shawntan

shawntan

shawntan

import torch
from torch import nn

class RNNCell(nn.Module):
    def __init__(self, hidden_size, dropout, activation=None):
        super(RNNCell, self).__init__()
        self.hidden_size = hidden_size
        self.activation = activation
        linear_transform = nn.Linear(hidden_size * 2, hidden_size)
        torch.nn.init.orthogonal_(linear_transform.weight)

shawntan

4kstogram
accounts-qml-module
acpica
acpid
adwaita-icon-theme
alex
alsa-utils
android-tools
android-udev
arandr

shawntan

import torch
import torch.nn as nn

class RNNOp(nn.Module):
    def __init__(self, nhid, dropout=0.):
        super(RNNOp, self).__init__()
        self.op = nn.Sequential(
            nn.Linear(2 * nhid, nhid),
            nn.Tanh(),
            nn.Dropout(dropout),

shawntan

rules = {'xxx': ' ','xx ': ' ', 'x  ': 'x','x x':' ',' xx':'x',' x ':'x', '  x': 'x', '   ':' '}
initial_string = " " * 100 + "x" + " " * 100 
prev_string = initial_string
for i in range(90):
    print(prev_string)
    prev_string = ''.join(rules.get(prev_string[i - 1:i + 2], ' ') for i in range(len(initial_string)))

shawntan

shawntan

# coding: utf-8
import torch
from torch import nn
import hinton
# import numpy as np


DEBUG = False
torch.autograd.set_detect_anomaly(DEBUG)
if DEBUG:

shawntan

#!/usr/bin/env python

import argparse
import os
import time

import numpy as np
import torch
import torch.nn as nn
import torch.optim.lr_scheduler as lr_scheduler

shawntan