1
00:00:00,090 --> 00:00:03,090
-: This lesson is all about the OpenAI Playground.

2
00:00:03,090 --> 00:00:04,200
So let's jump right in.

3
00:00:04,200 --> 00:00:06,720
Here, if you search for the OpenAI Playground,

4
00:00:06,720 --> 00:00:09,840
you will find the playground for OpenAI.

5
00:00:09,840 --> 00:00:12,750
platform.openai.com/playground.

6
00:00:12,750 --> 00:00:15,450
Now the default mode is Assistants,

7
00:00:15,450 --> 00:00:18,840
which is arguably the most exciting use case

8
00:00:18,840 --> 00:00:19,860
for the Playground.

9
00:00:19,860 --> 00:00:23,700
But let's go ahead and start here at this legacy version

10
00:00:23,700 --> 00:00:25,740
to go over all the different features

11
00:00:25,740 --> 00:00:27,330
that are available here in the Playground.

12
00:00:27,330 --> 00:00:29,070
Here in the legacy version,

13
00:00:29,070 --> 00:00:31,530
this used to be where you would go

14
00:00:31,530 --> 00:00:33,600
if ChatGPT was having trouble

15
00:00:33,600 --> 00:00:36,060
or you're running out of messages, or having some errors.

16
00:00:36,060 --> 00:00:38,430
The Playground was the place where you would go

17
00:00:38,430 --> 00:00:42,060
to continue using ChatGPT within a different interface

18
00:00:42,060 --> 00:00:44,190
that would usually sidestep the errors

19
00:00:44,190 --> 00:00:46,800
that you could get in the conventional interface.

20
00:00:46,800 --> 00:00:48,630
So here, you can use ChatGPT

21
00:00:48,630 --> 00:00:51,030
just like you would in the other interface.

22
00:00:51,030 --> 00:00:53,310
It says, "Write a tagline for an ice cream shop."

23
00:00:53,310 --> 00:00:55,740
Sure. So we're gonna go ahead and hit Submit.

24
00:00:55,740 --> 00:00:57,157
And here, we have our answer.

25
00:00:57,157 --> 00:00:58,980
"Scoops of happiness in every cone!"

26
00:00:58,980 --> 00:01:02,370
So this is just another interface where you can use ChatGPT.

27
00:01:02,370 --> 00:01:04,290
We select the different models here,

28
00:01:04,290 --> 00:01:08,010
and we can go ahead and tweak ChatGPT even further

29
00:01:08,010 --> 00:01:09,900
with this bar here on the right.

30
00:01:09,900 --> 00:01:13,860
So you have a little bit more control over the length,

31
00:01:13,860 --> 00:01:14,880
the temperature,

32
00:01:14,880 --> 00:01:19,110
and these different frequency, and presence penalties here.

33
00:01:19,110 --> 00:01:20,760
You're able to share here.

34
00:01:20,760 --> 00:01:24,060
You can view the code here, and you can save here.

35
00:01:24,060 --> 00:01:26,190
Let's go ahead and jump into Chat mode.

36
00:01:26,190 --> 00:01:29,010
This is a more user-friendly-looking chat interface

37
00:01:29,010 --> 00:01:30,450
where you type in a query here,

38
00:01:30,450 --> 00:01:32,160
and you'll get a response here.

39
00:01:32,160 --> 00:01:34,440
Please write me a recipe for a breakfast burrito.

40
00:01:34,440 --> 00:01:35,790
We're gonna go ahead and hit Submit.

41
00:01:35,790 --> 00:01:37,350
And here, we have our recipe.

42
00:01:37,350 --> 00:01:39,960
So this is, again, like using ChatGPT.

43
00:01:39,960 --> 00:01:41,820
The general system prompt here

44
00:01:41,820 --> 00:01:43,980
is that you are a helpful assistant.

45
00:01:43,980 --> 00:01:45,600
Of course, you can change this

46
00:01:45,600 --> 00:01:48,600
and give it some more specific parameters.

47
00:01:48,600 --> 00:01:50,130
You are a head chef at a restaurant.

48
00:01:50,130 --> 00:01:53,250
Now let's go ahead and ask this question again.

49
00:01:53,250 --> 00:01:58,140
So here, we have a slightly upgraded version of our recipe.

50
00:01:58,140 --> 00:02:01,950
Now, finally, we can jump over to the Assistant.

