subtitles/en/00_welcome-to-the-hugging-face-course.srt

1
00:00:05,850 --> 00:00:07,713
Welcome to the Hugging Face Course.

2
00:00:08,550 --> 00:00:10,320
This course has been designed to teach you

3
00:00:10,320 --> 00:00:12,750
all about the Hugging Face ecosystem,

4
00:00:12,750 --> 00:00:14,700
how to use the dataset and model hub

5
00:00:14,700 --> 00:00:16,803
as well as all our open-source libraries.

6
00:00:18,300 --> 00:00:19,950
Here is the Table of Contents.

7
00:00:19,950 --> 00:00:22,770
As you can see, it's
divided in three sections

8
00:00:22,770 --> 00:00:25,110
which become progressively more advanced.

9
00:00:25,110 --> 00:00:28,500
At this stage, the first two
sections have been released.

10
00:00:28,500 --> 00:00:30,120
So first, we'll teach you the basics

11
00:00:30,120 --> 00:00:32,250
of how to use a Transformer model,

12
00:00:32,250 --> 00:00:34,230
fine-tune it on your own data set

13
00:00:34,230 --> 00:00:36,960
and share the result with the community.

14
00:00:36,960 --> 00:00:39,420
So second, we'll dive
deeper into our libraries

15
00:00:39,420 --> 00:00:42,360
and teach you how to tackle any NLP task.

16
00:00:42,360 --> 00:00:44,430
We're actively working on the last one

17
00:00:44,430 --> 00:00:47,280
and hope to have it ready for
you for the spring of 2022.

18
00:00:48,510 --> 00:00:50,880
The first chapter requires
no technical knowledge

19
00:00:50,880 --> 00:00:52,320
and is a good introduction to learn

20
00:00:52,320 --> 00:00:54,180
what Transformers models can do

21
00:00:54,180 --> 00:00:56,883
and how it could be of use
to you or your company.

22
00:00:58,050 --> 00:01:01,110
The next chapters require
a good knowledge of Python

23
00:01:01,110 --> 00:01:02,130
and some basic knowledge of

24
00:01:02,130 --> 00:01:04,350
Machine Learning and Deep Learning.

25
00:01:04,350 --> 00:01:07,110
If you don't know what a
training and validation set are

26
00:01:07,110 --> 00:01:09,360
or what gradient decent means,

27
00:01:09,360 --> 00:01:11,340
you should look at an introductory course

28
00:01:11,340 --> 00:01:14,863
such as the ones published by
deeplearning.ai or fast.ai.

29
00:01:16,200 --> 00:01:17,910
It's also best if you have some basics

30
00:01:17,910 --> 00:01:21,150
in one Deep Learning Framework,
PyTorch or TensorFlow.

31
00:01:21,150 --> 00:01:23,520
Each part of the material
introduced in this course

32
00:01:23,520 --> 00:01:25,590
has a version in both those frameworks,

33
00:01:25,590 --> 00:01:26,730
so you will be able to pick the one

34
00:01:26,730 --> 00:01:28,230
you are most comfortable with.

35
00:01:29,550 --> 00:01:31,740
This is the team that
developed this course.

36
00:01:31,740 --> 00:01:33,120
I'll now let each of the speakers

37
00:01:33,120 --> 00:01:34,570
introduce themselves briefly.

38
00:01:37,230 --> 00:01:38,880
- Hi, my name is Matthew,

39
00:01:38,880 --> 00:01:41,610
and I'm a Machine Learning
Engineer at Hugging Face.

40
00:01:41,610 --> 00:01:43,200
I work on the open-source team

41
00:01:43,200 --> 00:01:45,180
and I'm responsible for
maintaining particularly

42
00:01:45,180 --> 00:01:47,280
the TensorFlow code there.

43
00:01:47,280 --> 00:01:50,130
Previously, I was a Machine
Learning Engineer at Parsley,

44
00:01:50,130 --> 00:01:52,620
who've recently been
acquired by Automatic,

45
00:01:52,620 --> 00:01:54,210
and I was a postdoctoral researcher

46
00:01:54,210 --> 00:01:57,000
before that at Trinity
College, Dublin in Ireland

47
00:01:57,000 --> 00:02:00,093
working on computational
genetics and retinal disease.

48
00:02:02,400 --> 00:02:03,870
- Hi, I'm Lysandre.

49
00:02:03,870 --> 00:02:05,640
I'm a Machine Learning
Engineer at Hugging Face

50
00:02:05,640 --> 00:02:08,700
and I'm specifically part
of the open-source team.

