[B! tensorflow] ishideoã®ãƒ–ãƒƒã‚¯ãƒžãƒ¼ã‚¯

ishideo id:ishideo

tensorflowã«é–¢ã™ã‚‹ishideoã®ãƒ–ãƒƒã‚¯ãƒžãƒ¼ã‚¯ (18)

${{author_name}}$

{{{comment_expanded}}}

{{label}}

{{#is_bookmark}}ãƒªã‚¹ãƒˆ{{/is_bookmark}}{{^is_bookmark}}ãƒªãƒ³ã‚¯{{/is_bookmark}}

${{author_name}}$
{{author_name}}{{created}}
{{ #comment }}{{ comment }}{{ /comment }}
- {{ label }}

${{author_name}}$

{{{comment_expanded}}}

{{label}}

{{#is_bookmark}}ãƒªã‚¹ãƒˆ{{/is_bookmark}}{{^is_bookmark}}ãƒªãƒ³ã‚¯{{/is_bookmark}}

ãƒ•ã‚©ãƒ³ãƒˆãƒ•ã‚¡ã‚¤ãƒ«ã«åŽéŒ²ã•ã‚Œã¦ãªã„å—ã‚’æ©Ÿæ¢°å¦ç¿’ã§ç”Ÿæˆã—ã¦ã¿ãŸ | Qiita
çµæžœ ã“ã®ç”»åƒã¯çµæžœã§ã™ã€ã“ã® document ã®ã€‚ ã“ã®æ–¹æ³•ã¯ã€ convert ã™ã‚‹ã“ã¨ã§ã™ã€ source-han-sans Japanese ã‚’ GLT-ã”ã¬ã‚“ãï¼ˆã‚¢ãƒ‹ãƒ¡ã€Œå°‘å¥³çµ‚æœ«æ—…è¡Œã€å‰¯é¡Œé¢¨ãƒ•ã‚©ãƒ³ãƒˆï¼‰ ã¸ã€‚ 'æ˜¨' isn't included in GLT-ã”ã¬ã‚“ãï¼ˆã‚¢ãƒ‹ãƒ¡ã€Œå°‘å¥³çµ‚æœ«æ—…è¡Œã€å‰¯é¡Œé¢¨ãƒ•ã‚©ãƒ³ãƒˆï¼‰ --max_epochs 45 ä¸»ã«ä½¿ã£ãŸã‚‚ã® Google Colab https://colab.research.google.com Google user ã¯å¯èƒ½ã§ã™ã€ ä½¿ã†ã“ã¨ãŒ ( GPU ã‚„ TPU ã‚’) ã€ Google Colab ä¸Šã§ã€‚ Google user ã¯å¯èƒ½ã§ã™ã€ã—ãªã„ ( pay ã™ã‚‹ã“ã¨ã‚’ ) ã€ Google Colab ã¸ã€‚ ( è¦å‘Š ) Google Colabã¯æœ‰ã—ã¾ã™ã€æ§˜ã€…ãªåˆ¶ç´„ã‚’ã€‚ æŒã¤ã“ã¨ã¯ ( Google account
ishideo 2022/08/31
kanji

font

ml

python

qiita

google

colaboratory

tensorflow

pix2pix
ãƒªãƒ³ã‚¯
Pythonä»¥å¤–ã§ã‚‚Deep Learningã—ãŸã„ï¼ - Qiita
ã¯ã˜ã‚ã« ã“ã®è¨˜äº‹ã¯ã€NTTãƒ†ã‚¯ãƒŽã‚¯ãƒã‚¹ Advent Calendar 2019 ã®23æ—¥ç›®ã§ã™ã€‚ ã“ã‚“ã«ã¡ã¯ã€NTTãƒ†ã‚¯ãƒŽã‚¯ãƒã‚¹ã®åºƒç€¬ã¨ç”³ã—ã¾ã™ã€‚æ¥å‹™ã§ã¯é«˜ç²¾ç´°VRé…ä¿¡ã‚¨ãƒ³ã‚¸ãƒ³1ã‚„æ©Ÿæ¢°å¦ç¿’ã«ã‚ˆã‚‹ç”»åƒèªè˜AIã‚¨ãƒ³ã‚¸ãƒ³ã®ç ”ç©¶é–‹ç™ºã«å–ã‚Šçµ„ã‚“ã§ã„ã¾ã™ã€‚ ä»Šå›žã¯ã¡ã‚‡ã£ã¨æŠ€è¡“ã‚„é–‹ç™ºã‹ã‚‰é›¢ã‚Œã¦ã€Deep Learningã‚’çµ„ã¿è¾¼ã‚“ã ã‚¢ãƒ—ãƒªé–‹ç™ºç¾å ´ã§ã€é€Ÿåº¦ã‚’æ±‚ã‚ãŸã‚ŠGUIãŒæ¬²ã—ã‹ã£ãŸã‚Šã¨ã‚¢ãƒ—ãƒªé–‹ç™ºã¯Pythonä»¥å¤–ã§é–‹ç™ºã—ãŸã„æ™‚ã¯ã©ã†ã™ã‚‹ã®ï¼Ÿã¨ã„ã£ãŸç–‘å•ã‚’è§£æ±ºã™ã‚‹å…¥å£éƒ¨åˆ†ã«è§¦ã‚Œã¦ã„ããŸã„ã¨æ€ã„ã¾ã™ã€‚ å…¥é–€æ›¸ã‚’èªã¿çµ‚ã‚ã£ã¦ã‚¢ãƒ—ãƒªã‚’ä½œã‚ã†ï¼ã¨ã„ã†æ–¹ã®ã€æ¬¡ã®å‹‰å¼·ã‚¹ãƒ†ãƒ¼ã‚¸é¸æŠžã®æ‰‹åŠ©ã‘ã«ãªã‚Œã‚Œã°å¹¸ã„ã§ã™ã€‚ æˆ‘ã€…ã¯Pythonã«ç¸›ã‚‰ã‚Œã¦ã„ã‚‹ ä¸–ã®ä¸ã«ã¯TensorFlowã‚„PyTorch/Chainerã¨ã„ã£ãŸDeep Learningã‚’æ‰±ã†ãƒ•ãƒ¬ãƒ¼ãƒ ãƒ¯ãƒ¼ã‚¯ã¯æ•°å¤šãå˜åœ¨ã™ã‚‹ã®ã§ã™ãŒã€é–‹ç™ºã«ã‚ãŸã£ã¦ã¯Pythonã‚’å‰
ishideo 2022/08/20
onnx

