create account

6-29 RNN과 CNN을 결합한 Hybrid 알고리듬에 의한 MNIST 머신러닝 by codingart

View this thread on: hive.blogpeakd.comecency.com
· @codingart · (edited)
$0.56
6-29 RNN과 CNN을 결합한 Hybrid 알고리듬에 의한 MNIST 머신러닝
![noname01.png](https://cdn.steemitimages.com/DQmc14x12tJpPPiwgit33jR25CFvkyCAKAsJwWsUMcx99kf/noname01.png)

RNN 에 의한 MNIST  문자 데이터 처리를 위한 구조이다. MNIST  문자 데이터 베이스를 읽어 들이는 명령은 다음과 같다. 즉 하나의 문자를 784개의 1차원 픽셀 배열이 아니라 28X28 매트릭스 배열 형태로 읽어 들이게 된다.

x = tf.placeholder("float",[None, time_steps, n_input])

이 데이터를 RNN 처리하기 위해서는 위 그림과 같이 28개 열로 나누어 각가의 28개 픽셀 데이터로 구성된 개개의 데이터를 28개 Cell 들이 연결되어 구성된 RNN에  입력하도록 하자.

28X28 매트릭스 형태로 읽어 들인 데이터 x를 아래와 같이 unstack  명령을 사용하여 28개 열(row)로 분할하도록 한다. time_steps 는 28개의 연결된 Cell  상태 수를 의미한다.

다음 url 주소로 이동하여 첨부된 코드 와 함께 마저 읽어 보십시오.

http://blog.daum.net/ejleep1/868

![마나마인로고.png]
https://cdn.steemitimages.com/DQmeePhYx37SUt2zaQJZjJZenWLWSyeK2zKiEp2erB77Lhb/%EB%A7%88%EB%82%98%EB%A7%88%EC%9D%B8%EB%A1%9C%EA%B3%A0.png
👍  , , , , , , , , , , , ,
properties (23)
authorcodingart
permlink2ejgac-6-28-rnn-mnist
categorykr
json_metadata{"tags":["kr","kr-new","manamine","jjangjjangman","kr-dev"],"image":["https://cdn.steemitimages.com/DQmc14x12tJpPPiwgit33jR25CFvkyCAKAsJwWsUMcx99kf/noname01.png","https://cdn.steemitimages.com/DQmeePhYx37SUt2zaQJZjJZenWLWSyeK2zKiEp2erB77Lhb/%EB%A7%88%EB%82%98%EB%A7%88%EC%9D%B8%EB%A1%9C%EA%B3%A0.png"],"links":["http://blog.daum.net/ejleep1/868"],"app":"steemit/0.1","format":"markdown"}
created2019-09-23 12:52:27
last_update2019-09-25 10:38:24
depth0
children1
last_payout2019-09-30 12:52:27
cashout_time1969-12-31 23:59:59
total_payout_value0.283 HBD
curator_payout_value0.280 HBD
pending_payout_value0.000 HBD
promoted0.000 HBD
body_length750
author_reputation41,833,111,979,966
root_title"6-29 RNN과 CNN을 결합한 Hybrid 알고리듬에 의한 MNIST 머신러닝"
beneficiaries[]
max_accepted_payout1,000,000.000 HBD
percent_hbd10,000
post_id90,853,217
net_rshares2,243,733,003,840
author_curate_reward""
vote details (13)
@steemitboard ·
Congratulations @codingart! You have completed the following achievement on the Steem blockchain and have been rewarded with new badge(s) :

<table><tr><td><img src="https://steemitimages.com/60x70/http://steemitboard.com/@codingart/posts.png?201909231340"></td><td>You published more than 650 posts. Your next target is to reach 700 posts.</td></tr>
</table>

<sub>_You can view [your badges on your Steem Board](https://steemitboard.com/@codingart) and compare to others on the [Steem Ranking](https://steemitboard.com/ranking/index.php?name=codingart)_</sub>
<sub>_If you no longer want to receive notifications, reply to this comment with the word_ `STOP`</sub>


To support your work, I also upvoted your post!


**Do not miss the last post from @steemitboard:**
<table><tr><td><a href="https://steemit.com/steemfest/@steemitboard/steemitboard-supports-the-steemfest-travel-reimbursement-fund"><img src="https://steemitimages.com/64x128/https://cdn.steemitimages.com/DQmXDHs9xfx8ZZ3DESFUqHRUQAcQT5kUWobArsRoJg2Yz1F/image.png"></a></td><td><a href="https://steemit.com/steemfest/@steemitboard/steemitboard-supports-the-steemfest-travel-reimbursement-fund">SteemitBoard supports the SteemFest⁴ Travel Reimbursement Fund.</a></td></tr></table>

###### [Vote for @Steemitboard as a witness](https://v2.steemconnect.com/sign/account-witness-vote?witness=steemitboard&approve=1) to get one more award and increased upvotes!
properties (22)
authorsteemitboard
permlinksteemitboard-notify-codingart-20190923t140620000z
categorykr
json_metadata{"image":["https://steemitboard.com/img/notify.png"]}
created2019-09-23 14:06:21
last_update2019-09-23 14:06:21
depth1
children0
last_payout2019-09-30 14:06:21
cashout_time1969-12-31 23:59:59
total_payout_value0.000 HBD
curator_payout_value0.000 HBD
pending_payout_value0.000 HBD
promoted0.000 HBD
body_length1,422
author_reputation38,975,615,169,260
root_title"6-29 RNN과 CNN을 결합한 Hybrid 알고리듬에 의한 MNIST 머신러닝"
beneficiaries[]
max_accepted_payout1,000,000.000 HBD
percent_hbd10,000
post_id90,855,180
net_rshares0