Skip to content

Conversation

@wckim
Copy link
Contributor

@wckim wckim commented Oct 2, 2019

added guide doc is korean translation of site/en/guide/tensor.md

@googlebot googlebot added the cla: yes CLA has been signed label Oct 2, 2019
@tfdocsbot tfdocsbot added ko translation Community-provided translation for certain languages labels Oct 2, 2019
@tfdocsbot
Copy link
Collaborator

Reviewers added, please take a look.
@rickiepark, @cre8tor, @Choiuijin1125, @JKIsaacLee, @NoelBird

When your review is finished, approve the pull request or include "LGTM" in your comment.

Copy link
Contributor

@JKIsaacLee JKIsaacLee left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks for your nice translation. It's pretty nice translation. 👍
I checked some typo and some of removable original word.

@wckim wckim changed the title add site/ko/guide/tensor.md KO: add site/ko/guide/tensor.md Oct 10, 2019
@wckim
Copy link
Contributor Author

wckim commented Oct 18, 2019

@JKIsaacLee, please check my current revision for your review.
And let me know other comments if you have.

@rickiepark
Copy link
Contributor

Hi @wckim, I'll check this PR asap. :)

Copy link
Contributor

@rickiepark rickiepark left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thank you @wckim , I'm late.
I continue to review ASAP.
Please check my comments. 👍

Copy link
Contributor

@rickiepark rickiepark left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@wckim , I'll continue to review this PR later.
Please check my suggestions.
Thank you so much! :)

Copy link
Contributor

@rickiepark rickiepark left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@wckim, Thank you for the translation. 👍
I finished to review. But I'm not sure tf.Print is suitable in this documetation. :)

Comment on lines 238 to 239
일단 계산 그래프를 만들었다면, 특정 `tf.Tensor`를 생성하고 그것에 지정된 값을 가져오는 계산을 실행할 수 있습니다.
이것은 대부분의 텐서플로가 작업하는데 필요할 뿐 아니라 디버깅에 종종 유용합니다.
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

To read easily I suggest '생성하고 그것에 지정된 값을' --> '생성하거나 텐서에 할당된 값을',
'대부분의 텐서플로가 작업하는데' --> '텐서플로 작업의 많은 부분에서'

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks, fixed it

Comment on lines 254 to 256
때때로 그 값이 이용할 수 없는 동적인 정보에 의존하기 때문에
컨텍스트(context)가 없는 `tf.Tensor`는 계산할 수 없는 경우가 있습니다.
예를 들어, `placeholder`인 텐서는 그 `placeholder`에 해당하는 값이 제공되지 않으면 계산할 수 없습니다.
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I suggest '그 값이 이용할 수 없는' --> '현재 사용할 수 없는' and 'placeholder인 텐서는 ' --> 'placeholder에 의존하는 텐서는'

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks, fixed it

다른 모델 구조는 `tf.Tensor`를 계산하는 것을 복잡하게 만들 수 있습니다.
텐서플로는 함수안이나 제어 흐름안에 정의된 `tf.Tensor`를 직접 계산할 수 없습니다.
만약에 `tf.Tensor`가 큐(queue)에 있는 값을 사용한다면,
무언가가 큐에 들어간 후에 만 `tf.Tensor` 계산을 할 수 있습니다; 그렇지 않으면 계산은 중단될 것입니다.
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Hang menas blocking the computation. I suggest ''; 그렇지 않으면 계산은 중단될 것입니다' --> ''. 그렇지 않으면 계산은 멈출(hang) 것입니다'

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks, fixed it.

Comment on lines 274 to 302
## 텐서 출력하기

디버깅을 위해서 `tf.Tensor` 값을 출력하고 싶을 것입니다.
고급 디버깅에 대해 [tfdbg](../guide/debugger.md)에서 가이드를 제공하지만,
텐서플로는 `tf.Tensor`값을 직접 출력할 수 있는 연산자를 가지고 있습니다.

`tf.Tensor`를 출력할 때 다음과 같은 패턴을 쓰고자 하는 경우는 거의 없습니다:

``` python
t = <<some tensorflow operation>>
print(t) # 이것은 그래프가 생성되어질 때 기호화된 텐서(symbolic tensor)를 출력할 것입니다.
# 이 텐서는 이 컨텍스트(context)안에서 값을 가지고 있지 않습니다.
```

이 코드는 `tf.Tensor`의 값이 아닌 객체(지연 계산으로 표현)를 출력합니다.
실제로 텐서플로는 두번째 인수로 전달된 `tf.Tensor` 집합을 출력하는 동안
변경되지 않는 첫번째 텐서 인수를 반환하는 `tf.Print` 연산을 제공합니다.

`tf.Print`을 제대로 사용하기 위해서는 반환된 값을 사용해야 합니다. 아래 예를 보면

``` python
t = <<some tensorflow operation>>
tf.Print(t, [t]) # 어떤 일도 하지 않습니다
t = tf.Print(t, [t]) # 여기서 tf.Print에 의해 반환된 값을 사용할 수 있습니다.
result = t + 1 # 이제 결과를 계산할 때 `t` 값이 출력될 것입니다.
```

`result`를 계산할 때 이와 관련된 모든 것이 계산될 것입니다.
`result`는 `t`와 의존성이 있고, `t`를 계산하는 것이 그 입력(`t`의 이전 값)을 출력하는 부가 효과가 있기 때문에 `t`는 출력됩니다.
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

tf.Print is deprecated in TF 1.5 and not included in TF 2.
I suggest to remove this section.

@lamberta , What do you think about this section? :)

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

If it is important to keep consistency with the original document(en), it is better not to change.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Yes, just waiting @lamberta's comment. :)

Copy link
Contributor

@rickiepark rickiepark Oct 28, 2019

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@wckim, I talked on this issue with Billy.
Please remove this section.
I'll make a pr to english version. :)
Thanks

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks, removed this section.

@rickiepark
Copy link
Contributor

@wckim, Great work. 👍
Thanks a lot!
LGTM!

@tfdocsbot tfdocsbot added kokoro:force-run lgtm Community-added approval labels Oct 29, 2019
Copy link
Contributor

@JKIsaacLee JKIsaacLee left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pretty nice 👍 LGTM

@MarkDaoust MarkDaoust added the ready to pull Start merge process label Nov 10, 2019
lamberta added a commit that referenced this pull request Nov 12, 2019
@lamberta lamberta merged commit dfca1e7 into tensorflow:master Nov 12, 2019
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

cla: yes CLA has been signed ko lgtm Community-added approval ready to pull Start merge process translation Community-provided translation for certain languages

8 participants