subtitles/zh-CN/34_the-push-to-hub-api-(tensorflow).srt

1
00:00:00,587 --> 00:00:02,670
（嗖嗖）
(swoosh)

2
00:00:05,100 --> 00:00:07,080
- [旁白] 嗨，本视频将带领大家了解
- [Narrator] Hi, this is going to be a video

3
00:00:07,080 --> 00:00:09,420
push_to_hub API 适用于 
about the push_to_hub API

4
00:00:09,420 --> 00:00:10,670
Tensorflow 和 Keras 的版本。
for Tensorflow and Keras.

5
00:00:11,820 --> 00:00:14,850
那么首先，打开我们的 jupyter。
So, to get started, we'll open up our notebook.

6
00:00:14,850 --> 00:00:16,920
你需要做的第一件事就是
And the first thing you'll need to do is log in to

7
00:00:16,920 --> 00:00:18,170
登录你的 HuggingFace 帐户，
your HuggingFace account,

8
00:00:19,043 --> 00:00:20,663
例如调用 notebook_login 函数。
for example with the notebook login function.

9
00:00:21,570 --> 00:00:24,630
所以要使用它，你只需调用函数，
So to use that, you simply call the function,

10
00:00:24,630 --> 00:00:26,010
随后会弹出窗口。
the popup will emerge.

11
00:00:26,010 --> 00:00:28,800
你需要输入你的用户名和密码，
You will enter your username and password,

12
00:00:28,800 --> 00:00:31,425
这里我使用密码管理器拷贝密码，
which I'm going to pull out of my password manager here,

13
00:00:31,425 --> 00:00:33,108
然后就登录了。
and you log in.

14
00:00:33,108 --> 00:00:35,670
接下来的两个单元格只是
The next two cells are just

15
00:00:35,670 --> 00:00:37,080
为训练做好准备。
getting everything ready for training.

16
00:00:37,080 --> 00:00:38,940
所以我们只是要加载一个数据集，
So we're just going to load a dataset,

17
00:00:38,940 --> 00:00:41,100
我们要对数据集进行分词，
we're going to tokenize that dataset,

18
00:00:41,100 --> 00:00:42,990
然后我们将加载我们的模型并
and then we're going to load our model and compile

19
00:00:42,990 --> 00:00:45,660
使用标准 Adam 优化器进行编译。
it with the standard Adam optimizer.

20
00:00:45,660 --> 00:00:47,560
现在我们会运行这些代码。
So I'm just going to run all of those.

21
00:00:49,830 --> 00:00:52,080
我们等几秒钟，
We'll wait a few seconds,

22
00:00:52,080 --> 00:00:54,280
一切都应该为训练做好准备。
and everything should be ready for training.

23
00:00:57,983 --> 00:00:58,816
好的。
Okay.

24
00:00:58,816 --> 00:01:01,440
所以现在我们准备好训练了。
So now we're ready to train.

25
00:01:01,440 --> 00:01:03,030
我将向你展示两种方法
I'm going to show you the two ways

26
00:01:03,030 --> 00:01:05,130
你可以将你的模型推送到 Hub。
you can push your model to the Hub.

27
00:01:05,130 --> 00:01:08,190
所以第一个是 PushToHubCallback。
So the first is with the PushToHubCallback.

28
00:01:08,190 --> 00:01:10,107
所以 Keras 中的回调
So a callback in Keras

29
00:01:10,107 --> 00:01:13,710
也会在训练过程中被频繁调用。
is a function that's called regularly during training.

30
00:01:13,710 --> 00:01:17,400
你可以将其设置为在一定数量的 step 后或者在每个 epoch 被调用，
You can set it to be called after a certain number of steps,

31
00:01:17,400 --> 00:01:21,427
甚至只是在训练结束时调用一次。
or every epoch, or even just once at the end of training.

32
00:01:21,427 --> 00:01:25,080
所以 Keras 中有很多回调，例如，
So a lot of callbacks in Keras, for example,

33
00:01:25,080 --> 00:01:28,050
在平稳期控制学习率衰减，
control learning rate decaying on plateau,

34
00:01:28,050 --> 00:01:30,047
诸如此类的事情。
and things like that.