51
00:02:01,950 --> 00:02:05,280
This here is the most exciting use case,

52
00:02:05,280 --> 00:02:06,900
but also the most in-depth.

53
00:02:06,900 --> 00:02:09,539
This is similar to creating a custom GPT,

54
00:02:09,539 --> 00:02:11,520
just a little bit more advanced.

55
00:02:11,520 --> 00:02:13,530
Here, you create the name

56
00:02:13,530 --> 00:02:16,020
of the assistant you're trying to create,

57
00:02:16,020 --> 00:02:18,630
and then you give it its set of custom instructions.

58
00:02:18,630 --> 00:02:21,480
You select the different model here that you want to use,

59
00:02:21,480 --> 00:02:23,520
and then you can add different functions.

60
00:02:23,520 --> 00:02:25,620
Let's say a Code interpreter, Retrieval,

61
00:02:25,620 --> 00:02:28,050
and you can upload some custom data here.

62
00:02:28,050 --> 00:02:29,640
If you wanna test your Assistant,

63
00:02:29,640 --> 00:02:32,460
let's go ahead and type in a query here.

64
00:02:32,460 --> 00:02:35,700
So the Assistant here is creating images

65
00:02:35,700 --> 00:02:36,900
from a video script.

66
00:02:36,900 --> 00:02:39,750
So let's go ahead and paste in an example video script here.

67
00:02:39,750 --> 00:02:43,200
Hit Run, and we see that I had instructions labeled one,

68
00:02:43,200 --> 00:02:46,320
two, and three, and we are running these instructions

69
00:02:46,320 --> 00:02:47,340
to get our answer.

70
00:02:47,340 --> 00:02:51,720
Now this GPT was to create image prompts based on a script,

71
00:02:51,720 --> 00:02:52,770
and here we have it.

72
00:02:52,770 --> 00:02:54,900
We have four different image prompts

73
00:02:54,900 --> 00:02:57,540
that are based on the script that I uploaded,

74
00:02:57,540 --> 00:03:00,510
that we can use to put in a image generator

75
00:03:00,510 --> 00:03:03,240
like Midjourney or even OpenAI's DALL-E.

76
00:03:03,240 --> 00:03:06,270
This menu bar on the left lets you click from Playground

77
00:03:06,270 --> 00:03:11,270
to Assistants, Fine-tuning, API keys, Files, Usage,

78
00:03:12,540 --> 00:03:15,660
which has to do with how many API calls you've made.

79
00:03:15,660 --> 00:03:18,750
And Organization, can go into Teams,

80
00:03:18,750 --> 00:03:20,910
Limits, Billing, Profile,

81
00:03:20,910 --> 00:03:23,430
and some Documentation here on the bottom.

82
00:03:23,430 --> 00:03:25,080
You can click onto all products,

83
00:03:25,080 --> 00:03:27,780
check out your personal information, even log out,

84
00:03:27,780 --> 00:03:28,950
and ask for help,

85
00:03:28,950 --> 00:03:31,590
which will prompt a little Help menu here on the right.

86
00:03:31,590 --> 00:03:34,020
Basically, you'll use this bar here on the left

87
00:03:34,020 --> 00:03:35,730
if you need to update your billing,

88
00:03:35,730 --> 00:03:37,650
if you want to update your usage,

89
00:03:37,650 --> 00:03:39,840
if you want to set limits on your usage

90
00:03:39,840 --> 00:03:42,300
to make sure that you don't spend too much money

91
00:03:42,300 --> 00:03:43,620
on your API calls.

92
00:03:43,620 --> 00:03:47,100
And if you're unfamiliar, if you use the OpenAI API,

93
00:03:47,100 --> 00:03:49,350
basically, it costs you a little bit of money

94
00:03:49,350 --> 00:03:50,220
every time you use it.

95
00:03:50,220 --> 00:03:53,580
So I encourage setting a limit to make sure

96
00:03:53,580 --> 00:03:55,650
that you don't spend more than you were looking to.

97
00:03:55,650 --> 00:03:58,800
We will do a deep dive into the Playground

98
00:03:58,800 --> 00:04:02,010
and all of its features in future lessons to come.

99
00:04:02,010 --> 00:04:04,050
I just wanted to do a quick overview

100
00:04:04,050 --> 00:04:06,423
so you can see what's available to you.

