title
stringlengths 12
112
| published
stringlengths 19
23
| url
stringlengths 28
28
| video_id
stringlengths 11
11
| channel_id
stringclasses 5
values | id
stringlengths 16
31
| text
stringlengths 0
596
| start
float64 0
37.8k
| end
float64 2.18
37.8k
|
|---|---|---|---|---|---|---|---|---|
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1690.0
|
to figure that out
| 1,690 | 1,694 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1692.0
|
what we do is write
| 1,692 | 1,696 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1694.0
|
let me do this
| 1,694 | 1,698 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1696.0
|
torch
| 1,696 | 1,700 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1698.0
|
device CUDA.
| 1,698 | 1,702 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1700.0
|
So this is we say
| 1,700 | 1,704 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1702.0
|
we want to use a CUDA enabled GPU
| 1,702 | 1,706 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1704.0
|
if torch
| 1,704 | 1,708 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1706.0
|
dot CUDA
| 1,706 | 1,710 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1708.0
|
is available.
| 1,708 | 1,714 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1712.0
|
So this will check our
| 1,712 | 1,716 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1714.0
|
environment and check if we have
| 1,714 | 1,718 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1716.0
|
a CUDA enabled GPU.
| 1,716 | 1,720 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1718.0
|
If it isn't available
| 1,718 | 1,722 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1720.0
|
we want to use a
| 1,720 | 1,724 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1722.0
|
torch device
| 1,722 | 1,726 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1724.0
|
CPU.
| 1,724 | 1,728 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1726.0
|
So run that and see
| 1,726 | 1,730 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1728.0
|
for me I have a CUDA enabled GPU
| 1,728 | 1,732 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1730.0
|
so it comes up with this.
| 1,730 | 1,734 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1732.0
|
So we saw that in
| 1,732 | 1,736 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1734.0
|
device. And then
| 1,734 | 1,738 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1736.0
|
what we can do is move
| 1,736 | 1,740 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1738.0
|
our model and also move our tensors
| 1,738 | 1,742 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1740.0
|
later on to that
| 1,740 | 1,744 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1742.0
|
device for training.
| 1,742 | 1,746 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1744.0
|
So we just write model
| 1,744 | 1,748 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1746.0
|
to device.
| 1,746 | 1,750 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1748.0
|
And we'll get
| 1,748 | 1,752 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1750.0
|
a lot of output from that.
| 1,750 | 1,754 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1752.0
|
Just ignore that we don't need to worry about
| 1,752 | 1,756 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1754.0
|
it.
| 1,754 | 1,758 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1756.0
|
And we can also
| 1,756 | 1,760 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1758.0
|
activate our models training mode
| 1,758 | 1,762 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1760.0
|
like that.
| 1,760 | 1,764 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1762.0
|
OK. So
| 1,762 | 1,766 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1764.0
|
we've moved our model over to
| 1,764 | 1,768 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1766.0
|
GPU activate training mode.
| 1,766 | 1,770 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1768.0
|
Now what we need to do
| 1,768 | 1,772 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1770.0
|
is initialize our optimizer.
| 1,770 | 1,774 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1772.0
|
So we're going to be using
| 1,772 | 1,776 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1774.0
|
Adam with weighted decay for our
| 1,774 | 1,778 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1776.0
|
optimizer. So
| 1,776 | 1,780 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1778.0
|
to use that we need to import
| 1,778 | 1,782 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1780.0
|
it from transformers.
| 1,780 | 1,784 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1782.0
|
So from transformers
| 1,782 | 1,786 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1784.0
|
import
| 1,784 | 1,788 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1786.0
|
AdamW.
| 1,786 | 1,792 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1790.0
|
And we
| 1,790 | 1,794 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1792.0
|
initialize the
| 1,792 | 1,796 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1794.0
|
optimizer like
| 1,794 | 1,798 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1796.0
|
this. So we
| 1,796 | 1,800 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1798.0
|
AdamW we pass our model
| 1,798 | 1,802 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1800.0
|
parameters. And we
| 1,800 | 1,804 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1802.0
|
also want to pass the learning rate
| 1,802 | 1,806 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1804.0
|
which is going to be
| 1,804 | 1,808 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1806.0
|
5e to the minus 5.
| 1,806 | 1,810 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1808.0
|
5e to the minus
| 1,808 | 1,812 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1810.0
|
5. OK that's a
| 1,810 | 1,814 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1812.0
|
pretty common one for training transformers.
| 1,812 | 1,816 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1814.0
|
And
| 1,814 | 1,818 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1816.0
|
that
| 1,816 | 1,820 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1818.0
|
looks pretty good to me.
| 1,818 | 1,822 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1820.0
|
So now we can begin
| 1,820 | 1,824 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1822.0
|
our training loop. So
| 1,822 | 1,826 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1824.0
|
first I want to
| 1,824 | 1,828 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1826.0
|
import something called TQDM.
| 1,826 | 1,830 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1828.0
|
Now this is purely
| 1,828 | 1,832 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1830.0
|
for aesthetics. We don't need it
| 1,830 | 1,834 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1832.0
|
for training. This is so
| 1,832 | 1,836 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1834.0
|
that we see a little progress bar during
| 1,834 | 1,838 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1836.0
|
training otherwise we don't see anything.
| 1,836 | 1,840 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1838.0
|
So I just want to include
| 1,838 | 1,842 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1840.0
|
that so we can actually see what is going on.
| 1,840 | 1,844 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1842.0
|
So from TQDM
| 1,842 | 1,846 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1844.0
|
import TQDM. So this is optional
| 1,844 | 1,848 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1846.0
|
we don't need to include it. It's up to you.
| 1,846 | 1,850 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1848.0
|
But I would recommend it.
| 1,848 | 1,852 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1850.0
|
We'll train
| 1,850 | 1,854 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1852.0
|
for
| 1,852 | 1,856 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1854.0
|
let's go with 2
| 1,854 | 1,858 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1856.0
|
epochs.
| 1,856 | 1,860 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1858.0
|
Again we don't want to train transformers
| 1,858 | 1,862 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1860.0
|
too much because they will easily
| 1,860 | 1,864 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1862.0
|
over fit. And to be honest they'll probably
| 1,862 | 1,866 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1864.0
|
over fit on this dataset because it's very
| 1,864 | 1,868 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1866.0
|
small.
| 1,866 | 1,870 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1868.0
|
But that's fine. We just want
| 1,868 | 1,872 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1870.0
|
to use this as an example.
| 1,870 | 1,874 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1872.0
|
So we're going
| 1,872 | 1,876 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1874.0
|
to train for 2 epochs.
| 1,874 | 1,878 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1876.0
|
And because we're using TQDM
| 1,876 | 1,880 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1878.0
|
we want to set up our
| 1,878 | 1,882 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1880.0
|
training loop like this. So we wrap
| 1,880 | 1,884 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1882.0
|
it within a TQDM
| 1,882 | 1,886 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1884.0
|
instance.
| 1,884 | 1,888 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1886.0
|
And all we do here is pass our
| 1,886 | 1,890 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1888.0
|
data loader. So we
| 1,888 | 1,892 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1890.0
|
create that up here.
| 1,890 | 1,894 |
Training BERT #4 - Train With Next Sentence Prediction (NSP)
|
2021-05-27 16:15:39 UTC
|
https://youtu.be/x1lAcT3xl5M
|
x1lAcT3xl5M
|
UCv83tO5cePwHMt1952IVVHw
|
x1lAcT3xl5M-t1892.0
|
That's our PyTorch data loader.
| 1,892 | 1,896 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.