►
Description
Join Tanmayee Kamath on a tour of GitHub Codespaces to walk through setting up your data science and machine learning projects. Codespaces is a cloud-powered product with zero configuration so this may just be your favorite new productivity hack.
As always, feel free to leave us a comment below and don't forget to subscribe: http://bit.ly/subgithub
Thanks!
Connect with us.
Facebook: http://fb.com/github
Twitter: http://twitter.com/github
LinkedIn: http://linkedin.com/company/github
About GitHub
GitHub is the best place to share code with friends, co-workers, classmates, and complete strangers. Millions of people use GitHub to build amazing things together. For more info, go to http://github.com
A
Hey
all
wow,
it's
amazing
to
be
here
in
person
after
almost
two
years
welcome.
My
name
is
tanmayi
I'm,
a
product
manager
working
on
code
spaces
here
at
GitHub,
and
today
we're
going
to
learn
how
code
spaces
empowers
developers.
Data
scientists,
like
you
and
me,
be
super
productive
as
we
get
started
on
or
as
we
work
on
our
data,
science
and
machine
learning
projects.
So
with
that,
imagine
it's
your
best
friend's
birthday,
all
right
now.
A
What's
the
first
thing
that
comes
to
mind
when
you
think
of
a
birthday,
a
cake
right,
so
you're
gonna
have
to
get
this
cake
for
your
friend
and
let's
talk
about
some
of
the
ways
you
can
go
about
it.
Your
first
option
is
to
bake
the
cake
from
scratch.
In
this
scenario,
you
would
first
pick
up
your
phone
look
for
some
cake
recipes
zero
in
on
one.
A
Then
you
would
head
to
the
kitchen
to
see
if
you
have
all
of
the
ingredients,
then
I'm
gonna
bet
that
you're
not
gonna
have
all
the
ingredients,
so
you're
gonna
head
to
the
grocery
store,
buy
all
of
the
required
ingredients.
Come
back
home,
maybe
take
a
nap.
After
all
that
shopping,
then
you
would
spend
the
next
few
hours
mixing
all
of
the
ingredients,
baking
the
cake
in
your
oven
and
then
figure
out
how
you
want
to
decorate
it.
Let's
just
say
end
to
end.
A
All
of
this
might
end
up
taking
you
a
day
or
two
before
that
cake
is
ready
to
go
now.
Your
second
option
to
speed
things
up
a
little
bit
might
be
to
get
a
pre-mix.
Instead,
in
this
scenario,
you
don't
have
to,
you
know,
spend
time
figuring
out
the
recipe
or
the
ingredients
per
se,
but
you'd
still
head
to
the
grocery
store,
buy
that
box
of
Ghirardelli
Betty
Crocker,
whichever
one
you
love
and
then
come
back
home
bake
the
cake
and
decorate
it
after
but
hey.
A
If
you're
someone
like
me,
who
has
no
time
to
bake
a
cake
from
scratch,
but
who
loves
eating
cakes
regardless,
then
you
might
head
to
your
favorite
Bakery
in
town
and
simply
place
an
order
for
the
perfectly
decorated
tastiest
cake.
That's
going
to
be
baked
for
you.
So
in
this
scenario,
all
you
have
to
do
is
pick
it
up
right
now.
A
code
space
is
essentially
a
fully
baked
cake
with
all
of
the
ingredients
flavors
toppings
decorations,
created
for
you
just
the
way
you
want
without
you
having
to
bake
that
cake
yourself.
A
We
take
care
of
all
of
that
for
you
and
as
long
as
you
have
a
browser
and
an
internet
connection
at
hand,
you're
good
to
go,
you
can
access
these
environments
from
anywhere
all
right.
So
now
that
we
know
what
code
spaces
are
at
a
high
level,
let
me
tell
you
about
the
time
when
I
started
my
first
job
ever
that's.
Actually
a
picture
of
my
first
badge,
I
started
out
as
a
data
analyst
at
Cummins.
A
Cummins
is
a
company
that
builds
engines
for
context
and,
interestingly
enough,
the
first
thing
that
I
was
going
to
work
on
was
a
data
science
project.
Now,
if
I
remember
correctly,
our
team
was
tasked
with
creating
an
inventory
of
all
the
spare
parts
that
go
into
making
an
engine
and
then
optimize
their
usage
to
you
know
reduce
some
space
in
our
warehouses.
A
Now,
as
someone
who
is
new
to
the
field
New
to
the
job,
the
very
first
question
that
came
to
my
mind
was
how
do
I
get
started
right,
I'm
sure
you
all
have
asked
that
yourselves
as
well.
So
how
do
I
look
at
the
various
data
science
projects?
