cuda.is_available():
     device = "cuda"

enflows/CNF/cnf.py → flowcon/CNF/cnf.py

@@ -2,7 +2,7 @@
 import torch.nn as nn
 from torchdiffeq import odeint_adjoint as odeint
 
-from enflows.CNF.neural_odes.wrappers.cnf_regularization import RegularizedODEfunc
+from flowcon.CNF.neural_odes.wrappers.cnf_regularization import RegularizedODEfunc
 
 __all__ = ["CNF"]
 

enflows/CNF/neural_odes/odefunc.py → flowcon/CNF/neural_odes/odefunc.py

@@ -2,7 +2,7 @@
 import numpy as np
 import torch
 import torch.nn as nn
-from enflows.transforms import ActNorm
+from flowcon.transforms import ActNorm
 
 from . import diffeq_layers
 from .squeeze import squeeze, unsqueeze

flowcon/__init__.py

@@ -0,0 +1 @@
+from flowcon.version import VERSION as __version__

flowcon/datasets/__init__.py

@@ -0,0 +1,2 @@
+from flowcon.datasets.uci import load_uci_dataset, get_uci_dataset_range, get_uci_dataset_max_abs_value
+from flowcon.datasets.base import load_plane_dataset, InfiniteLoader, Dataset

enflows/datasets/plane.py → flowcon/datasets/plane.py

@@ -8,7 +8,7 @@
 
 from torch.utils.data import Dataset
 import sklearn.datasets
-import enflows.utils as utils
+import flowcon.utils as utils
 from sklearn.utils import shuffle as util_shuffle
 
 
@@ -336,7 +336,7 @@ def _test():
     torch.set_default_tensor_type('torch.cuda.FloatTensor')
     dataset = DiamondDataset(num_points=int(1e6), width=20, bound=2.5, std=0.04)
 
-    from enflows.utils import torchutils
+    from flowcon.utils import torchutils
     from matplotlib import pyplot as plt
     data = torchutils.tensor_to_np(dataset.data)
     fig, ax = plt.subplots(1, 1, figsize=(5, 5))

flowcon/datasets/uci/__init__.py

@@ -0,0 +1,2 @@
+from flowcon.datasets.uci.download import download_and_preprocess_uci_data
+from flowcon.datasets.uci.uci import load_uci_dataset, get_uci_dataset_range, get_uci_dataset_max_abs_value

enflows/datasets/uci/bsds300.py → flowcon/datasets/uci/bsds300.py

@@ -1,6 +1,6 @@
 import numpy as np
 import os
-import enflows.datasets.uci.utils as utils
+import flowcon.datasets.uci.utils as utils
 
 from matplotlib import pyplot as plt
 from torch.utils import data

enflows/datasets/uci/download.py → flowcon/datasets/uci/download.py

@@ -20,7 +20,7 @@
 
 
 def preprocess_uci_data() -> None:
-    from enflows.datasets.uci import gas, power, miniboone, hepmass
+    from flowcon.datasets.uci import gas, power, miniboone, hepmass
 
     gas.save_splits()
     power.save_splits()

enflows/datasets/uci/gas.py → flowcon/datasets/uci/gas.py

@@ -3,7 +3,7 @@
 
 from matplotlib import pyplot as plt
 from torch.utils.data import Dataset
-import enflows.datasets.uci.utils as utils
+import flowcon.datasets.uci.utils as utils
 
 def load_gas():
     import pandas as pd

enflows/datasets/uci/hepmass.py → flowcon/datasets/uci/hepmass.py

@@ -1,6 +1,6 @@
 import numpy as np
 import os
-import enflows.datasets.uci.utils as utils
+import flowcon.datasets.uci.utils as utils
 
 from collections import Counter
 from matplotlib import pyplot as plt

enflows/datasets/uci/miniboone.py → flowcon/datasets/uci/miniboone.py

@@ -1,6 +1,6 @@
 import numpy as np
 import os
-import enflows.datasets.uci.utils as utils
+import flowcon.datasets.uci.utils as utils
 
 from matplotlib import pyplot as plt
 from torch.utils.data import Dataset

enflows/datasets/uci/power.py → flowcon/datasets/uci/power.py

@@ -1,6 +1,6 @@
 import numpy as np
 import os
-import enflows.datasets.uci.utils as utils
+import flowcon.datasets.uci.utils as utils
 
 
 from matplotlib import pyplot as plt

enflows/datasets/uci/uci.py → flowcon/datasets/uci/uci.py

@@ -17,11 +17,11 @@ def load_uci_dataset(name, split, frac=None):
          ValueError: If any of the arguments has an invalid value.
     """
 