python

savedmodel

tensorflow

keras

ml

menoh

tf2onnx

qiita
ãƒªãƒ³ã‚¯
Go ã«ã‚ˆã‚‹æ©Ÿæ¢°å¦ç¿’ æŽ¨è«–ãƒ•ãƒ¬ãƒ¼ãƒ ãƒ¯ãƒ¼ã‚¯ã®æœ€æ–°å‹•å‘ 2019 - Qiita
Deleted articles cannot be recovered. Draft of this article would be also deleted. Are you sure you want to delete this article? Golang ã§æŽ¨è«– æ˜¨ä»Šã§ã¯ã€Œæ©Ÿæ¢°å¦ç¿’ã¨è¨€ãˆã° Pythonã€ã€ŒPython ã¨è¨€ãˆã°æ©Ÿæ¢°å¦ç¿’ã€ã¨æ€ã‚ã‚ŒãŒã¡ãªã®ã§ã™ãŒã€æŽ¨è«–ç”¨é€”ã§ã‚ã‚Œã°å¦ç¿’æ¸ˆã¿ãƒ¢ãƒ‡ãƒ«ã‚’åˆ©ç”¨ã—ã¦è‰²ã€…ãªãƒ—ãƒã‚°ãƒ©ãƒŸãƒ³ã‚°è¨€èªžã‹ã‚‰æ‰±ãˆã¾ã™ã€‚Go ã‹ã‚‰æ‰±ãˆã‚‹æ©Ÿæ¢°å¦ç¿’ãƒ©ã‚¤ãƒ–ãƒ©ãƒªã®å†…ã€å¦ç¿’æ¸ˆã¿ãƒ¢ãƒ‡ãƒ«ãŒåˆ©ç”¨ã§ãã‚‹ç‰©ã¨ã—ã¦ã¯ãŠãŠã‚ˆãä»¥ä¸‹ã®3ã¤ã«çµžã‚‰ã‚Œã¾ã™ã€‚ TensorFlow TensorFlow Lite ONNX Golang ã§ TensorFlow Golang ã§ TensorFlow ã‚’åˆ©ç”¨ã™ã‚‹å ´åˆã¯ã‚ªãƒ•ã‚£ã‚·ãƒ£ãƒ«ã‹ã‚‰ Go ã® binding ãŒæä¾›ã•ã‚Œã¦ã„ã‚‹ã®ã§ Go
ishideo 2022/08/20
golang

ml

onnx

tensorflow

onnx-go

qiita

go-menoh

go-tflite
ãƒªãƒ³ã‚¯
Detecting Malicious Requests with Keras & Tensorflow
Security is a concern for any public facing web application. Good development practices can assist with defending against attempts from users looking to expose data or bring down an app. However, sometimes not all attack vectors are handled and new exploits are bound to be discovered. This is where security software can assist with monitoring and preventing unforeseen attacks. So what if you could
ishideo 2022/08/19
malicious

keras

tensorflow

request

detection

score

ml

python
ãƒªãƒ³ã‚¯
Phishing Predictions with Deep Learning and TensorFlow â€” CSIRT Gadgets, LLC
ishideo 2022/08/19
phishing

tensorflow

domain

score

example

github

python
ãƒªãƒ³ã‚¯
How to Hunt for Phishing Domains using TensorFlow â€” CSIRT Gadgets, LLC
ishideo 2022/08/19
phishing

tensorflow

domain

score

example

github

python
ãƒªãƒ³ã‚¯
GitHub - csirtgadgets/tf-domains-example: Tensorflow Phishing Domains Example
ishideo 2022/08/19
phishing