35
00:01:30,047 --> 00:01:32,520
所以这个回调，默认情况下，
So this callback, by default,

36
00:01:32,520 --> 00:01:35,760
在每个 epoch 都会将你的模型保存到 Hub 一次。
will save your model to the Hub once every epoch.

37
00:01:35,760 --> 00:01:37,080
这真的非常有用，
And that's really helpful,

38
00:01:37,080 --> 00:01:38,790
特别是如果你的训练时间很长，
especially if your training is very long,

39
00:01:38,790 --> 00:01:40,800
因为那意味着你可以从那个存储中恢复，
because that means you can resume from that save,

40
00:01:40,800 --> 00:01:43,290
所以你可以实现将你的模型进行自动云存储。
so you get this automatic cloud-saving of your model.

41
00:01:43,290 --> 00:01:45,027
你甚至可以利用
And you can even run inference

42
00:01:45,027 --> 00:01:47,730
该回调所上传的模型存储点(checkpoint) 
with the checkpoints of your model

43
00:01:47,730 --> 00:01:50,208
进行推断过程。
that have been uploaded by this callback.

44
00:01:50,208 --> 00:01:52,260
这意味着你可以，
And that means you can,

45
00:01:52,260 --> 00:01:54,150
你懂的，运行一些测试输入
y'know, run some test inputs

46
00:01:54,150 --> 00:01:56,100
并实际查看你的模型在训练的各个阶段
and actually see how your model works

47
00:01:56,100 --> 00:01:57,990
是如何工作的，
at various stages during training,

48
00:01:57,990 --> 00:01:59,540
这是一个非常好的功能。
which is a really nice feature.

49
00:02:00,390 --> 00:02:03,960
所以我们要添加 PushToHubCallback，
So we're going to add the PushToHubCallback,

50
00:02:03,960 --> 00:02:05,670
它只需要几个参数。
and it takes just a few arguments.

51
00:02:05,670 --> 00:02:08,250
第一个参数是临时目录
So the first argument is the temporary directory

52
00:02:08,250 --> 00:02:10,260
文件在上传到 Hub 之前
that files are going to be saved to

53
00:02:10,260 --> 00:02:12,150
将被保存到该目录。
before they're uploaded to the Hub.

54
00:02:12,150 --> 00:02:14,127
第二个参数是分词器，
The second argument is the tokenizer,

55
00:02:14,127 --> 00:02:15,808
第三个参数在这里
and the third argument here

56
00:02:15,808 --> 00:02:19,080
是关键字参数 hub_model_id。
is the keyword argument hub_model_id.

57
00:02:19,080 --> 00:02:21,330
也就是在 HuggingFace Hub 上
So that's the name it's going to be saved under

58
00:02:21,330 --> 00:02:23,006
它将被保存的名称。
on the HuggingFace Hub.

59
00:02:23,006 --> 00:02:26,267
你还可以上传到组织帐户
You can also upload to an organization account

60
00:02:26,267 --> 00:02:29,370
只需在带有斜杠的仓库名称之前
just by adding the organization name

61
00:02:29,370 --> 00:02:32,460
添加组织名称，就像这样。
before the repository name with a slash, like this.

62
00:02:32,460 --> 00:02:34,020
所以你可能没有权限
So you probably don't have permissions

63
00:02:34,020 --> 00:02:36,000
上传到 HuggingFace 组织，
to upload to the HuggingFace organization,

64
00:02:36,000 --> 00:02:37,170
如果你有权限可以上传，
if you do please file a bug

65
00:02:37,170 --> 00:02:38,973
请提交 bug 并尽快通知我们。
and let us know extremely urgently.

66
00:02:40,830 --> 00:02:42,960
但如果你确实有权限访问你自己的组织，
But if you do have access to your own organization,

67
00:02:42,960 --> 00:02:44,730
那么你可以使用相同的方法
then you can use that same approach

68
00:02:44,730 --> 00:02:46,650
将模型上传到他们的帐户
to upload models to their account

69
00:02:46,650 --> 00:02:50,760
而不是你自己的个人模型集。
instead of to your own personal set of models.

