Skip to content

Commit

Permalink
reviewed ep.58 (#405)
Browse files Browse the repository at this point in the history
  • Loading branch information
tyisme614 authored Dec 12, 2022
1 parent 4dfa823 commit 87b77eb
Showing 1 changed file with 29 additions and 29 deletions.
58 changes: 29 additions & 29 deletions subtitles/zh-CN/58_what-is-domain-adaptation.srt
Original file line number Diff line number Diff line change
Expand Up @@ -15,7 +15,7 @@

4
00:00:05,910 --> 00:00:07,923
- 什么是领域适应
- 什么是域适配
- What is domain adaptation?

5
Expand All @@ -25,12 +25,12 @@ When fine-tuning a pre-trained model on a new dataset,

6
00:00:12,540 --> 00:00:15,480
我们获得的微调模型将做出预测
我们适配新的数据集所获得的微调模型
the fine-tuned model we obtain will make predictions

7
00:00:15,480 --> 00:00:17,433
适应这个新数据集
将做出预测
that are attuned to this new dataset.

8
Expand All @@ -40,47 +40,47 @@ When the two models are trained with the same task,

9
00:00:21,840 --> 00:00:25,320
然后我们可以比较他们对相同输入的预测
我们可以使用相同的输入比较他们的预测结果
we can then compare their predictions on the same input.

10
00:00:25,320 --> 00:00:27,870
两个模型的预测会有所不同
两个模型的预测结果
The predictions of the two models will be different

11
00:00:27,870 --> 00:00:29,790
以反映差异的方式
会以一种方式反映
in a way that reflects the differences

12
00:00:29,790 --> 00:00:31,680
在两个数据集之间,
两个数据集之间的差别
between the two datasets,

13
00:00:31,680 --> 00:00:34,053
我们称之为领域适应的现象
就是我们称之为域适配的现象
a phenomenon we call domain adaptation.

14
00:00:35,310 --> 00:00:38,640
让我们看一个带有掩码语言建模的例子
让我们通过带有版本微调
Let's look at an example with masked language modeling

15
00:00:38,640 --> 00:00:41,910
通过比较预训练的 DistilBERT 模型的输出
比较预训练的 DistilBERT 模型的输出
by comparing the outputs of the pre-trained DistilBERT model

16
00:00:41,910 --> 00:00:43,080
版本微调
看一个和掩码语言建模相关的例子
with the version fine-tuned

17
00:00:43,080 --> 00:00:45,273
在课程的第 7 章中,链接如下。
该内容在课程的第 7 章中,链接如下。
in chapter 7 of the course, linked below.

18
Expand All @@ -100,27 +100,27 @@ has its first two predictions linked to cinema.

21
00:00:54,390 --> 00:00:57,210
由于它在电影评论数据集上进行了微调
由于它是基于电影评论数据集上进行了微调
Since it was fine-tuned on a movie reviews dataset,

22
00:00:57,210 --> 00:00:58,680
看到是完全正常的
因此它像这样调整它的推荐结果
it's perfectly normal to see

23
00:00:58,680 --> 00:01:01,440
它像这样调整了它的建议
是完全正常的
it adapted its suggestions like this.

24
00:01:01,440 --> 00:01:03,090
注意它如何保持相同的预测
注意它作为之后的预训练模型
Notice how it keeps the same prediction

25
00:01:03,090 --> 00:01:05,220
作为之后的预训练模型
如何保持相同的预测
as the pre-trained model afterward.

26
Expand All @@ -130,7 +130,7 @@ Even if the fine-tuned model adapts to the new dataset,

27
00:01:08,100 --> 00:01:10,450
它不会忘记预先训练的内容
它不会遗失预先训练的内容
it's not forgetting what it was pre-trained on.

28
Expand All @@ -140,57 +140,57 @@ This is another example on a translation task.

29
00:01:14,220 --> 00:01:17,310
最重要的是,我们使用预训练的法语 / 英语模型,
在上面的代码里,我们使用预训练的法语 / 英语模型,
On top, we use a pre-trained French/English model,

30
00:01:17,310 --> 00:01:21,330
在底部,我们在第 7 章中微调的版本。
在下面的代码里,是我们在第 7 章中微调的版本。
and at the bottom, the version we fine-tuned in chapter 7.

31
00:01:21,330 --> 00:01:23,610
顶级模型在大量文本上进行了预训练
上面的模型在大量文本上进行了预训练
The top model is pre-trained on lots of texts,

32
00:01:23,610 --> 00:01:25,170
并留下技术英语术语
并保留了英文中的技术术语
and leaves technical English terms,

33
00:01:25,170 --> 00:01:28,350
像插件和电子邮件,翻译不变
像 plugin 和 email 这样的单词,是不会被翻译的
like plugin and email, unchanged in the translation.

34
00:01:28,350 --> 00:01:31,350
两者都被法国人完全理解
法国用户都可以很好地理解两者
Both are perfectly understood by French people.

35
00:01:31,350 --> 00:01:33,780
为微调选择的数据集是一个数据集
为微调模型选择的数据集是
The dataset picked for the fine-tuning is a dataset

36
00:01:33,780 --> 00:01:36,660
特别注意的技术文本
一个包含技术文本的数据集
of technical texts where special attention was picked

37
00:01:36,660 --> 00:01:39,150
用法语翻译一切
其中特别将所有内容都翻译为法语
on translating everything in French.

38
00:01:39,150 --> 00:01:42,090
结果,经过微调的模型选择了那个习惯
结果,经过微调的模型适应了该特征
As a result, the fine-tuned model picked that habit

39
00:01:42,090 --> 00:01:44,193
并翻译了插件和电子邮件
并翻译了 plugin 和 email 两个词
and translated both plugin and email.

40
Expand Down

0 comments on commit 87b77eb

Please sign in to comment.