tensorflow

domain

score

example

github

python
ãƒªãƒ³ã‚¯
AWS Lambda + Docker + TensorFlowã‚’ä½¿ã£ã¦ã‚µã‚¯ãƒƒã¨æŽ¨è«–APIã‚’ã¤ãã‚‹ - aptpod Tech Blog
ã¿ãªã•ã¾ã€ã“ã‚“ã«ã¡ã¯ã€‚ç ”ç©¶é–‹ç™ºã‚°ãƒ«ãƒ¼ãƒ—ã¨è£½å“é–‹ç™ºã‚°ãƒ«ãƒ¼ãƒ— ã«å…¼å‹™ã§æ‰€å±žã—ã¦ãŠã‚Šã¾ã™ã€ãã—ã ã§ã™ã€‚aptpod Advent Calendar 2020 11æ—¥ç›®ã‚’æ‹…å½“ã—ã¾ã™ã€‚ä»Šå›žã¯æ©Ÿæ¢°å¦ç¿’ã«é–¢ã‚ã‚‹ã‚¨ãƒ³ã‚¸ãƒ‹ã‚¢å‘ã‘ã«ã€æœ€è¿‘AWSãŒãƒªãƒªãƒ¼ã‚¹ã—ãŸAWS Lambdaã®æ–°æ©Ÿèƒ½ã‚’åˆ©ç”¨ã—ã¦ã€ã‚µã‚¯ãƒƒã¨æŽ¨è«–APIã‚’ä½œã‚‹ãƒã‚¿ã‚’ã”ç´¹ä»‹ã—ã¾ã™ã€‚ aws.amazon.com æŽ¨è«–ã‚’ã™ã°ã‚„ãAPIåŒ–ã™ã‚‹æ„ç¾© ãã®å‰ã«ã€æŽ¨è«–ç®‡æ‰€ã‚’APIã¨ã—ã¦ã™ã°ã‚„ãç”¨æ„ã§ãã‚‹å¿…è¦æ€§ã«ã¤ã„ã¦ç°¡å˜ã«è§¦ã‚ŒãŸã„ã¨æ€ã„ã¾ã™ã€‚ æ©Ÿæ¢°å¦ç¿’é–¢ä¿‚ã®æ¡ˆä»¶ã§ã¯ã€ãŠå®¢æ§˜å´ã‚‚ç†è§£ã§ãã‚‹KPIã‚’ç«‹ã¦ã‚‹ã“ã¨ãŒéžå¸¸ã«é‡è¦è¦–ã•ã‚Œã¦ã„ã¾ã™ã€‚ä¾‹ãˆã°ã€Œãƒ¢ãƒ‡ãƒ«ã®æ£è§£çŽ‡ã‚’ã€œã«ã—ãŸã„ã€ã‚„ã€Œãƒ¢ãƒ‡ãƒ«ã®å‹•ãé€Ÿã•ã‚’ã€œã«ã—ãŸã„ã€ãªã©ã§ã™ãã€‚ã“ã‚Œã‚‰ã®å…·ä½“çš„ãªæ•°å€¤ã¯ãŠå®¢æ§˜ã¨è°è«–ã‚’é‡ãã€ãŠäº’ã„ã«ã—ã£ãã‚Šãã‚‹æ•°å€¤ã«è½ã¨ã—è¾¼ã‚€å¿…è¦ãŒã‚ã‚‹ã®ã§ã™ãŒã€ã“ã‚ŒãŒã¨ã¦ã‚‚ã‚€ãšã‹ã—ã„ã®ã§ã™ã€‚ãªãœã§ã—ã‚‡ã†ã‹ã€‚ ãŠå®¢
ishideo 2020/12/17
api

aws

lambda

docker

tensorflow

ml

serverless
ãƒªãƒ³ã‚¯
python ã§ã‚ˆãã‚ã‹ã‚‰ãªã„ã‚ªãƒ–ã‚¸ã‚§ã‚¯ãƒˆã®ä¸èº«ã‚’è¡¨ç¤ºã™ã‚‹ - Qiita
Tensorflow ç‰ã‚’ä½¿ã£ã¦ã„ã‚‹ã¨ã€ã“ã®ã‚ªãƒ–ã‚¸ã‚§ã‚¯ãƒˆã¯ä¸€ä½“ãªã‚“ã®ã‚¯ãƒ©ã‚¹ã®ã‚¤ãƒ³ã‚¹ã‚¿ãƒ³ã‚¹ã§ã€ã©ã‚“ãªæƒ…å ±ï¼ˆãƒ—ãƒãƒ‘ãƒ†ã‚£ï¼‰ã‚’æŒã£ã¦ã„ã‚‹ã®ã ã‚ã†ã€ã¨ã‚ã‹ã‚‰ãªããªã‚‹ã¨ããŒã—ã°ã—ã°ã‚ã‚Šã¾ã™ã€‚ ã—ã‹ã‚‚ã€ã¾ã¨ã‚‚ãªãƒ‰ã‚ãƒ¥ãƒ¡ãƒ³ãƒˆãŒç„¡ã„ã“ã¨ã‚‚ã—ã°ã—ã°ã‚ã‚Šã¾ã™ã€‚ ï¼ˆä¾‹ãˆã° BeamSearchDecoderOutput ãªã©ï¼‰ ã“ã®ã‚ˆã†ãªæ™‚ã«ä¾¿åˆ©ãªé–¢æ•° explain ã‚’ä½œã‚Šã¾ã—ãŸã€‚ ä¾‹ tf.constant ã®æˆ»ã‚Šå€¤ã‚’èª¿ã¹ã¦ã¿ã¾ã—ã‚‡ã†ã€‚ In [7]: tensor = tf.constant([[1, 2], [3, 4]]) In [8]: explain(tensor) EXPLAIN ------------------ Tensor("Const_2:0", shape=(2, 2), dtype=int32) <class 'tensorflow.python.framework.ops.Tens
ishideo 2019/05/28
python

