1
00:00:00,000 --> 00:00:00,000
Hello guys.

2
00:00:00,000 --> 00:00:03,000
So we are going to continue the discussion with respect to Lang Chain.

3
00:00:03,000 --> 00:00:06,000
And now we are going to get started to work with hugging Face.

4
00:00:06,000 --> 00:00:11,000
Already in our previous videos we have seen how we can use hugging face embeddings.

5
00:00:11,000 --> 00:00:15,000
But uh, in this module itself I will be talking more about hugging face.

6
00:00:15,000 --> 00:00:17,000
And uh, you can see over here.

7
00:00:17,000 --> 00:00:22,000
On May 14th, 2024, there was a article that was published by Hugging Face Itself.

8
00:00:22,000 --> 00:00:26,000
Hugging Face and Lang chain are new partner package in Lang chain okay.

9
00:00:26,000 --> 00:00:29,000
And previously like before May 14th.

10
00:00:29,000 --> 00:00:29,000
Right.

11
00:00:29,000 --> 00:00:36,000
Uh, everybody who really wanted to use LM models from Huggingface, uh, they had different set of

12
00:00:36,000 --> 00:00:41,000
libraries that they usually used from Huggingface itself to call it, and it was really a daunting task.

13
00:00:41,000 --> 00:00:47,000
But by this specific integration of Huggingface and Lang Chain, now it has become really, really easy.

14
00:00:47,000 --> 00:00:51,000
And since our course in our course, we are completely focusing on the Lang chain ecosystem.

15
00:00:51,000 --> 00:00:54,000
So Lang Chain has this hugging face integration also.

16
00:00:54,000 --> 00:00:56,000
So we'll be talking about this as we go ahead.

17
00:00:56,000 --> 00:01:01,000
So first of all, what I really want to do is that I want to go to Hugging face.co.

18
00:01:01,000 --> 00:01:07,000
So please make sure that if you do not have an account go ahead and create an account in hugging face.co.

19
00:01:07,000 --> 00:01:09,000
Now what exactly is hugging face for?

20
00:01:10,000 --> 00:01:11,000
It's an amazing company.

21
00:01:11,000 --> 00:01:12,000
It has.

22
00:01:12,000 --> 00:01:16,000
If you just go ahead and click on the models over here, you'll be able to see it has tons and tons

23
00:01:16,000 --> 00:01:17,000
of models.

24
00:01:17,000 --> 00:01:19,000
It's just not like 1 or 2 model.

25
00:01:19,000 --> 00:01:23,000
It has a huge repository of every kind of models that are probably available.

26
00:01:23,000 --> 00:01:27,000
So here you can see Google Gamma two 9 billion parameter model is also available right.

27
00:01:27,000 --> 00:01:28,000
Let it be open source.

28
00:01:28,000 --> 00:01:29,000
Let it be paid.

29
00:01:29,000 --> 00:01:30,000
Everything is there.

30
00:01:30,000 --> 00:01:34,000
Uh, some of the models like OpenAI models will not be available over here.

31
00:01:34,000 --> 00:01:38,000
But other than that I feel every model is available uh, paid models.

32
00:01:38,000 --> 00:01:43,000
Uh uh, and over here you'll also be able to find both open source and paid models, also paid paid

33
00:01:43,000 --> 00:01:44,000
models itself.

34
00:01:44,000 --> 00:01:50,000
And to access paid models you have to probably use hugging face API endpoints.

35
00:01:50,000 --> 00:01:55,000
You can create your endpoint with respect to any models itself, but there you will definitely require

36
00:01:55,000 --> 00:01:56,000
a credit card and all.

37
00:01:56,000 --> 00:02:00,000
Other than that, uh, hugging face still provides you so many different open source models which you

38
00:02:00,000 --> 00:02:03,000
can completely, easily, uh, use it at any point of time.

39
00:02:03,000 --> 00:02:10,000
So, uh, just to give you an example, let's say I want to probably perform something like image text

40
00:02:10,000 --> 00:02:10,000
to text.

41
00:02:10,000 --> 00:02:14,000
So see over here there are three categories of model that you will be able to see.

42
00:02:14,000 --> 00:02:15,000
One is multi model.

43
00:02:15,000 --> 00:02:21,000
Multi model basically means all those models which will be able to uh work both with text and images.

44
00:02:21,000 --> 00:02:24,000
So here they will be able to understand both text and images.

45
00:02:24,000 --> 00:02:26,000
So some of the model is over here.

46
00:02:26,000 --> 00:02:28,000
Visual question answering document question answering.

47
00:02:28,000 --> 00:02:31,000
Let's consider this particular model like image text to text.

48
00:02:31,000 --> 00:02:35,000
Now we will go ahead and see one example over here Florence two large right.

49
00:02:35,000 --> 00:02:38,000
And here you can actually see this is the model.

50
00:02:38,000 --> 00:02:39,000
All the requirements is basically given.

51
00:02:39,000 --> 00:02:42,000
Code is given how to call this particular model right.

52
00:02:42,000 --> 00:02:47,000
And you can also use this particular model with the help of transformer library which is available in

53
00:02:47,000 --> 00:02:47,000
hugging face.

54
00:02:47,000 --> 00:02:48,000
Okay.

55
00:02:48,000 --> 00:02:54,000
But again as our main aim is that our main aim is to simplify this calls to the LM models that is available

56
00:02:54,000 --> 00:02:55,000
in the hugging face.

57
00:02:55,000 --> 00:02:59,000
And for that we will be using a Lang chain package which is integrated with hugging face.