70
00:02:50,760 --> 00:02:53,520
所以，一旦你实现了回调，
So, once you've made your callback,

71
00:02:53,520 --> 00:02:56,310
你只需在调用 model.fit 时
you simply add it to the callbacks list

72
00:02:56,310 --> 00:02:58,080
将它添加到回调列表。
when you're calling model.fit.

73
00:02:58,080 --> 00:03:01,110
所有内容均为从那里上传，
And everything is uploaded for you from there,

74
00:03:01,110 --> 00:03:02,610
没有什么可担心的。
there's nothing else to worry about.

75
00:03:02,610 --> 00:03:04,530
上传模型的第二种方式，
The second way to upload a model, though,

76
00:03:04,530 --> 00:03:07,020
就是调用 model.push_to_hub。
is to call model.push_to_hub.

77
00:03:07,020 --> 00:03:09,086
所以这更像是一种一次性的方法。
So this is more of a once-off method.

78
00:03:09,086 --> 00:03:11,550
在训练期间不会定期调用它。
It's not called regularly during training.

79
00:03:11,550 --> 00:03:13,680
你可以随时手动调用它
You can just call this manually whenever you want to

80
00:03:13,680 --> 00:03:15,240
将模型上传到 hub。
upload a model to the hub.

81
00:03:15,240 --> 00:03:18,949
所以我们建议在训练结束后运行这个，
So we recommend running this after the end of training,

82
00:03:18,949 --> 00:03:21,870
只是为了确保你有一条 commit 消息
just to make sure that you have a commit message

83
00:03:21,870 --> 00:03:24,060
保证这是训练结束时
to guarantee that this was the final version

84
00:03:24,060 --> 00:03:26,143
模型的最终版本。
of the model at the end of training.

85
00:03:26,143 --> 00:03:27,930
它只是为了确保，你懂的，
And it just makes sure that, you know,

86
00:03:27,930 --> 00:03:30,480
你当前正在使用的是最终训练结束的模型
you're working with the definitive end-of-training model

87
00:03:30,480 --> 00:03:32,190
而不是意外从某个地方拿到的
and not accidentally using a checkpoint

88
00:03:32,190 --> 00:03:34,224
模型的某个存储点的版本。
from somewhere along the way.

89
00:03:34,224 --> 00:03:37,173
所以我要运行这两个单元格。
So I'm going to run both of these cells.

90
00:03:39,299 --> 00:03:41,716
然后我要在这里剪视频，
And then I'm going to cut the video here,

91
00:03:41,716 --> 00:03:43,080
只是因为训练需要几分钟。
just because training is going to take a couple of minutes.

92
00:03:43,080 --> 00:03:44,580
所以我会直接跳到最后，
So I'll skip forward to the end of that,

93
00:03:44,580 --> 00:03:46,320
当模型全部上传后，
when the models have all been uploaded,

94
00:03:46,320 --> 00:03:48,390
我会告诉你
and I'm gonna show you how you can

95
00:03:48,390 --> 00:03:50,010
如何访问 Hub 中的模型，
access the models in the Hub,

96
00:03:50,010 --> 00:03:52,713
以及其他能够在那里利用模型做到的事情。
and the other things you can do with them from there.

97
00:03:55,440 --> 00:03:56,700
好的，我们回来了，
Okay, we're back,

98
00:03:56,700 --> 00:03:59,160
我们的模型已上传。
and our model was uploaded.

99
00:03:59,160 --> 00:04:00,750
两者都由 PushToHubCallback
Both by the PushToHubCallback

100
00:04:00,750 --> 00:04:04,251
以及我们在训练后调用 model.push_to_hub。
and also by our call to model.push_to_hub after training.

101
00:04:04,251 --> 00:04:05,910
所以一切看起来都很好。
So everything's looking good.