object

explain

tensorflow

qiita
ãƒªãƒ³ã‚¯
ã€ç§’é€Ÿã§ç„¡æ–™GPUã‚’ä½¿ã†ã€‘TensorFow(Keras)/PyTorch/Chainerç’°å¢ƒæ§‹ç¯‰ on Colaboratory - Qiita
2019/5/11 PR: ã“ã¡ã‚‰ã®å†…å®¹ã‚’å«ã‚ã€2019å¹´5æœˆ11æ—¥ç™ºåˆŠã® å›³è§£é€Ÿç¿’DEEP LEARNINGã¨ã„ã†æœ¬ãŒã§ãã¾ã—ãŸã€‚[2019å¹´5æœˆç‰ˆ] æ©Ÿæ¢°å¦ç¿’ãƒ»æ·±å±¤å¦ç¿’ã‚’å¦ã³ã€ãƒˆãƒ¬ãƒ³ãƒ‰ã‚’è¿½ã†ãŸã‚ã®ãƒªãƒ³ã‚¯150é¸ - Qiitaã§ã‚‚ã€ä¸€éƒ¨å†…å®¹ã‚’ã”è¦§ã„ãŸã ã‘ã¾ã™ 19/1/11 18å¹´1æœˆã®å…¬é–‹å¾Œã€TensorFlowæœ¬ä½“ã«Kerasçµ±åˆã€ChainerãŒãƒ‡ãƒ•ã‚©ãƒ«ãƒˆã§æä¾›ã¨ãªã‚‹ãªã©ã€çŠ¶æ³ãŒå¤‰åŒ–ã—ãŸãŸã‚ã€å¤§å¹…ã«åŠ ç†ã—ã¾ã—ãŸã€‚TensorFlow 2.0 Previewã«ã¤ã„ã¦ã‚‚è¿½è¨˜ã—ã¾ã—ãŸã€‚ 19/1/31 PyTorchãŒæ¨™æº–ã‚¤ãƒ³ã‚¹ãƒˆãƒ¼ãƒ«ã¨ãªã£ãŸã“ã¨ã€PyTorch/ TensorFlowã®Colabç‰ˆãƒãƒ¥ãƒ¼ãƒˆãƒªã‚¢ãƒ«ã‚’è¿½è¨˜ã€‚ 2019/3/9 Colaboratoryã«é–¢ã™ã‚‹æƒ…å ±äº¤æ›Slackã‚’è©¦é¨“çš„ã«ç«‹ã¡ä¸Šã’ã¾ã—ãŸã€‚ãƒªãƒ³ã‚¯ã‚ˆã‚Šã€ç™»éŒ²ãƒ»ã”å‚åŠ ãã ã•ã„ã€‚ TL;DR Google Colab
ishideo 2018/05/02
google