How
do
I,
acclimate
myself
before
jumping
into
this
real
world
project
now,
like
most
of
us,
would
do
I
pretty
much
went
to
the
internet
started.
A
Looking
up
everything,
data
science
right,
I
came
across
videos,
white
papers,
a
ton
of
GitHub
projects,
so
I
just
selected
a
project
and
decided
to
run
it
on
my
machine.
I
cloned
the
project
and
I
attempted
to
run
it,
and
that's
when
I
was
hit
with
a
slew
of
errors.
First
up,
my
conda
environment
wasn't
set
up
correctly
turns
out,
so
I
had
to
First
understand
what
that
meant
and
then
figure
out.
Good
time
to
you
know,
get
that
set
up
next
up.
Docker
started.
A
I
started,
hitting
into
other
issues
around
specific
machine
learning
modules
and
libraries
that
were
just
missing
on
my
laptop
so
to
go
back
to
the
internet
and
you
know
find
the
syntax
to
get
them
installed
on
my
machine
in
a
nutshell,
I
was
really
struggling
to
get
to
a
point
where
I
could
run
the
sample
and
start
learning
about
data
science,
which
was
my
initial
goal.
Right
so,
if
I
were
to
summarize
how
I
was
feeling
back
when
I
got
started,
I
would
say
it
was
a
very
daunting
experience.
A
There
were
too
many
barriers
and
I
wanted
to
give
up.
That
was
the
feeling
now,
with
code
spaces.
Our
mission
is
to
really
lower
these
barriers
to
entry
and
Empower
you
to
explore.
Learn,
get
started
immediately
out
of
the
box,
so
we're
going
to
replay
the
same
experience
that
I
went
through,
but
with
the
power
of
code
spaces
this
time.
So
let
me
show
you
my
screen
here:
real
quick
and
we're
gonna
head
over
to
this
intro
to
machine
learning
project.
A
Now,
if
you're
just
getting
started,
this
is
I'm
sure
you
know
repo
that
you
might
come
across
right
now.
It
has
two
notebooks
there's
one
for
scikit-learn
one
for
matplotlib
and
then
there's
a
readme.
There
is
no
configuration
or
a
requirements.txt
file,
telling
me
how
to
set
a
environment
up
to
run
this
project
with
code
spaces,
though
all
I
have
to
do,
is
hit
on
this
green
code
button
and
we're
going
to
fire
up
a
code
space
for
this
project.
A
Now,
as
that
spins
up,
what
we
are
doing
is
we
are
creating
a
container
in
the
back
end,
using
a
default
image
that
is
hosted
and
managed
by
GitHub.
For
you,
then
we
are
deploying
that
container
onto
a
Linux
virtual
machine,
which
is
also
managed
by
us.
Now
the
awesome
part
about
this
default
or
Universal
image
as
we
call
it,
is
that
it
has
a
set
of
tools
and
runtimes
think
of
Docker
python
node.
A
All
of
that
already
set
up
for
you
and
in
addition
to
that,
we
also
pre-install
a
set
of
commonly
used
machine
learning,
libraries
and
toolkits
also
in
that
environment,
so
that
you
don't
have
to
worry
about
creating
a
conda
environment
or
setting
up
Docker
desktop
on
your
laptop
before
you
can.
Actually,
you
know,
start
exploring
so
here.
In
a
few
seconds
you
saw
that
my
code
space
is
ready.
It's
it's.
A
fully
featured
IDE
I'll
have
to
highlight
that
I
have
access
to
a
terminal.
A
So
I
can
scroll
down
as
I
come
across
a
line
of
code
here,
you'll
see
that
VSCO
recommends
I
install
the
python
extension
since
we're
going
to
work
with
jupyter
notebooks
we're
going
to
need
a
kernel,
let's
actually
go
ahead
and
take
that
recommendation
and
then
I'm
gonna,
head
back
to
the
same
scikit-learn
notebook
and
at
this
point
I
have
a
very
rich
notebook.
Editing
experience
right
inside
this
code.
Space
and
I
can
simply
come
here
and
run
this
block
of
code.
A
You'll
see
that
this
just
works
out
of
the
box,
primarily
because
scikit-learn
is
already
pre-installed
in
this
environment
and
I
can
focus
on
running
these
cells,
really
learning
about
what
scikit-learn
is
and
get
started
on.
My
Learning
Journey
there
very
similar
to
that
if
I
head
over
to.
Hopefully
this
runs
real,
quick
okay.
A
If
I
head
over
to
this
matplotlib
notebook
again
very
similar
to
the
scikit-learn
this,
this
particular
notebook
tells
me
how
to
get
started
with
matplotlib
and
I
can
also
come
in
here
and
hit
run.