102
00:04:05,910 --> 00:04:09,960
那么现在如果我们转到我在 HuggingFace 上的个人资料，
So now if we drop over to my profile on HuggingFace,

103
00:04:09,960 --> 00:04:12,630
你只需点击 profile 按钮就可以
and you can get there just by clicking the profile button

104
00:04:12,630 --> 00:04:13,680
在下拉列表中打开。
in the dropdown.

105
00:04:13,680 --> 00:04:16,860
我们可以看到 bert-fine-tuned-cola 模型在这里，
We can see that the bert-fine-tuned-cola model is here,

106
00:04:16,860 --> 00:04:18,369
并于 3 分钟前更新。
and was updated 3 minutes ago.

107
00:04:18,369 --> 00:04:20,520
所以它总是在你的列表的顶部，
So it'll always be at the top of your list,

108
00:04:20,520 --> 00:04:23,340
因为它们是按最近更新时间排序的。
because they're sorted by how recently they were updated.

109
00:04:23,340 --> 00:04:25,740
我们可以立即开始查询我们的模型。
And we can start querying our model immediately.

110
00:04:30,564 --> 00:04:32,939
所以我们训练的数据集
So the dataset we were training on

111
00:04:32,939 --> 00:04:34,320
是 Glue CoLA 数据集，
is the Glue CoLA dataset,

112
00:04:34,320 --> 00:04:36,210
CoLA 是 Corpus of Linguistic Acceptability(语言可接受性语料库)
and CoLA is an acronym standing for

113
00:04:36,210 --> 00:04:39,420
的首字母缩写。
the Corpus of Linguistic Acceptability.

114
00:04:39,420 --> 00:04:42,480
所以这意味着正在训练模型用来决定
So what that means is the model is being trained to decide

115
00:04:42,480 --> 00:04:46,350
一个句子在语法或语言上是正确的呢，
if a sentence is grammatically or linguistically okay,

116
00:04:46,350 --> 00:04:48,171
还是有问题的呢。
or if there's a problem with it.

117
00:04:48,171 --> 00:04:52,890
例如，我们可以说，“This is a legitimate sentence.”
For example, we could say, "This is a legitimate sentence."

118
00:04:52,890 --> 00:04:54,180
并且希望它判断
And hopefully it realizes that

119
00:04:54,180 --> 00:04:56,080
这实际上是一个合理的句子。
this is in fact a legitimate sentence.

120
00:04:57,630 --> 00:05:00,240
所以当你第一次调用模型时，
So it might take a couple of seconds for the model to load

121
00:05:00,240 --> 00:05:03,060
加载模型可能需要几秒钟。
when you call it for the first time.

122
00:05:03,060 --> 00:05:05,960
所以我可能会在这里从这个视频中剪掉几秒钟。
So I might cut a couple of seconds out of this video here.

123
00:05:07,860 --> 00:05:09,060
好的，我们回来了。
Okay, we're back.

124
00:05:09,060 --> 00:05:12,407
所以模型加载了，我们得到了一个输出结果，
So the model loaded and we got an output,

125
00:05:12,407 --> 00:05:14,340
但这里有一个明显的问题。
but there's an obvious problem here.

126
00:05:14,340 --> 00:05:16,888
这些标签并没有真正告诉我们
So these labels aren't really telling us

127
00:05:16,888 --> 00:05:19,740
模型实际为这个输入的句子
what categories the model has actually assigned

128
00:05:19,740 --> 00:05:21,655
分配了哪些类别。
to this input sentence.

129
00:05:21,655 --> 00:05:23,520
所以如果我们想解决这个问题，
So if we want to fix that,

130
00:05:23,520 --> 00:05:26,010
我们要确保模型配置
we want to make sure the model config

131
00:05:26,010 --> 00:05:28,980
针对每个标签类别都有正确的名称，
has the correct names for each of the label classes,

132
00:05:28,980 --> 00:05:30,707
然后我们要上传该配置。
and then we want to upload that config.

133
00:05:30,707 --> 00:05:32,220
所以我们可以在这里实现。
So we can do that down here.