gpu

colaboratory

qiita

deeplearning

chainer

keras

tensorflow
ãƒªãƒ³ã‚¯
ã€ç§’é€Ÿã§ç„¡æ–™GPUã‚’ä½¿ã†ã€‘æ·±å±¤å¦ç¿’å®Ÿè·µTips on Colaboratory - Qiita
2019/5/11 PR: ã“ã¡ã‚‰ã®å†…å®¹ã‚’å«ã‚ã€2019å¹´5æœˆ11æ—¥ç™ºåˆŠã® å›³è§£é€Ÿç¿’DEEP LEARNINGã¨ã„ã†æœ¬ãŒã§ãã¾ã—ãŸã€‚[2019å¹´5æœˆç‰ˆ] æ©Ÿæ¢°å¦ç¿’ãƒ»æ·±å±¤å¦ç¿’ã‚’å¦ã³ã€ãƒˆãƒ¬ãƒ³ãƒ‰ã‚’è¿½ã†ãŸã‚ã®ãƒªãƒ³ã‚¯150é¸ - Qiitaã§ã‚‚ã€ä¸€éƒ¨å†…å®¹ã‚’ã”è¦§ã„ãŸã ã‘ã¾ã™ 2019/3/9 Colaboratoryã«é–¢ã™ã‚‹æƒ…å ±äº¤æ›Slackã‚’è©¦é¨“çš„ã«ç«‹ã¡ä¸Šã’ã¾ã—ãŸã€‚ãƒªãƒ³ã‚¯ã‚ˆã‚Šã€ç™»éŒ²ãƒ»ã”å‚åŠ ãã ã•ã„ã€‚ 2019/3/3 TensorBoardã«å…¬å¼å¯¾å¿œã—ã¾ã—ãŸã€‚ã¾ãŸã€ãƒ©ãƒ³ã‚¿ã‚¤ãƒ ã®RAM/ãƒ‡ã‚£ã‚¹ã‚¯ç©ºãå®¹é‡ãŒä¸€ç›®ã§ç¢ºèªã§ãã‚‹ã‚ˆã†ã«ãªã‚Šã¾ã—ãŸã€‚å¾Œã»ã©è¨˜äº‹ã«è¿½è¨˜ã—ã¾ã™ã€‚ ã¯ã˜ã‚ã« Colaboratoryã¯ã€ç„¡æ–™ã§ä½¿ã†ã“ã¨ãŒã§ãã€ã»ã¨ã‚“ã©ã®ä¸»è¦ãƒ–ãƒ©ã‚¦ã‚¶ã§å‹•ä½œã™ã‚‹ã€è¨å®šä¸è¦ã®JupyterãƒŽãƒ¼ãƒˆãƒ–ãƒƒã‚¯ç’°å¢ƒã§ã™ã€‚GoogleãŒã€æ©Ÿæ¢°å¦ç¿’ã®æ•™è‚²ã€ç ”ç©¶ç”¨ã«ä½¿ã‚ã‚Œã‚‹ã“ã¨ã‚’ç›®çš„ã«ã€ç„¡å„Ÿæä¾›ã—ã¦ã„ã¾ã™ã€‚ã–ã£ãã‚Šã¨ã„ã†ãªã‚‰ã€
ishideo 2018/05/02
google

gpu

colaboratory

qiita

deeplearning

chainer

keras

tensorflow
ãƒªãƒ³ã‚¯
Google Colaboratoryã‚’ä½¿ã£ãŸ GPUå¦ç¿’ç’°å¢ƒã®å®Ÿé¨“ - script day.log
ä»Šå›žã¯Google Colaboratoryã‚’ä½¿ã£ã¦ã€ DeepLearningã®GPUå¦ç¿’ç’°å¢ƒãŒå‡ºæ¥ã‚‹ã¨twitterã§è¦‹ãŸã®ã§ã€å®Ÿé¨“ã—ã¾ã—ãŸã€‚ ã¨ã‚Šã‚ãˆãšå˜ç´”ã«MNISTã®CNNã‚„ã‚Šã¾ã™ã€‚ æä¾›ã•ã‚Œã‚‹ãƒžã‚·ãƒ³ã‚¹ãƒšãƒƒã‚¯ã¯æ¬¡ã®é€šã‚Šã‚‰ã—ã„ã§ã™ã€‚ n1-highmem-2 instance 2vCPU @ 2.2GHz 13GB RAM 33GB Free Space idle cut-off 90 minutes maximum 12 hours stackoverflow.com ã•ã‚ã€å§‹ã‚ã¾ã—ã‚‡ã†ã€‚ ã¯ã˜ã‚ã«æ–°è¦ãƒŽãƒ¼ãƒˆãƒ–ãƒƒã‚¯ã®ä½œæˆã‚’è¡Œã„ã¾ã™ã€‚ ãƒŽãƒ¼ãƒˆãƒ–ãƒƒã‚¯ã‚’æ–°è¦ä½œæˆã‚ˆã‚ŠPython3ã®æ–°ã—ã„ãƒŽãƒ¼ãƒˆãƒ–ãƒƒã‚¯ã‚’é¸ã³ã¾ã™ã€‚ æ¬¡ã«ãƒ©ãƒ³ã‚¿ã‚¤ãƒ ã‚ˆã‚Šãƒ©ãƒ³ã‚¿ã‚¤ãƒ ã®ã‚¿ã‚¤ãƒ—ã‚’å¤‰æ›´ã€ãƒŽãƒ¼ãƒˆãƒ–ãƒƒã‚¯ã®è¨å®šã‚’é–‹ãã€ ãƒãƒ¼ãƒ‰ã‚¦ã‚§ã‚¢ã‚¢ã‚¯ã‚»ãƒ©ãƒ¬ãƒ¼ã‚¿ã«GPUã‚’é¸æŠžã—ã€ä¿å˜ã—ã¾ã™ã€‚ æ–°ã—ãã‚»ãƒ«ã‚’ä½œã‚Šã€ä¸‹è¨˜ã‚’å…¥åŠ›ã—ã€å®Ÿè¡Œã™ã‚‹ã€‚ i
ishideo 2018/05/01
google

colaboratory

gpu

tensorflow

python
ãƒªãƒ³ã‚¯
Colaboratoryã§GPUè¨ˆç®— | Jun Nishii
ishideo 2018/05/01
colaboratory