All
and
you'll
see
in
a
few
seconds
that
everything
just
works,
because
matplotlib
is
also
pre-installed
in
this
environment.
So
all
of
the
common
libraries
that
you
use,
whether
it's
numpy,
pandas
and
so
on
and
so
forth,
are
already
set
up
for
you.
A
So
the
power
of
this
experience
is
that,
if
you
are
someone
who's
just
getting
started,
and
you
come
across
one
of
the
thousand
plus
notebooks
that
are
hosted
on
GitHub
today-
that
don't
necessarily
tell
you
how
to
get
started
with
the
project
right,
then
you
can
simply
create
a
code
space
for
them
and
we'll
take
care
of.
You
know
configuring,
that
environment
for
you
and
you
can
focus
on
learning
and
getting
started
all
right.
Now,
exploration
just
doesn't
stop
at
trying
existing
projects
right.
A
You
might
want
to
experiment
prototype
with
your
own
sandbox
environment,
to
sort
of
continue
building
on
your
knowledge
there.
So
to
help
you
with
that,
we
are
adding
support
for
what
we
call
Quick
Start
templates,
now,
the
one
that
we
are
going
to
take
a
look
at
is
this
Jupiter
notebook,
quick,
start
template,
but
I'm
going
to
hit
see
all
here
for
a
second
just
to
show
you
the
wide
range
of
scenarios
that
we've
added
support.
For
now.
A
So,
let's
go
to
this
pre-created
sandbox
codespace
that
I
created
using
the
Jupiter
template
again
you'll
see
that
we
are
using
the
same
default
image,
which
means
myconda
environment
is
set
up
all
of
the
machine
learning
libraries
are
there
in
addition
to
that,
let's
actually
hide
this
terminal,
real
quick.
In
addition
to
that,
we
also
load
this
notebook
up
with
a
set
of
sample
notebooks,
so
that
you
know
you
have
some
starter
code
to
explore
form
instead
of
getting
started
from
scratch.
A
So
I'm
going
to
take
a
look
at
this
population
notebook
here
we
are
essentially
using
matplotlib
to
plot
a
bar
plotty
line.
Chart
so
I'm
going
to
hit
run
all
to
show
you
what
that
looks
like.
So
we
are
plotting
the
population
of
Atlantis
over
the
past
few
years.
Let's
say
I
want
to
continue
experimenting
with
this
notebook
right
and
instead
of
plotting
the
population
for
Atlantis.
Let's
say
I
want
to
plot
it.
For
Vancouver
Vancouver
is
my
city,
so
you
know,
let's,
let's
do
that
now.
A
So
really,
it's
very
simple
for
me
to
jump
into
this
sandbox
playground,
type
of
an
environment
where
I
can
add
more
notebooks
I
can
get
rid
of
these
I,
can
update
them
and
really
continue
building
on
my
knowledge
now,
let's
say
that
you
end
up
loving
what
you're
creating
here-
and
you
know
you
want
to
share
this
with
your
colleagues
or
friends.
You
want
to
collaborate
with
more
folks.
All
you
have
to
do
is
head
over
to
this
Source
control
tab
here,
let's
add
a
commit
message:
added
a
CSV
file.
A
Let's
go
ahead
and
commit
it,
and
at
this
point
I
can
publish
this
code
space
to
a
GitHub
project,
so
I'm
going
to
say,
publish
Branch,
let's
just
call
it
sandbox
environment
and
then
I'm
going
to
zoom
in
so
that
folks
can
see
here,
and
there
are
two
options:
either
I
can
publish
to
a
private
or
a
public
repo.
So
let's
just
select
private
in
this
case
and
within
a
few
seconds
we're
going
to
upload
all
your
files
and
create
a
GitHub
repository
for
you.
A
So
this
is
where
you
can
now
add
more
folks
submit
PRS.
You
know
the
drill
right
and
who
knows
this
might
end
up
becoming
a
fun
project,
a
passion
project
that
you
end
up
maintaining
over
time,
all
right.
So
now
that
we've
seen
all
of
this
I
also
want
to
highlight
that,
in
addition
to
all
of
these
awesome
functionalities,
we
are
also
partnering
with
LinkedIn
learning,
to
offer
a
wide
range
of
courses
right
from
programming
languages,
to
data
science,
machine
learning,
python
that
are
all
seamlessly
integrated
with
code
spaces
to
further
help.
A
A
A
What
this
meant
was
I
had
to
wipe
out
my
laptop
of
all
of
the
configurations
that
I
had
done
during
the
getting
started
phase
and
pretty
much
reset
it
up
with
a
configuration
that
my
team
shared
with
me
I
vividly
remember:
it
was
a
doc
with
about
20
art
steps,
not
fun
at
all.