134
00:05:32,220 --> 00:05:34,050
要获取标签名称，
To get the label names,

135
00:05:34,050 --> 00:05:36,547
我们可以从我们加载的数据集中得到它，
we can get that from the dataset we loaded,

136
00:05:36,547 --> 00:05:39,627
它其中包含特性属性。
from the features attribute it has.

137
00:05:39,627 --> 00:05:42,217
然后我们可以创建字典
And then we can create dictionaries

138
00:05:42,217 --> 00:05:44,865
“id2label” 和 “label2id”，
"id2label" and "label2id",

139
00:05:44,865 --> 00:05:47,452
并将它们设置到模型配置中。
and just assign them to the model config.

140
00:05:47,452 --> 00:05:50,790
然后我们可以推送我们更新的配置，
And then we can just push our updated config,

141
00:05:50,790 --> 00:05:54,690
这将覆盖 Hub 仓库中的现有配置。
and that'll override the existing config in the Hub repo.

142
00:05:54,690 --> 00:05:56,368
所以这已经完成了。
So that's just been done.

143
00:05:56,368 --> 00:05:58,320
那么现在，如果我们回到这里，
So now, if we go back here,

144
00:05:58,320 --> 00:06:00,000
我要用一个稍微不同的句子
I'm going to use a slightly different sentence

145
00:06:00,000 --> 00:06:03,540
因为句子的输出有时会被缓存。
because the outputs for sentences are sometimes cached.

146
00:06:03,540 --> 00:06:06,030
所以，如果我们想生成新的结果
And so, if we want to generate new results

147
00:06:06,030 --> 00:06:07,590
我将使用一些稍微不同的东西。
I'm going to use something slightly different.

148
00:06:07,590 --> 00:06:09,783
那么，让我们尝试换一个不正确的句子。
So let's try an incorrect sentence.

149
00:06:10,830 --> 00:06:12,640
所以这不是有效的英语语法
So this is not valid English grammar

150
00:06:13,538 --> 00:06:15,030
希望模型能发现这一点。
and hopefully the model will see that.

151
00:06:15,030 --> 00:06:16,958
它会在这里重新加载，
It's going to reload here,

152
00:06:16,958 --> 00:06:18,630
所以在这里稍微快几秒，
so I'm going to cut a couple of seconds here,

153
00:06:18,630 --> 00:06:20,933
然后我们会看到模型会返回什么结果。
and then we'll see what the model is going to say.

154
00:06:22,860 --> 00:06:23,820
好的。
Okay.

155
00:06:23,820 --> 00:06:26,580
所以这个模型，它的置信度不是很好，
So the model, it's confidence isn't very good,

156
00:06:26,580 --> 00:06:28,830
因为我们还没有真正优化
because of course we didn't really optimize

157
00:06:28,830 --> 00:06:30,630
我们的超参数。
our hyperparameters at all.

158
00:06:30,630 --> 00:06:32,190
但它返回的结果显示这句话
But it has decided that this sentence

159
00:06:32,190 --> 00:06:35,094
不可接受的程度大于可接受的程度。
is more likely to be unacceptable than acceptable.

160
00:06:35,094 --> 00:06:38,160
如果我们更努力地训练
Presumably if we tried a bit harder with training

161
00:06:38,160 --> 00:06:40,080
我们可以获得更低的验证集 loss，
we could get a much lower validation loss,

162
00:06:40,080 --> 00:06:43,830
因此模型的预测会更准确。
and therefore the model's predictions would be more precise.

163
00:06:43,830 --> 00:06:46,260
但是让我们再试一次我们原来的句子。
But let's try our original sentence again.

164
00:06:46,260 --> 00:06:49,140
当然，因为缓存的问题，
Of course, because of the caching issue,

165
00:06:49,140 --> 00:06:52,740
我们看到原来的答案没有改变。
we're seeing that the original answers are unchanged.

166
00:06:52,740 --> 00:06:55,196
所以让我们尝试一个不同的，有效的句子。
So let's try a different, valid sentence.

167
00:06:55,196 --> 00:06:58,767
所以让我们试试，“This is a valid English sentence”。
So let's try, "This is a valid English sentence".

