1
00:00:00,658 --> 00:00:04,614
In a co-lab with this data and
these parameters, using a GPU,

2
00:00:04,614 --> 00:00:08,134
it typically takes about 20
minutes to train a model.

3
00:00:08,134 --> 00:00:12,580
Once it's done, try a a seed sentence and
get it to give you 100 words.

4
00:00:12,580 --> 00:00:15,610
Note that there are no line
breaks in the prediction, so

5
00:00:15,610 --> 00:00:19,418
you'll have to add them manually to
turn the word stream into poetry.

6
00:00:19,418 --> 00:00:21,038
Here's a simple example.

7
00:00:21,038 --> 00:00:23,940
I used a famous quote from
a very famous movie and

8
00:00:23,940 --> 00:00:26,126
let's see if you can recognize it.

9
00:00:26,126 --> 00:00:29,953
And I tried that to see what type of
poem it would give me and I got this.

10
00:00:29,953 --> 00:00:34,226
And it almost make sense,
[LAUGH] but that's poetry for you.

11
00:00:34,226 --> 00:00:37,491
Help me Obi-Wan Kenobi,
you're my only hope.

12
00:00:37,491 --> 00:00:42,292
My dear and hope as I did fly with
its flavors, along with all its joys.

13
00:00:42,292 --> 00:00:47,347
But sure I will build,
love you still, gold it did join.

14
00:00:47,347 --> 00:00:51,099
Do mans run away cross our country,
our wedding I was down to.

15
00:00:51,099 --> 00:00:53,575
Off holyhead wished myself
down among the pigs.

16
00:00:53,575 --> 00:00:56,531
Played some hearty Riggs, me embarrass.

17
00:00:56,531 --> 00:01:01,093
Find me brother,
me chambers she gave me her story.

18
00:01:01,093 --> 00:01:03,374
Be Irishmen to greet you lovely Molly.

19
00:01:03,374 --> 00:01:07,223
Gone away from me home,
home to leave the old tin cans.

20
00:01:07,223 --> 00:01:09,554
The foreman's chain once was shining.

21
00:01:09,554 --> 00:01:13,049
Sky above I think I love.