You
know,
setting
up
a
conda
environment,
really
understanding
Docker,
you
know
specific
to
my
Project's
needs
cloning.
A
A
Two
weeks
to
15
seconds:
that's
that's
amazing.
Right
now,
with
code
spaces,
we
really
want
to
eliminate
the
pain
from
you
to
configure
complex
projects
with
very
specific
configurations
that
also
have
a
ton
of
changing
dependencies
on
your
local
machine
and
really
enable
you
start
coding
on
day
one.
In
short,
we
want
you
to
get
coding
faster.
So
let's
talk
about
how
we
can
do
that
right.
A
First,
up
with
configurationist
code
using
Dev
container
configuration,
you
can
spin
up
fully
configured
environments
that
are
standardized
for
a
given
project
with
all
the
right
dependencies.
Extensions
and
scripts
set
up
for
you.
So
again
we
are
going
to
replay
the
same
scenario
that
I
went
through
with
the
power
of
code
spaces
this
time.
A
So
let
me
head
over
to
this
simple
image:
classifier
project,
real
quick,
now
I
already
have
created
a
code
space
for
it
right
here
and
before
we
see
the
notebook,
let's
actually
take
a
look
at
that
shared
configuration
or
the
dev
container.json.
Now
there
are
a
couple
of
ways
to
look
at
this.
This
is
one
way
that
you're
seeing
on
screen
the
other
way
that
we've
added
support
for
very
recently
is,
if
you
click
on
this
green
code
button
here
and
click
on
these
three
ellipses,
you
can
say
configure
Dev
container.
A
So
here
you
get
a
nice
Dev
container,
editing
experience
right
inside
your
repository,
where
you
can
make
all
the
changes
to
your
heart's
content
and
once
you're
satisfied
with
it.
All
you
have
to
do
is
commit
these
changes
come
with
this
file
back
to
your
repository,
and
once
you
do
that,
every
user,
every
data
scientist
requesting
for
a
code
space
off
of
this
repository,
the
simple
image
classifier
project,
will
have
a
very
standard,
consistent
environment
which
is
based
on
this
very
configuration.
A
Now
you
can
go
very
simple
or
very
complex
with
this
config.
It's
really
up
to
you.
In
this
case,
you
see
it's
fairly,
simple,
right,
I'm
still
using
the
default
image,
which
is
why
you
don't
see
an
image
attribute
here,
but
hey.
If
you
have
a
custom
image
or
a
Docker
file
or
a
Docker
compose
yaml
that
you
want
to
bring
us,
you
can
specify
it
in
this
configuration
and
then
we'll
use
that
to
create
your
code
Space
versus
the
default
image.
Then
I
include
some
extensions.
A
Of
course
we
need
python
Jupiter,
since
we
are
working
with
jupyter,
notebooks
and
then
I
also
have
copilot,
which
is
my
favorite
AI
pair
programmer
and
we'll
see
how
copilot
helps
me
not
have
to
worry
about
syntax,
and
you
know
really
focus
on
the
business
logic
as
I
write
code
on
a
day-to-day
basis
and
then,
in
addition
to
that,
I
also
want
to
highlight.
Dev
container
features
real
quick,
so
you
can
think
of
features
as
automatic
installations
of
shareable
code
that
you'd
otherwise
have
to
maintain
lengthy
scripts.
A
For
on
your
side
now,
the
great
part
is.
We
have
created
this
amazing
Marketplace
here
of
features
that
are
powered
by
us.
So
if
you
see
it
says,
buy
Dev
containers,
which
means
we
host
and
manage
these.
But
then
there
are
some
that
you'll
see
the
community
contributing
as
well
now,
let's
say
you're
working
with
a
tensorflow
project
right
and
you
need
Cuda
installed
and
set
up
in
your
environment.
A
So
now
that
we've
seen
Dev
containers
I'm
going
to
go
back
here,
I
also
want
to
quickly
highlight
that
this
project
also
has
a
requirements.txt
file.
So,
in
addition
to
looking
at
the
dev
container
configuration
when
we
create
your
code
space,
we
also
make
sure
that
we
Auto
install
any
requirements.txt
file
that
lives
in
your
project.
So
there
are
additional
modules
that
you
want
to
specify
in
here.
You
can
feel
free
to
do
that
and
then
we'll
use
that,
to
you
know,
set
up
your
environment
when
you
request
for
a
code
space
awesome.
A
So
now
that
we've
seen
all
of
that
I'm
going
to
head
back
to
this
notebook
here,
real
quick,
let's
get
rid
of
this
and
then
go
to
this
notebook
and
really
quick.
This
is
essentially
simple
image:
classification
using
pytorch
right.
