crowsonkb

# The name of the Pygments (syntax highlighting) style to use.
sys.path.append('.')
import pygments.styles
import base16_spacemacs_kat
# pygments.styles.base16_spacemacs_kat = base16_spacemacs_kat
sys.modules['pygments.styles.' + 'base16_spacemacs_kat'] = base16_spacemacs_kat
pygments.styles.STYLE_MAP['base16-spacemacs-kat'] = 'base16_spacemacs_kat::Base16SpacemacsStyle'
pygments_style = 'base16-spacemacs-kat'  # 'friendly'

crowsonkb

from prettyprinter import register_pretty, pretty_call
import pyparsing as pp

@register_pretty(pp.ParseResults)
def pretty_parse_results(value, ctx):
    return pretty_call(ctx, pp.ParseResults, value.asList(), value.asDict())

crowsonkb

from torch import nn


class Lambda(nn.Module):
    """Wraps a callable in an :class:`nn.Module` without registering it."""

    def __init__(self, func):
        super().__init__()
        object.__setattr__(self, 'forward', func)

crowsonkb

from torch.nn import functional as F
from torchvision import transforms


def random_shift(input, max_shift, mode='replicate', value=0):
    padded = F.pad(input, max_shift, mode=mode, value=value)
    return transforms.RandomCrop(input.shape[2:])(padded)

crowsonkb

"""Computes the covariance matrix in PyTorch."""


def cov_mean(input, unbiased=True, keepdims=False):
    n = input.shape[-1] - unbiased
    mean = input.mean(dim=-1, keepdims=True)
    dev = input - mean
    mean = mean if keepdims else mean[..., 0]
    return dev @ dev.transpose(-1, -2) / n, mean

crowsonkb

import torch
from torch import nn, optim


class Constraint(nn.Module):
    def __init__(self, fn, maximum, damping=1e-2):
        super().__init__()
        self.fn = fn
        self.register_buffer('maximum', torch.as_tensor(maximum))
        self.register_buffer('damping', torch.as_tensor(damping))

crowsonkb

import abc

import torch
from torch import nn, optim


class Constraint(nn.Module, metaclass=abc.ABCMeta):
    def __init__(self, fn, damping):
        super().__init__()
        self.fn = fn

crowsonkb

from math import ceil

import torch
from torch import nn
from torch.nn import functional as F


class Downsample2d(nn.Module):
    kernels = {
        'binomial2': [0.25, 0.5, 0.25],

crowsonkb

"""Good differentiable image resampling for PyTorch."""

from functools import update_wrapper
import math

import torch
from torch.nn import functional as F


def sinc(x):

crowsonkb

"""Trains IMLE on the MNIST dataset."""

import torch
from torch import optim, nn
from torch.nn import functional as F
from torch.utils import data
from torchvision import datasets, transforms, utils
from torchvision.transforms import functional as TF
from tqdm import tqdm