-    from enflows.datasets.uci.gas import GasDataset
-    from enflows.datasets.uci.power import PowerDataset
-    from enflows.datasets.uci.hepmass import HEPMASSDataset
-    from enflows.datasets.uci.miniboone import MiniBooNEDataset
-    from enflows.datasets.uci.bsds300 import BSDS300Dataset
+    from flowcon.datasets.uci.gas import GasDataset
+    from flowcon.datasets.uci.power import PowerDataset
+    from flowcon.datasets.uci.hepmass import HEPMASSDataset
+    from flowcon.datasets.uci.miniboone import MiniBooNEDataset
+    from flowcon.datasets.uci.bsds300 import BSDS300Dataset
 
     if split not in ['train', 'val', 'test']:
         raise ValueError('Split must be one of \'train\', \'val\' or \'test\'.')

flowcon/distributions/__init__.py

@@ -0,0 +1,9 @@
+from flowcon.distributions.base import Distribution, NoMeanException
+from flowcon.distributions.discrete import ConditionalIndependentBernoulli
+from flowcon.distributions.mixture import MADEMoG
+from flowcon.distributions.normal import (
+    ConditionalDiagonalNormal,
+    DiagonalNormal,
+    StandardNormal,
+)
+from flowcon.distributions.uniform import LotkaVolterraOscillating, MG1Uniform

enflows/distributions/base.py → flowcon/distributions/base.py

@@ -3,8 +3,8 @@
 import torch
 from torch import nn, optim
 
-from enflows.utils import torchutils
-import enflows.utils.typechecks as check
+from flowcon.utils import torchutils
+import flowcon.utils.typechecks as check
 
 
 class NoMeanException(Exception):

enflows/distributions/discrete.py → flowcon/distributions/discrete.py

@@ -3,8 +3,8 @@
 import torch
 from torch.nn import functional as F
 
-from enflows.distributions.base import Distribution
-from enflows.utils import torchutils
+from flowcon.distributions.base import Distribution
+from flowcon.utils import torchutils
 
 
 class ConditionalIndependentBernoulli(Distribution):

enflows/distributions/mixture.py → flowcon/distributions/mixture.py

@@ -1,7 +1,7 @@
 from torch.nn import functional as F
 
-from enflows.distributions.base import Distribution
-from enflows.nn.nde import MixtureOfGaussiansMADE
+from flowcon.distributions.base import Distribution
+from flowcon.nn.nde import MixtureOfGaussiansMADE
 
 
 class MADEMoG(Distribution):

enflows/distributions/normal.py → flowcon/distributions/normal.py

@@ -4,8 +4,8 @@
 import torch
 from torch import nn
 
-from enflows.distributions.base import Distribution
-from enflows.utils import torchutils
+from flowcon.distributions.base import Distribution
+from flowcon.utils import torchutils
 
 
 class StandardNormal(Distribution):

flowcon/flows/__init__.py

@@ -0,0 +1,3 @@
+from flowcon.flows.autoregressive import MaskedAutoregressiveFlow
+from flowcon.flows.base import Flow
+from flowcon.flows.realnvp import SimpleRealNVP

enflows/flows/autoregressive.py → flowcon/flows/autoregressive.py

@@ -2,12 +2,12 @@
 
 from torch.nn import functional as F
 
-from enflows.distributions.normal import StandardNormal
-from enflows.flows.base import Flow
-from enflows.transforms.autoregressive import MaskedAffineAutoregressiveTransform
-from enflows.transforms.base import CompositeTransform
-from enflows.transforms.normalization import BatchNorm
-from enflows.transforms.permutations import RandomPermutation, ReversePermutation
+from flowcon.distributions.normal import StandardNormal
+from flowcon.flows.base import Flow
+from flowcon.transforms.autoregressive import MaskedAffineAutoregressiveTransform
+from flowcon.transforms.base import CompositeTransform
+from flowcon.transforms.normalization import BatchNorm
+from flowcon.transforms.permutations import RandomPermutation, ReversePermutation
 
 
 class MaskedAutoregressiveFlow(Flow):

enflows/flows/base.py → flowcon/flows/base.py

@@ -2,8 +2,8 @@
 
 import torch.nn
 
-from enflows.distributions.base import Distribution
-from enflows.utils import torchutils
+from flowcon.distributions.base import Distribution
+from flowcon.utils import torchutils
 from inspect import signature
 