Essentially,
we
are
First
downloading
a
data
set
and
then
previewing
images
from
that
data
set.
Then
we
are
configuring,
a
neural
network.
This
is
really
the
Crux
of
my
notebook
right.
This
is
where
we
convert
a
specific
image
to
a
description.
A
Then
I'm
going
to
train
my
network
with
that
data
set
and
save
a
model,
then
we
will
test
that
trained
model
and
then,
lastly,
we
are
dumping.
Some
accuracy
numbers
into
this
notebook
right
here
now,
let's
say
I
want
to
continue
working
on
this
particular
notebook
right.
Let's
say:
I
want
to
optimize
my
model
accuracy.
So,
first
up,
let's
get
all
of
these
accuracy
numbers
into
a
CSV
file
and
then
let's
plot
a
bar
chart
for
it,
so
that
we
can
take
a
look
look
at
how
well
we're
doing
with
the
accuracy.
A
Currently
so
I'm
gonna
add
a
light,
a
block
of
code
here,
let's
say
import,
CSV
and
you'll-
see
here,
hopefully
folks
can
see
how
copilot
is
pretty
much
predicting.
What
I
should
do
next,
so
I'm
just
going
to
hit
the
Tab
Key
and
complete
that
and
then
copilot
pretty
much
tells
me
the
entire
function
to
be
able
to
create
an
accuracy.csv
file
and
then
calculate
the
accuracy
numbers
and
put
them
into
that
file.
So
let's
make
sure
this
actually
works.
A
I'm
going
to
run
this
block
of
code
and
you'll
see
here
that
accuracy.csv
file
was
created.
Awesome
thanks,
copilot!
Let's
go
back
to
the
notebook
here,
I'm
going
to
X
this
out
next
up,
I
want
to
import
matplotlib,
since
we
need
to
draw
that
bar
chart
and
then
we're
going
to
import
pandas
instead
of
numpy
here
and
we're
just
going
to
see
copilot
do
its
magic.
A
A
A
Okay,
that's!
Okay,
no
worries
we
can.
We
can
just
move
ahead
instead
of
trying
to
oh
there
you
go
thanks,
go
violet
a
little
late,
but
there
it
is
so
yeah
I
can
effectively
just
hit
tab
right
and
complete
those
lines
of
code
without
having
to
worry
about
the
syntax,
to
go
figure
out
how
to
plot
that
bar
chart.
A
Let's
again
run
this
block
of
code
to
make
sure
it
looks
good
and
there
you
go
I
have
this
bar
chart
created
for
me
within
a
few
seconds,
and
I
can
focus
on
optimizing
the
model
focusing
on
the
business
logic.
Think
about
how
I
productionize
this
notebook.
You
know
from
my
Dev
environment
ultimately
to
production
now
talking
about
productionizing
it.
The
other
thing
I
love
to
highlight
is
that
this
is
a
fully
configured
IDE,
which
means,
in
addition
to
you,
know,
working
with
your
notebooks
updating
cells
running
them.
A
You
also
have
access
to
a
terminal,
a
debugger,
so
that
you
can
think
about
how
you
would
move
this
notebook
ultimately
into
production.
So
let's
do
some
code
refactoring
here
right,
something
that
you
would
probably
do
in
your
day-to-day.
Let's
head
back
to
this
second
step
here
where
we
were
creating
the
model
right.
This
is
the
Crux
of
our
notebook,
so
I'm
going
to
copy
paste
pretty
much
this
function
here
and
we
are
going
to
create
a
source
folder
in
my
project.
Inside
of
that,
we
are
going
to
create
two
files.
A
We
are
going
to
create
a
init
inits.py
file
and
then
we're
also
going
to
create
a
configure
model,
dot,
py
file
and
then
inside
that
we're
going
to
paste
all
of
these
contents
I'm
going
to
head
back
here.
Probably,
we
also
need
some
of
these
Imports.
So
let's
move
those
Imports
as
well.
Let
me
just
expand
this:
let's
add
that,
let's
also
add
import
torch
in
here
cool
and
then
we
can
head
back
to
this.
A
Notebook
probably
get
rid
of
this
line
of
code
from
here
and
then
the
one
thing
that
I
need
to
do
real,
quick
is
I
also
need
to
set
a
python
path
so
that
we
can
call
into
our
module
correctly
and
in
the
interest
of
time.
I
do
have
that
line
of
code
copied
to
my
clipboard
here.
So
let's
just
try
to
paste
it.
A
Okay,
there
you
go
and
then
lastly,
I
need
to
Simply
import
the
model
that
we
created
the
configure.model
pi.
So
we're
going
to
say:
oh
thanks
to
copilot
I,
don't
need
to
worry
about
that
syntax
and
at
this
point
I
can
essentially
restart
my
kernel.