google

gpu

keras

tensorflow
ãƒªãƒ³ã‚¯
æ©Ÿæ¢°å¦ç¿’ç´ äººãŒ2ã‹æœˆåŠã§æ©Ÿæ¢°å¦ç¿’ã‚’å…¥é–€ã—ãŸã“ã¨ã¾ã¨ã‚ - ã‚ã•ã®ã²ã¨ã‚Šã”ã¨
ã¡ã¾ãŸã§ã¯ã€æ©Ÿæ¢°å¦ç¿’ãŒãƒ–ãƒ¼ãƒ ã®ã‚ˆã†ã§ã™ã€‚ ãŒã€ã€ã¾ã£ãŸãæ™‚ä»£ã«ã¤ã„ã¦ã„ã‘ã¦ã„ã¾ã›ã‚“ã€‚ ã—ã‹ã—ã€æ©Ÿæ¢°å¦ç¿’ã€ç‰¹ã«è‡ªç„¶è¨€èªžå‡¦ç†ã«ç²¾é€šã—ãŸäººã®æŽ¡ç”¨ã«ã‹ã‹ã‚ã‚‹ä»•äº‹ã‚’ã—ã¦ã„ã‚‹ã€ã€ã€ ã«ã‚‚ã‹ã‹ã‚ã‚‰ãšã€è‡ªç„¶è¨€èªžå‡¦ç†ã©ã“ã‚ã‹æ©Ÿæ¢°å¦ç¿’ãŒå…¨ãåˆ†ã‹ã‚‰ãªã„ã€‚ ã“ã‚Œã§ã¯ã€ã„ã‘ãªã„ã€‚ã¨ã„ã†ã“ã¨ã§ ã€Œæ©Ÿæ¢°å¦ç¿’ã‚’ãŸã—ãªã‚€å¦ç”Ÿã®çš†ã•ã‚“ã¨ã€ãµã‚ã£ã¨é›‘è«‡ãŒã§ãã‚‹ãƒ¬ãƒ™ãƒ«ã€ ã‚’ç›®æŒ‡ã—ã¦ã€2017å¹´æ£æœˆæ˜Žã‘ã‹ã‚‰å‹‰å¼·ã‚’å§‹ã‚ã¾ã—ãŸã€‚ ã¡ãªã¿ã«ã€ã©ã‚“ãªã«ã‚ãƒªãŒæ‚ªãã¦ã‚‚1æ—¥3æ™‚é–“ã¾ã§ï¼ã¨æ±ºã‚ã¦ã„ã¾ã™ã€‚ ãã‚‚ãã‚‚æ©Ÿæ¢°å¦ç¿’ã«èˆˆå‘³é–¢å¿ƒãŒã‚ã‚‹ã‚ã‘ã§ã¯ãªã„ ã‚„ã‚‰ãªã‘ã‚Œã°ãªã‚‰ãªã„ä»–ã®ä»•äº‹ãŒã‚ã‚‹ å®¶äº‹è‚²å…ãŒå„ªå…ˆ ãªã®ã§ã€ã™ã“ã—ã§ã‚‚ç„¡ç†ã™ã‚‹ã¨ç¶šã‹ãªã„ãŸã‚ã§ã™ã€‚ ã€ŒAIã§ä¸–ç•Œã‚’å¤‰ãˆã‚‰ã‚Œã‚‹ï¼ã€ ã€Œäººå·¥çŸ¥èƒ½ã§æƒ³åƒã‚‚ã§ããªã„æœªæ¥ãŒã€ã€ã€ã€ ã¿ãŸã„ãªã”æ™‚ä¸–ã®ä¸ã€ã‚ã‚Šãˆãªã„ã»ã©ã®ä½Žãƒ†ãƒ³ã‚·ãƒ§ãƒ³ã§æ·¡ã€…ã¨å‹‰å¼·ã—ã¦ã„ã‚‹ã‚ã‘ã§ é€†ã«ã€ãã†ã„ã†æ„è˜ä½Žã„ç³»äººé–“ã¯ãã‚“ãªã«å¤šããªã„ã§ã—ã‚‡
ishideo 2018/03/16
deep-learning