168
00:07:00,150 --> 00:07:02,100
我们看到现在模型的结果是正确的
And we see that now the model correctly decides

169
00:07:02,100 --> 00:07:04,290
它的可接受度非常高，
that it has a very high probability of being acceptable,

170
00:07:04,290 --> 00:07:06,900
并且被拒绝的可能性非常低。
and a very low probability of being unacceptable.

171
00:07:06,900 --> 00:07:09,930
所以你可以使用这个 inference API
So you can use this inference API

172
00:07:09,930 --> 00:07:12,810
甚至可用于训练期间上传的存储点，
even with the checkpoints that are uploaded during training,

173
00:07:12,810 --> 00:07:14,546
能够看到在每个训练的 epoch 时 
so it can be very interesting to see how

174
00:07:14,546 --> 00:07:17,690
不同的样本输入所输出的预测结果  
the model's predictions for sample inputs change

175
00:07:17,690 --> 00:07:20,579
也是非常有意思的。
with each epoch of training.

176
00:07:20,579 --> 00:07:23,370
另外，你也可以访问
Also, the model we've uploaded

177
00:07:23,370 --> 00:07:25,740
我们上传的模型
is going to be accessible to you and,

178
00:07:25,740 --> 00:07:28,046
如果公开分享，其它任何人都可以访问。
if it's shared publicly, to anyone else.

179
00:07:28,046 --> 00:07:29,788
所以如果你想加载那个模型，
So if you want to load that model,

180
00:07:29,788 --> 00:07:32,500
你或其他任何人所需要做的
all you or anyone else needs to do

181
00:07:34,290 --> 00:07:37,440
就是将它加载到 pipeline，
is just to load it in either a pipeline,

182
00:07:37,440 --> 00:07:40,925
或者你可以使用其它方式加载它，例如，
or you can just load it with, for example,

183
00:07:40,925 --> 00:07:43,203
TFAutoModelForSequenceClassification。
TFAutoModelForSequenceClassification.

184
00:07:46,920 --> 00:07:49,989
然后对于名称，你只需传入
And then for the name you would just simply pass

185
00:07:49,989 --> 00:07:53,325
想要上传的 repo 的路径即可。
the path to the repo you want to upload.

186
00:07:53,325 --> 00:07:55,890
或者下载，不好意思。
Or to download, excuse me.

187
00:07:55,890 --> 00:07:58,710
所以如果我想再次使用这个模型，
So if I want to use this model again,

188
00:07:58,710 --> 00:08:00,667
如果我想从 hub 加载它，
if I want to load it from the hub,

189
00:08:00,667 --> 00:08:01,763
仅需运行这一行代码。
I just run this one line of code.

190
00:08:02,813 --> 00:08:03,773
该模型将被下载。
The model will be downloaded.

191
00:08:07,757 --> 00:08:10,080
而且，运气好的话，它会准备好
And, with any luck, it'll be ready to

192
00:08:10,080 --> 00:08:12,450
对不同的数据集进行微调，进行预测，
fine-tune on a different dataset, make predictions with,

193
00:08:12,450 --> 00:08:14,340
或者做任何你想做的事情。
or do anything else you wanna do.

194
00:08:14,340 --> 00:08:17,700
以上内容就是关于 
So that was a quick overview of how,

195
00:08:17,700 --> 00:08:19,470
在你训练期间或者训练之后
after your training or during your training,

196
00:08:19,470 --> 00:08:21,420
如何将模型上传到 Hub，
you can upload models to the Hub,

197
00:08:21,420 --> 00:08:22,440
可以添加存储点，
you can checkpoint there,

198
00:08:22,440 --> 00:08:24,240
可以恢复训练， 
you can resume training from there,

199
00:08:24,240 --> 00:08:26,790
以及从所上传模型 
and you can get inference results

200
00:08:26,790 --> 00:08:28,384
获得推理结果。
from the models you've uploaded.

201
00:08:28,384 --> 00:08:31,084
谢谢大家，希望在以后的视频再会。
So thank you, and I hope to see you in a future video.

202
00:08:32,852 --> 00:08:34,935
（嗖嗖）
(swoosh)