 

enflows/flows/realnvp.py → flowcon/flows/realnvp.py

@@ -3,15 +3,15 @@
 import torch
 from torch.nn import functional as F
 
-from enflows.distributions.normal import StandardNormal
-from enflows.flows.base import Flow
-from enflows.nn import nets as nets
-from enflows.transforms.base import CompositeTransform
-from enflows.transforms.coupling import (
+from flowcon.distributions.normal import StandardNormal
+from flowcon.flows.base import Flow
+from flowcon.nn import nets as nets
+from flowcon.transforms.base import CompositeTransform
+from flowcon.transforms.coupling import (
     AdditiveCouplingTransform,
     AffineCouplingTransform,
 )
-from enflows.transforms.normalization import BatchNorm
+from flowcon.transforms.normalization import BatchNorm
 
 
 class SimpleRealNVP(Flow):

flowcon/nn/__init__.py

@@ -0,0 +1,2 @@
+from flowcon.nn.nets import *
+

flowcon/nn/nde/__init__.py

@@ -0,0 +1 @@
+from flowcon.nn.nde.made import MixtureOfGaussiansMADE

enflows/nn/nde/made.py → flowcon/nn/nde/made.py

@@ -8,7 +8,7 @@
 from torch.nn import functional as F
 from torch.nn import init
 
-from enflows.utils import torchutils
+from flowcon.utils import torchutils
 
 
 def _get_input_degrees(in_features):

enflows/nn/nets/__init__.py → flowcon/nn/nets/__init__.py

@@ -1,8 +1,8 @@
-from enflows.nn.nets.mlp import MLP, FCBlock
-from enflows.nn.nets.resnet import ConvResidualNet, ResidualNet
-from enflows.nn.nets.activations import Swish, CLipSwish, LeakyLSwish, FullSort, LipschitzCube, MaxMin, Sin, CSin, \
+from flowcon.nn.nets.mlp import MLP, FCBlock
+from flowcon.nn.nets.resnet import ConvResidualNet, ResidualNet
+from flowcon.nn.nets.activations import Swish, CLipSwish, LeakyLSwish, FullSort, LipschitzCube, MaxMin, Sin, CSin, \
     LipSwish
-from enflows.nn.nets.invertible_densenet import (DenseNet,
+from flowcon.nn.nets.invertible_densenet import (DenseNet,
                                                  MixedConditionalDenseNet,
                                                  InputConditionalDenseNet,
                                                  LastLayerConditionalDenseNet,

enflows/nn/nets/activations.py → flowcon/nn/nets/activations.py

@@ -160,7 +160,7 @@ def __init__(self, k=5, device=None):
         super(Pila, self).__init__()
         assert k > 0
 
-        from enflows.nn.nets.pila import PilaFunction
+        from flowcon.nn.nets.pila import PilaFunction
         self.PilaFunction = PilaFunction
         self.k = k
 

enflows/nn/nets/invertible_densenet.py → flowcon/nn/nets/invertible_densenet.py

@@ -8,11 +8,11 @@
 from abc import ABC, abstractmethod
 from pprint import pformat
 
-from enflows.nn.nets import activations
-from enflows.nn.nets.extended_basic_nets import ExtendedSequential, ExtendedLinear
-from enflows.nn.nets.spectral_norm import scaled_spectral_norm
-from enflows.nn.nets.lipschitz_dense import LipschitzDenseLayer
-from enflows.nn.nets import MLP
+from flowcon.nn.nets import activations
+from flowcon.nn.nets.extended_basic_nets import ExtendedSequential, ExtendedLinear
+from flowcon.nn.nets.spectral_norm import scaled_spectral_norm
+from flowcon.nn.nets.lipschitz_dense import LipschitzDenseLayer
+from flowcon.nn.nets import MLP
 
 ACTIVATION_FNS = {
     'relu': torch.nn.ReLU,