machine-learning

ai

python

tensorflow

ml
ãƒªãƒ³ã‚¯
å…¥é–€ Keras (5) å¦ç¿’æ¸ˆã¿ãƒ¢ãƒ‡ãƒ«ã¨ Flask ã§ API ã‚µãƒ¼ãƒ“ã‚¹ã‚’ä½œã‚‹ï½œæŠ€è¡“ãƒ–ãƒã‚°ï½œåŒ—æµ·é“æœå¹Œå¸‚ãƒ»å®®åŸŽçœŒä»™å°å¸‚ã®VRãƒ»ã‚²ãƒ¼ãƒ ãƒ»ã‚·ã‚¹ãƒ†ãƒ é–‹ç™º ã‚¤ãƒ³ãƒ•ã‚£ãƒ‹ãƒƒãƒˆãƒ«ãƒ¼ãƒ—
å…¥é–€ Keras (5) å¦ç¿’æ¸ˆã¿ãƒ¢ãƒ‡ãƒ«ã¨ Flask ã§ API ã‚µãƒ¼ãƒ“ã‚¹ã‚’ä½œã‚‹ é€£è¼‰ä¸€è¦§ å…¥é–€ Keras (1) Windows ã« Tensorflow ã¨ Keras ã‚’ã‚»ãƒƒãƒˆã‚¢ãƒƒãƒ— å…¥é–€ Keras (2) ãƒ‘ãƒ¼ã‚»ãƒ—ãƒˆãƒãƒ³ã¨ãƒã‚¸ã‚¹ãƒ†ã‚£ãƒƒã‚¯å›žå¸° å…¥é–€ Keras (3) ç·šå½¢åˆ†é›¢ã¨å¤šå±¤ãƒ‘ãƒ¼ã‚»ãƒ—ãƒˆãƒãƒ³ å…¥é–€ Keras (4) å¤šã‚¯ãƒ©ã‚¹åˆ†é¡ž â€“ Iris ãƒ‡ãƒ¼ã‚¿ã‚’å¦ç¿’ã™ã‚‹ å…¥é–€ Keras (5) å¦ç¿’æ¸ˆã¿ãƒ¢ãƒ‡ãƒ«ã¨ Flask ã§ API ã‚µãƒ¼ãƒ“ã‚¹ã‚’ä½œã‚‹ ç¬¬5å›žã¯æ¯›è‰²ã‚’å¤‰ãˆã¦ã€å¦ç¿’æ¸ˆã¿ã®ãƒ¢ãƒ‡ãƒ«ã¨ Python ã® Flask ãƒ•ãƒ¬ãƒ¼ãƒ ãƒ¯ãƒ¼ã‚¯ã‚’ä½¿ã£ã¦ã€ç°¡æ˜“çš„ãª API ã‚µãƒ¼ãƒ“ã‚¹ã‚’æ§‹ç¯‰ã—ã¾ã™ã€‚ ã¾ã ã¾ã ãƒ‡ã‚£ãƒ¼ãƒ—ãƒ©ãƒ¼ãƒ‹ãƒ³ã‚°ã®å…¥ã‚Šå£ã‚’ã¡ã‚‡ã£ã¨è¦—ã„ãŸã ã‘ã§ã™ãŒã€ã›ã£ã‹ããƒ‹ãƒ¥ãƒ¼ãƒ©ãƒ«ãƒãƒƒãƒˆã«å¦ç¿’ã•ã›ã‚‹ã“ã¨ãŒå‡ºæ¥ã‚‹ã‚ˆã†ã«ãªã‚Šã¾ã—ãŸã®ã§ã€è‡ªåˆ†ã® PC ä¸Šã® Jupyter Notebook ã§æ£è§£ï¼
ishideo 2018/02/28
python

keras

flask

api

tensorflow

webapi
ãƒªãƒ³ã‚¯
TensorFlowã§ç”»åƒã‚’TFRecordsã§èªã¿æ›¸ãã—ã¦ã¿ã‚‹ - SSJB's blog
TensorFlowã§CNNã‚’è©¦ã—ã¦ã„ã‚‹ã®ã§ã™ãŒã€æ•™å¸«ãƒ‡ãƒ¼ã‚¿ã®JPEGç”»åƒã‚’TFRecordså½¢å¼ã¸ã®å¤‰æ›ã¨èªã¿æ›¸ãã«å››è‹¦å…«è‹¦ã—ãŸã®ã§ã€ãã®ãƒ¡ãƒ¢ã€‚ ã‚‚ãã˜ ã¾ãˆãŠã TFRecordså½¢å¼ã«æ›¸ãå‡ºã— ä¸èº«ã‚’å¾©å…ƒã—ã¦ç¢ºèªã™ã‚‹ å‚è€ƒ ã¾ãˆãŠã å¦ç¿’æ™‚ã«ã„ã‚ã„ã‚ãªç”»åƒã‚µã‚¤ã‚ºã§è©¦ã—ãŸãã¦ã€1æžšã®ç”»åƒã«ã¤ãã€ 32x32 64x64 96x96 128x128 ã‚’ãã‚Œãžã‚Œç”¨æ„ã—ã¦ã„ãŸã®ã§ã™ãŒã€ã©ã†ã«ã‚‚é¢å€’ãã•ã„ã€‚ ã•ã‚‰ã«ã€ãƒ©ãƒ™ãƒ«ã¯è¨³ã‚ã£ã¦ç”»åƒåˆ†ã®XMLãƒ•ã‚¡ã‚¤ãƒ«ã‹ã‚‰å¦ç¿’æ™‚ã«æ¯Žå›žèªã¿è¾¼ã‚“ã§ã„ãŸã®ã§ã€èªã¿è¾¼ã¿ã«ã‹ãªã‚Šæ™‚é–“ãŒã‹ã‹ã£ã¦ã„ã¾ã—ãŸã€‚ è¨³ã¨ã„ã†ã®ã¯ã€æ•™å¸«ãƒ‡ãƒ¼ã‚¿ã‚’ä½œã‚‹ãŸã‚ã«ã€ã¾ãšæœ€åˆã«ã€SSD(Single Shot MultiBox Detector)ã§å¯¾è±¡ã®é ˜åŸŸã‚’æŽ¨æ¸¬ã—ã€ãã®æŽ¨æ¸¬ã•ã‚ŒãŸé ˜åŸŸï¼ˆground truth bounding boxï¼‰ã®ãƒ‡ãƒ¼ã‚¿ã¨ãƒ©ãƒ™ãƒ«ã‚’PASCAL VOCã®Annotati
ishideo 2017/09/14
tensorflow

