Python Introduction 2 Deep Learning.pptx

The indentation
matters… First line with
less
indentation is considered to
be outside of the function
definition.
Defining Functions
No header file or declaration of types of function or
arguments
def get_final_answer(filename):
“““Documentation
String””” line1
line2
return total_counter
Function definition begins with
“def.”
Function name and its
arguments.
The keyword ‘return’ indicates
the value to be sent back to the
caller.
Colon.

Python and Types
 Dynamic typing: Python determines the
data types of variable bindings in a program
automatically
 Strong typing: But Python’s not casual
about types, it enforces the types of objects
 For example, you can’t just append an
integer to a string, but must first convert it
to a string
x = “the answer is ” # x bound to a
string y = 23 # y bound to an
integer. print x + y # Python will

Calling a Function
The syntax for a function call is:
>>> def myfun(x, y):
return x * y
>>> myfun(3, 4)
12
Parameters in
Python are Call by
Assignment
• Old values for the variables that are
parameter names are hidden, and these
variables are simply made to refer to
the new values
•

Functions without returns
 All functions in Python have a return
value, even if no return line inside the
code
 Functions without a return return the
special value None
• None is a special constant in the
language
• None is used like NULL, void, or nil in
other languages
• None is also logically equivalent to False
• The interpreter’s REPL doesn’t print None

Function overloading? No.
There is no function overloading in Python
• Unlike C++, a Python function is specified
by its name alone
The number, order, names, or types of
arguments cannot be used to distinguish
between two functions with the same name
• Two different functions can’t have the
same name, even if they have
different arguments
But: see operator overloading in later slides
(Note: van Rossum playing with function overloading for the

Default Values for Arguments
 You can provide default values
for a function’s arguments
 These arguments are optional when
the function is called
>>> def myfun(b, c=3,
d=“hello”): return
b + c
>>> myfun(5,3,”hello”)
>>> myfun(5,3)
>>> myfun(5)
All of the above function calls
return 8

Keyword Arguments
 Can call a function with some/all of its
arguments out of order as long as you specify
their names
>>> def foo(x,y,z): return(2*x,4*y,8*z)
>>> foo(2,3,4)
(4, 12, 32)
>>> foo(z=4, y=2,
x=3) (6, 8, 32)
>>> foo(-2, z=-4,
y=-3)
(-4, -12, -32)
Can be combined with
defaults, too
>>> def
foo(x=1,y=2,z=3):

Functions are first-class objects
Functions can be used as any
other datatype, eg:
• Arguments to function
• Return values of functions
• Assigned to variables
• Parts of tuples, lists, etc
>>> def square(x): return x*x
>>> def applier(q, x): return q(x)
>>> applier(square, 7)
49

Lambda Notation
 Python’s lambda creates
anonymous functions
>>> applier(lambda z: z *
42, 7) 14
 Note: only one expression in the
lambda body; its value is always
returned
 Python supports functional
programming idioms: map, filter,
closures, continuations, etc.

Lambda Notation
>>> f = lambda x,y
:
2 *
x
+ y
>>> f
Be careful with the syntax
<function <lambda> at 0x87d30>
>>> f(3, 4)
10
>>> v = lambda x: x*x(100)
>>> v
<function <lambda> at 0x87df0>
>>> v = (lambda x: x*x)(100)
>>> v
10000

Example: composition
>>> def square(x):
return x*x
>>> def twice(f):
return lambda x: f(f(x))
>>> twice
<function twice at 0x87db0>
>>> quad = twice(square)
>>> quad
>>> quad(5)
625

Example: closure
>>> def counter(start=0, step=1):
x = [start]
def _inc():
x[0]
+=
step
return x[0]
return _inc
>>>
>>>
c1
c2
= counter()
= counter(100, -10)
>>> c1()
1
>>> c2()
90

map, filter, reduce
>>> def add1(x): return x+1
>>> def odd(x): return x%2 == 1
>>> def add(x,y): return x + y
>>> map(add1, [1,2,3,4])
[2, 3, 4, 5]
>>> map(+,[1,2,3,4],[100,200,300,400])
map(+,[1,2,3,4],[100,200,300,400])
^
SyntaxError: invalid syntax
>>> map(add,[1,2,3,4],
[100,200,300,400]) [101, 202, 303, 404]
>>> reduce(add, [1,2,3,4])
10
>>> filter(odd, [1,2,3,4])
[1, 3]

Lambda Notation
Python’s lambda creates anonymous
functions
>>> lambda x: x + 1
<function <lambda> at 0x1004e6ed8>
>>> f = lambda x: x + 1
>>> f
<function <lambda> at 0x1004e6f50>
>>> f(100)
101

Lambda Notation
>>> f = lambda x,y: 2 *
x
+ y
>>> f
Be careful with the syntax
>>> f(3, 4)
10
>>> v = lambda x: x*x(100)
>>> v
<function <lambda> at 0x87df0>
>>> v = (lambda x: x*x)(100)
>>> v
10000

Lambda Notation Limitations
 Note: only one expression in the
lambda body; Its value is always
returned
 The lambda expression must fit on
one line!
 Lambda will probably be
deprecated in future versions of
python
Guido is not a lambda fanboy

Functional programming
 Python supports functional
programming idioms
 Builtins for map, reduce, filter,
closures, continuations, etc.
These are often used with lambda

map
>>> def add1(x): return x+1
>>> map(add1, [1,2,3,4])
[2, 3, 4, 5]
>>> map(lambda x: x+1,
[1,2,3,4]) [2, 3, 4, 5]
>>> map(+, [1,2,3,4],
[100,200,300,400])
map(+,[1,2,3,4],
[100,200,300,400])
^

map
 + is an operator, not a function
We can define a corresponding add
function
>>> def add(x, y): return x+y
>>> map(add,[1,2,3,4],
[100,200,300,400]) [101, 202, 303, 404]
Or import the operator
module
>>> from operator import *
>>> map(add, [1,2,3,4],
[100,200,300,400])
[101, 202, 303, 404]

filter, reduce
Python has buiting for reduce and
filter
>>> reduce(add, [1,2,3,4])
10
>>> filter(odd, [1,2,3,4])
[1, 3]
 The map, filter and reduce
functions are also at risk 

Python Introduction 2 Deep Learning.pptx

More Related Content

Similar to Python Introduction 2 Deep Learning.pptx (20)

Recently uploaded (20)

Python Introduction 2 Deep Learning.pptx