hccho2

import numpy as np

hccho2

def get_scope_name(scope_name,base_name,mod=1):
    if scope_name =='':
        scope_name = base_name
    else: scope_name = scope_name + '/' + base_name    
    
    all_vars = tf.global_variables()
    n = len([var for var in all_vars if var.name.startswith(scope_name)])
    if n==0:
        return base_name
    else:

hccho2

# coding: utf-8
import tensorflow as tf
import numpy as np
tf.reset_default_graph()

col_size=6
n = 3
a = np.arange(col_size*n*n).reshape(col_size*n,n)
x = tf.convert_to_tensor(a)
y = tf.reshape(x,[col_size*n*n])

hccho2

# coding: utf-8
import tensorflow as tf
import numpy as np
tf.reset_default_graph()

batch_size=2
T=3
input_dim=4
hidden_dim=50

hccho2

x = tf.placeholder(tf.float32,[3,5,5,7])
y = tf.placeholder(tf.float32,[3,5,5,8])
z1 = tf.layers.conv2d(x,filters=3,kernel_size=2,name='XXX')

z2 = tf.layers.conv2d(x,filters=3,kernel_size=2,name='XXX',reuse=True)

print(tf.trainable_variables())

hccho2

N=2   # batch size
T=20  # encoder time length
D1=30 # encoder hidden dim
D2=6 # decoder hidden dim
D3=11 # attention dim
h = np.random.randn(N,T,D1) # all encoder hidden
s = np.random.randn(N,D2) # decoder hidden at one time step
Wm = np.random.randn(D1,D3)
Wq = np.random.randn(D2,D3)
A = np.matmul(h,Wm) + np.expand_dims(np.matmul(s,Wq),axis=1)

hccho2

import numpy as np
import tensorflow as tf
tf.reset_default_graph()
x_data= tf.placeholder(tf.int32,shape=[None,6])
vocab_size=5
init = np.arange(5*3).reshape(vocab_size,-1)
embedding = tf.get_variable("embedding", initializer=init.astype(np.float32),dtype = tf.float32)

inputs = tf.nn.embedding_lookup(embedding, x_data)

hccho2

# coding: utf-8
import torch

w_true = torch.Tensor([1, 2, 3])

X = torch.cat([torch.ones(100, 1), torch.randn(100, 2)], 1)
Y = torch.mv(X, w_true) + torch.randn(100) * 0.5
w = torch.randn(3, requires_grad=True)

lr = 0.1    

hccho2

# coding: utf-8

'''
conv2d: weight(kernel_size,kernel_size,in_channel,out_channel)
conv2d_transpose: weight(kernel_size,kernel_size,out_channel,in_channel)

'''


import numpy as np

hccho2

import numpy as np
import tensorflow as tf
import matplotlib.pyplot as plt
tf.compat.v1.reset_default_graph()

x = np.sin(np.linspace(0,3*np.pi,30))+ np.random.randn(30)*0.1
filter = np.sin(np.linspace(np.pi/2-np.pi/6,np.pi/2+np.pi/6,5))

x = np.array([ 0.15056179,  0.30541199,  0.72807816,  0.940651  ,  0.97211703,
        1.01255197,  0.86129876,  0.71344136,  0.56986685,  0.24477384,