58
00:02:59,000 --> 00:03:05,000
Okay, other than this, if I go back, you will be also able to see other multiple models like, uh,

59
00:03:05,000 --> 00:03:06,000
computer vision.

60
00:03:06,000 --> 00:03:09,000
In NLP, you have uh, text classification.

61
00:03:09,000 --> 00:03:10,000
You have question answering.

62
00:03:10,000 --> 00:03:14,000
Let's say if I go ahead and click on question answering, we'll be seeing so many different models.

63
00:03:14,000 --> 00:03:16,000
Like like there's tons and tons of model.

64
00:03:16,000 --> 00:03:17,000
Right.

65
00:03:17,000 --> 00:03:22,000
So Hugging Face is one of the most popular company who is specifically working in this field of generative

66
00:03:22,000 --> 00:03:24,000
AI with respect to any open source model.

67
00:03:24,000 --> 00:03:27,000
Also, that comes, first of all, it will be available in Hugging Face and Kaggle.

68
00:03:27,000 --> 00:03:31,000
And there are also other companies like whichever company is basically launching it.

69
00:03:31,000 --> 00:03:35,000
They will also put there in their own website, but at the end of the day, they are going to do the

70
00:03:35,000 --> 00:03:36,000
integration in Huggingface.

71
00:03:36,000 --> 00:03:38,000
They're going to do the integration in long chain.

72
00:03:38,000 --> 00:03:39,000
And all right.

73
00:03:39,000 --> 00:03:41,000
So all this particular models is also available.

74
00:03:41,000 --> 00:03:42,000
You can go ahead and check it out.

75
00:03:42,000 --> 00:03:48,000
Text summarization is there let's say text summarization I want to go ahead and see gamma 29B I'll be

76
00:03:48,000 --> 00:03:48,000
able to see over here.

77
00:03:48,000 --> 00:03:51,000
I'll also able to see or check the inference API.

78
00:03:51,000 --> 00:03:58,000
I'll say, hey, um, my name is Chris and I like something like this.

79
00:03:58,000 --> 00:04:04,000
And if I go ahead and compute it, you will be able to see gamma two model is too large to be loaded.

80
00:04:04,000 --> 00:04:05,000
Uh, please use spaces.

81
00:04:05,000 --> 00:04:05,000
Okay.

82
00:04:05,000 --> 00:04:09,000
Now it is telling you to use spaces because 9 billion parameter is there, right?

83
00:04:09,000 --> 00:04:12,000
But you can go ahead and check out with any other models that you like.

84
00:04:12,000 --> 00:04:15,000
Let's say I want to go ahead and try meta llama 8 billion parameters.

85
00:04:15,000 --> 00:04:17,000
I don't know whether I'll be able to check or not.

86
00:04:17,000 --> 00:04:18,000
So let's see over here.

87
00:04:18,000 --> 00:04:19,000
Also the same issue right.

88
00:04:19,000 --> 00:04:20,000
But smaller model.

89
00:04:20,000 --> 00:04:23,000
You can actually go ahead and check the inferencing.

90
00:04:23,000 --> 00:04:23,000
Okay.

91
00:04:23,000 --> 00:04:29,000
Uh, so uh, what we are basically going to do is that now, I hope you have got an idea with respect

92
00:04:29,000 --> 00:04:35,000
to what hugging face exactly is, you can go ahead and see more about it, the kind of documentation,

93
00:04:35,000 --> 00:04:40,000
amazing posts are there, you know, and uh, slowly we'll be discussing about many things as we go

94
00:04:40,000 --> 00:04:45,000
ahead and how we can actually create an end to end generative AI application with the help of both,

95
00:04:45,000 --> 00:04:46,000
uh, long chain and hugging face.

96
00:04:46,000 --> 00:04:50,000
That is what I'm actually going to discuss in the next video.

97
00:04:50,000 --> 00:04:50,000
Okay.

98
00:04:50,000 --> 00:04:53,000
Uh, if I go ahead and see the documentation page.

99
00:04:53,000 --> 00:04:57,000
So here is what hugging face is basically there in the long chain documentation.

100
00:04:57,000 --> 00:04:59,000
You can see how the installation is basically done.

101
00:04:59,000 --> 00:05:01,000
And then we'll be talking about chat hugging face.

102
00:05:01,000 --> 00:05:03,000
We'll be talking about this.

103
00:05:03,000 --> 00:05:05,000
We'll be talking about hugging face embeddings.

104
00:05:05,000 --> 00:05:05,000
Right.

105
00:05:05,000 --> 00:05:10,000
Hugging face embeddings I think I have already spoken about and uh, some of the examples after doing

106
00:05:10,000 --> 00:05:14,000
this will create an end to end project with the help of hugging face itself.

107
00:05:14,000 --> 00:05:16,000
So, uh, yeah, this was it.

108
00:05:16,000 --> 00:05:21,000
Uh, and you can probably read more about it, you know, so there is a new Python package that is designed

109
00:05:21,000 --> 00:05:24,000
to bring the power of the latest development of hugging face into long chain.

110
00:05:24,000 --> 00:05:25,000
Right.

111
00:05:25,000 --> 00:05:30,000
So yeah, in the next video we will be discussing more about hugging Face and we will be seeing some

112
00:05:30,000 --> 00:05:31,000
practical implementation.

113
00:05:31,000 --> 00:05:34,000
And after that, finally we will implement an end to end project.

114
00:05:34,000 --> 00:05:36,000
So yes, this was it from my side.

115
00:05:36,000 --> 00:05:37,000
I will see you all in the next video.

116
00:05:37,000 --> 00:05:38,000
Thank you.