So
we
can
type
in
Ctrl
shift
p.
Here,
look
for
the
restart
kernel,
command
and
restart
the
kernel,
and
then
we
can
hit
run
all
and
make
sure
everything
looks
good.
Now,
in
case
we
do
come
across
errors
here,
you
can
always
hit
breakpoints.
A
So
that's
the
power
of
configurationist
code,
where
you
can,
with
a
one
click,
get
into
this
environment,
focus
on
the
task
and
just
be
done
with
it
end
to
end
and
then
simply
discard
this
code
space
and
move
on
to
your
next
task.
If
you
will
all
right
so
now
that
we've
seen
this,
let
me
continue
telling
you
my
story
now,
as
I
continued
working
on
my
project,
right,
I
managed
to
get
started,
get
it
set
up.
I
soon
started
running
into
out
of
memory,
not
enough
cores
on
your
machine
type
of
Errors.
A
Quite
often
now
I
will
admit.
The
data
set
that
we
were
working
on
was
quite
large.
Think
of
tens
of
thousands
of
rows,
so
it
used
to
take
us
hours
to
just
pre-process
it
and
then
do
feature
selection
and
so
on
and
so
forth,
I
and
then,
of
course,
the
the
model
that
we
were
working
with
was
also
super
complex,
so
I
remember,
we
used
to
start
our
model
training
with
that
large
data
set
in
the
morning
and
pretty
much
wait
until
end
of
the
day
to
get
some
meaningful
output
I'm.
A
A
So
we
are
actually
going
to
see
GPU
code
spaces
in
action
now.
So,
let's
head
back
to
a
trimmed
version
of
the
very
popular
stable
diffusion
project,
I'm
sure
some
of
you
are
already
familiar
with
this
one
now.
One
of
the
prerequisites
for
the
stable
diffusion
model
is
that
it
requires
a
GPU
SKU,
primarily
because
it's
a
very
complex
model
and
the
data
set
that
it's
trained
with
is
also
quite
large.
So
I
already
have
this
code
space,
that's
powered
by
a
GPU
created.
For
me.
A
This
first
cell
block
here
actually
checks
if
the
GPU
is
present.
So,
let's
hit
run
here
and
you'll
say
that
it
in
fact
says
GPU
detected,
awesome
and
then
for
context.
Stable
diffusion
is
a
text
to
image
diffusion
model
where
it's
trained
with
a
large
data
set
of
images
and
then
as
users
we
can
supply
it.
With
a
text
prompt
and
you'll
see,
you
can
get
very
creative
with
these
text
prompts
and
then
the
model
will
go
run
and
find
a
fun
image
and
the
most
relevant
image
for
you
and
I
will
say.
A
A
If
I
can
spell
pigeon
eating
a
birthday
cake
because
y'all
are
we
going
to
see
that
in
real
life?
I?
Don't
think
so?
So,
let's,
let's
have
a
model.
Do
it
for
us
and
you'll
see
that
you
know
it
runs
super
fast
looking
through
those
images
that
it
is
loaded
with,
and
hopefully
it
should
show
us
an
image
here.
Oh
there
you
go.
That
is
a
very
realistic
pigeon
sitting
on
a
cupcake
but
I'll
take
it.
It's
fun,
awesome
all
right.
A
So
now
that
we've
seen
GPU
code
spaces,
let
me
continue
telling
you
my
story
now,
I'm
sure,
as
developers,
all
of
us
have
our
favorite
or
preferred
Ides
tools
that
we
enjoy
working
with
right.
This
is
actually
a
picture
of
my
team
back
then,
when
we
had
an
opportunity
to
tour
one
of
the
Cummins
warehouses,
we
saw
how
engines
were
built
and
assembled
in
real
time.
It
was
an
incredible
experience
now
in
my
team.
I
have
personally
come
to
love
and
learn
vs
code.
A
That's
what
I
use
today
as
well,
but
then
I
also
had
some
team
members
who
preferred
working
with
Jupiter
lab
some
of
them
enjoyed
working
with
pycharm.
Some
of
them
just
preferred
SSH
right
now,
with
code
spaces,
we
really
want
to
meet
you
where
you
are
and
provide
you
access
to
the
tools,
editors
that
you
love,
that
you
enjoy
working
with,
so
with
that,
in
addition
to
vs
code,
which
is
the
default
editor
for
code
spaces,
we're
also
adding
support
for
Jupiter
lab
pycharm
and
other
jetbrains
IDs.
A
So
we're
again
going
to
head
back
to
the
same
stable
diffusion
model
that
we
took
a
look
at
and
instead
of
vs
code,
we
are
going
to
spin
it
up
using
Jupiter
lab.