51
00:02:08,700 --> 00:02:10,890
I've been at Hugging
Face for a few years now

52
00:02:10,890 --> 00:02:12,300
and alongside my team members,

53
00:02:12,300 --> 00:02:13,890
I've been working on most of the tools

54
00:02:13,890 --> 00:02:15,790
that you'll get to see in this course.

55
00:02:18,270 --> 00:02:20,130
- Hi, I'm Sylvain.

56
00:02:20,130 --> 00:02:22,140
I'm a Research Engineer at Hugging Face

57
00:02:22,140 --> 00:02:25,830
and one of the main maintainers
of the Transformers Library.

58
00:02:25,830 --> 00:02:28,110
Previously, I worked at fast.ai

59
00:02:28,110 --> 00:02:30,420
where I helped develop the fast.ai Library

60
00:02:30,420 --> 00:02:32,220
as well as the online book.

61
00:02:32,220 --> 00:02:35,340
Before that, I was a math
and computer science teacher

62
00:02:35,340 --> 00:02:36,173
in France.

63
00:02:38,550 --> 00:02:41,340
- Hi, my name is Sasha and I'm
a Researcher at Hugging Face,

64
00:02:41,340 --> 00:02:42,420
working on the ethical,

65
00:02:42,420 --> 00:02:46,230
environmental and social impacts
of machine learning models.

66
00:02:46,230 --> 00:02:49,020
Previously, I was a
postdoctoral researcher at Mila,

67
00:02:49,020 --> 00:02:50,400
University in Montreal

68
00:02:50,400 --> 00:02:53,040
and I also worked as an
Applied AI Researcher

69
00:02:53,040 --> 00:02:55,140
for the United Nations Global Pulse.

70
00:02:55,140 --> 00:02:57,300
I've been involved in
projects such as CodeCarbon

71
00:02:57,300 --> 00:02:59,790
and the Machine Learning
Impacts Calculator

72
00:02:59,790 --> 00:03:02,390
to measure the carbon
footprint of machine learning.

73
00:03:05,160 --> 00:03:07,650
- Hi, I'm Merve and I'm
a Developer Advocate

74
00:03:07,650 --> 00:03:09,390
at Hugging Face.

75
00:03:09,390 --> 00:03:12,480
Previously, I was working as
a Machine Learning Engineer

76
00:03:12,480 --> 00:03:15,360
building NLP tools and chat bots.

77
00:03:15,360 --> 00:03:17,670
Currently, I'm working to improve the hub

78
00:03:17,670 --> 00:03:19,563
and democratize machine learning.

79
00:03:22,140 --> 00:03:23,670
- Hello everyone.

80
00:03:23,670 --> 00:03:27,210
My name is Lucile and I'm
a Machine Learning Engineer

81
00:03:27,210 --> 00:03:28,353
at Hugging Face.

82
00:03:29,580 --> 00:03:32,550
To tell you in two sentences who I am,

83
00:03:32,550 --> 00:03:35,590
I work on the development and
support of open-source tools

84
00:03:36,600 --> 00:03:39,595
and I also participate in
several research project

85
00:03:39,595 --> 00:03:41,795
in the field of Natural
Language Processing.

86
00:03:44,610 --> 00:03:45,540
- Good day there.

87
00:03:45,540 --> 00:03:47,550
I'm Lewis and I'm a
Machine Learning Engineer

88
00:03:47,550 --> 00:03:50,130
in the open-source team at Hugging Face.

89
00:03:50,130 --> 00:03:53,490
I'm passionate about developing
tools for the NLP community

90
00:03:53,490 --> 00:03:55,050
and you'll see me at
many of Hugging Face's

91
00:03:55,050 --> 00:03:56,910
outreach activities.

92
00:03:56,910 --> 00:03:58,470
Before joining Hugging Face,

93
00:03:58,470 --> 00:03:59,790
I spent several years developing

94
00:03:59,790 --> 00:04:01,860
machine learning applications for startups

95
00:04:01,860 --> 00:04:04,230
and enterprises in the domains of NLP,

96
00:04:04,230 --> 00:04:07,260
topological data analysis and time series.

97
00:04:07,260 --> 00:04:10,110
In a former life, I was
a theoretical physicist,

98
00:04:10,110 --> 00:04:11,760
where I researched particle collisions

99
00:04:11,760 --> 00:04:13,560
at the Large Hadron Collider and so.

100
00:04:15,900 --> 00:04:18,450
- Hey, I'm Leandro and I'm
a Machine Learning Engineer

101
00:04:18,450 --> 00:04:21,030
in the open-source team at Hugging Face.

102
00:04:21,030 --> 00:04:23,460
Before joining Hugging Face,
I worked as a Data Scientist

103
00:04:23,460 --> 00:04:26,733
in Switzerland and have taught
Data Science at University.