So
there
are
a
couple
of
ways
you
can
go
about
it.
The
first
way
is
using
our
GitHub
CLI.
So
if
you
type
in
ghcs
Jupiter
and
hit
enter
that
should
give
you
a
list
of
running
code
spaces
that
you
have
the
one
that
I
want
is
stable
diffusion.
A
So
let's
go
there
and
hit
enter
and
within
a
few
seconds,
we'll
basically
spin
up
Jupiter
lab
in
your
browser.
There
you
go
as
that
loads
up
the
other
way
to
go
about.
This
is,
if
you
click
on
this
green
code
button
here
go
to
the
code
space.
Then
we've
added
this
open
in
tab
where
you
can
select
from
a
wide
range
of
IDs
and
Jupiter
lab
is
again
one
of
them
right
now,
let's
head
to
this
Jupiter
lab
instance.
Here,
hopefully
it
shows
up
if
the
demo
gods
are
with
me.
A
A
Let's
see
here,
I
mean
if
this
does
not
work,
we're
going
to
try
the
ghcs
Jupiter
command
one
more
time
because
I
have
faith,
you
guys
all
right.
Let's,
let's
try
again,
you
know,
let's
just
I'm
gonna
close
my
command
prompt,
we're
going
to
open
it
again
and
I'm
going
to
type
in
ghcs.
Jupiter
I
mean
this
experience
is
in
beta,
so
you
know
we
would
love
feedback.
This
is
let's
see,
let's
see
come
on.
A
A
All
right
so
you'll
see
that
you
know
that's
the
stable
diffusion
project
right
there
I
can
type
in
ghcs
Jupiter
and
that
should
show
us
a
list
of
existing
code
spaces,
as
we
saw
there.
I
can
just
hit
enter
and
then
that
should
open
up
my
code,
space
with
Jupiter
lab
within
a
few
seconds
and
I
think
I
show
a
couple
of
other
ways
to
also
create
it,
create
and
access
the
code
space
with
Jupiter
lab.
So
let's
take
a
look
at
that,
so
the
first
way
is,
as
I
showed
you
there.
A
You
can
go
to
an
existing
code.
Space
use
the
open
in
option,
and
that
should
give
you
the
open
in
Jupiter
lab
experience.
The
other
one
that
I
think
show
here
as
well
is,
if
you
go
to
github.com,
codespaces
You'll
also
see
some
of
your
code.
Spaces
show
up
there
too,
but
anyway,
right
now,
I'm
going
to
restart
the
kernel
and
run
all
cells
and
we're
then
going
to
go
to
github.com
forward
slash
code
spaces.
A
This
is
where
you'll
see
all
of
the
code
spaces
that
you've
created
across
different
projects,
and
this
is
another
place
where
you
can
open
it
up
with
Jupiter
Labs.
So
if
you'll
see
the
open
in
option
there
and
the
open
in
Jupiter
lab
command
all
right,
so,
let's
head
to
that
stable
diffusion
model,
it's
opened
up
in
Jupiter
lab
instance.
It's
still
powered
by
a
code
space
and
it's
still
using
the
same
configuration
that
you
know
lives
in
my
repository.
If
you
will
so,
let's
go
ahead
and
try
a
different
one
right.
A
Let's
say
I
think
this
was
a
photo
of
a
monkey
scuba
diving
if
I'm
not
wrong.
If
I
remember,
if
my
memory
helps
me
and
yes
there,
you
go,
that's
the
one!
So
let's
run
it
and
hopefully
we
should
see
another
fun
text
to
image
conversion
with
a
GPU
back
code
space
in
jupyter
lab.
Oh
there
you
go!
That's
nice!
That's
a
very
cute
monkey.
It's
underwater!
It
has
it's
wearing
a
scuba
dive
gear.
So
that's
an
awesome
awesome,
one,
all
right
cool!
Let
us
continue
talking
about
my
story
now.
Oh
that's
another
one!
A
That's
very
right!
That
there
were
two
models
that
are
that
I
had
in
the
devil.
All
right
can
we
switch
to
the
deck
awesome.
Now,
when
there's
a
team,
there
is
a
ton
of
sharing
and
collaboration
right.
Even
today,
when
I
come
across
an
awesome,
notebook
I
just
can't
wait
to
share
it
with
my
teammates
or
friends
so
that
they
can
try
it
out
as
well
now
with
code
spaces.
A
Our
mission
is
not
only
to
get
you
up
and
running
quickly,
but
to
also
enable
quick
sharing
collaboration
with
your
team
or
just
the
community,
as
we
all
learn
and
build
products
together.
So
with
that
I'm
going
to
share
a
sneak
peek
into
something
super
exciting
that
we're
working
on
before
we
play
a
pre-recorded
video
Let's
assume
that
one
of
my
teammates
shared
this
fun
notebook
that
they
came
across
via
slack.
A
So,
let's
roll
the
video
now
so
I
can
click
on
that
and
open
up
the
notebook
with
the
github.dev
web
editor
at
this
point,
I,
can
you
know,
look
at
all
of
the
awesome
notebooks
that
are
part
of
that
project.
In
addition
to
that,
I
can
you
know,
have
access
to
this
Rich
notebook,
editing
experience,
so
I
can
make
changes
to
the
cells
or
I
can
just
explore
this
notebook
now.
A
What
we're
adding
support
for
is
the
ability
to
also
run
this
notebook
directly
from
the
web
editor
by
connecting
it
to
a
code
space,
so
you
can
specify
exactly
how
much
compute
you
need
and
then
we'll
create
a
backing
code.
Space
start
the
Jupiter
server
attach
a
kernel
for
you,
so
that
you
can
run
the
cells
right
in
this
editor
without
losing
any
contacts.
A
A
The
awesome
part
is
this:
entire
repo
is
backed
by
a
code
space
at
this
point,
which
means
I
can
actually
head
to
a
different
notebook
in
the
same
project
and
again
hit
run
all
to
explore
all
of
the
other
notebooks
that
you
know
are
part
of
that
project.
So,
as
I
said,
this
experience
is
coming
soon.
A
However,
one
thing
that
was
missing,
though,
was
that
there
was
no
Wi-Fi
on
the
plane,
which
means
I
was
not
able
to
test
my
code
spaces
demos
with
no
internet
connection.
Now
this
is
where
the
dev
containers
extension
was
extremely
helpful.
To
get
me
up
and
running
with
a
very
code
spaces
like
experience,
but
on
my
local
machine
instead.
A
Now
the
death
containers
extension
follows
the
same
Dev
container
spec,
which
means
once
you
have
that
configuration
created
in
your
repo.
You
can
not
only
use
it
to
create
a
fully
configured
code
space
in
the
cloud,
but
you
can
also
use
it
to
create
a
fully
featured
Dev
environment
on
your
local
machine,
to
highlight
a
few
improvements
that
we
are
doing
for
that
Dev
container
spec.
A
First
up,
we've
opened
the
meta
Dev
container
spec
so
that
it
can
be
used
across
other
editors
and
Beyond
development
scenarios,
including
CI
CD,
test
Automation
in
board,
the
dev
containers
extension
and
the
code
space
extension
both
follow
the
same
Dev
container
CLI,
and
the
great
news
is
that
we
are
also
open
sourcing,
that
Dev
container
CLI
so
that
any
tool
or
any
one
can
contribute
to
it
to
make
it
work
for
your
own
needs.
In
addition
to
that,
we
also
saw
features
when
we
looked
at
that
Cuda
feature
in
my
previous
demo.
A
So
you
can
not
only
take
advantage
of
all
of
the
existing
features
that
we
support
you,
but
you
can
also
create
your
own
via
oci
artifacts
that
are
very
much
tailored
to
your
own
needs
now.
I
know
this
is
a
lot
of
information,
so
if
you
type
in
containers.dev
in
your
browser,
you'll
get
all
of
this
information.
In
addition
to
that,
we're
super
excited
to
just
you
know,
get
feedback
on
all
of
these
things
that
are
open
source
now.
A
So,
if
you
would
like
to
collaborate
with
us,
please
join
the
dev
containers,
Community
slack
Channel,
and
you
should
have
all
of
the
information
right
there
all
right
now,
as
we
near
the
end
of
my
talk
and
since
I
started
with
AKA
analogy.
I
would
love
to
end
with
one
too,
as
most
of
you
might
know,
already
code
spaces
are
rolling
out
SGA
this
week
for
all
users
and
as
part
of
that,
our
cake
to
you
is
this
free
included,
monthly
usage
across
free
and
pro
plans
that
I
highly
encourage
everyone
to
take
advantage
of.
A
With
that.
There
are
some
next
steps
that
I
want
to
end
you
with.
There
is
a
getting
started
guide
that
that
includes
all
of
these
repo
links,
so
definitely
check
that
out.
If
you
would
like
to
sign
up
to
get
GPU
access
in
preview,
that's
the
short
link
to
fill
out
a
form.
And
lastly,
if
you
just
want
to
stay
up
to
date
with
everything
that
we
are
doing
with
code
spaces,
you
can
subscribe
to
our
roadmap
with
that.
I
definitely
want
to
thank
everyone
to
be
here
and
have
a
great
rest
of
your
day.