tfrecords

python
ãƒªãƒ³ã‚¯
Haskellã§TensorFlowã€ã‚’å¯¾è©±çš„ã«è©¦ã—ã¦ã¿ã‚‹ - Qiita
ã¯ã˜ã‚ã« TensorFlowã®Haskellç”¨ãƒã‚¤ãƒ³ãƒ‡ã‚£ãƒ³ã‚°ãŒãƒªãƒªãƒ¼ã‚¹ã•ã‚Œã¦ã„ã¾ã™ã€‚ æœ¬æ ¼çš„ã«ä½¿ç”¨ã™ã‚‹å ´åˆã¯ã‚³ãƒ³ãƒ‘ã‚¤ãƒ«æ–¹å¼ã‚’ç”¨ã„ã‚‹ã¨æ€ã„ã¾ã™ãŒã€æœ¬è¨˜äº‹ã§ã¯ã€ç’°å¢ƒã«æ…£ã‚Œã‚‹ãŸã‚ã«å¯¾è©±çš„ã«ä½¿ç”¨ã™ã‚‹æ–¹æ³•ã«ã¤ã„ã¦ç°¡å˜ã«ç´¹ä»‹ã—ã¾ã™ã€‚ ä»¥ä¸‹ã§ã¯ã€Linux(ubuntu 16.04LTS 18.04LTS)ã‚’ç”¨ã„ãŸä¾‹ã‚’ç´¹ä»‹ã—ã¾ã™ã€‚ ãªãŠã€Dockerã¯ã‚¤ãƒ³ã‚¹ãƒˆãƒ¼ãƒ«æ¸ˆã¿ã§ã‚ã‚‹ã“ã¨ã‚’å‰æã¨ã—ã¦ã„ã¾ã™ã€‚ æœ¬è¨˜äº‹ã¯ã€2017å¹´8æœˆæ™‚ç‚¹ 2018å¹´8æœˆæ™‚ç‚¹ã®å†…å®¹ã§ã™ã€‚TensorFlowã®Haskellç”¨ãƒã‚¤ãƒ³ãƒ‡ã‚£ãƒ³ã‚°ã¯é–‹ç™ºä¸ã®ãŸã‚ã€ä»Šå¾Œã€ä½¿ã„å‹æ‰‹ãŒå¤‰ã‚ã‚‹ã¨æ€ã„ã¾ã™ã®ã§æ³¨æ„ã—ã¦ãã ã•ã„ã€‚ ä»¥ä¸‹ã¯å‚è€ƒè³‡æ–™ã§ã™ã€‚ Haskell bindings for TensorFlow Starting out with Haskell Tensor Flow ç’°å¢ƒã®æº–å‚™ 1. stackã‚³ãƒžãƒ³ãƒ‰ã®ã‚¤ãƒ³ã‚¹ãƒˆãƒ¼ãƒ« stackã‚’
ishideo 2017/08/24
haskell

tensorflow

qiita
ãƒªãƒ³ã‚¯
ã¾ãŸã‚‚ã‚„TensorFlowãŒå¼·åŒ–ï¼ï¼æ·±å±¤å¦ç¿’ãƒ©ã‚¤ãƒ–ãƒ©ãƒªã€Œsonnetã€ã®ç™»å ´ã€ä½¿ã£ã¦ã¿ãŸè¨˜äº‹ç´¹ä»‹ã‚’è¿½åŠ ã€‘ - HELLO CYBERNETICS
æ–°ãŸãªãƒ©ã‚¤ãƒ–ãƒ©ãƒªsonnet sonnetã¨ã¯ DeepMindç¤¾è£½ã§ã‚ã‚‹ã“ã¨ TensorFlowã¨å…±ã«ä½¿ãˆã‚‹ TensorFlow TensorFlowã®å½¹å‰² TensorFlowã®è¿½åŠ ãƒ©ã‚¤ãƒ–ãƒ©ãƒª Keras TensorFlow-Fold edward sonnet sonnetä½¿ã£ã¦ã¿ãŸè¨˜äº‹ã¾ã¨ã‚ å¤šåˆ†æ—¥æœ¬èªžæœ€é€Ÿã§å‡ºãŸä½¿ã£ã¦ã¿ãŸè¨˜äº‹ ã‚¤ãƒ³ã‚¹ãƒˆãƒ¼ãƒ«ã‹ã‚‰ä¸å¯§ã« ã‚³ãƒ¼ãƒ‰ã«å¯¾ã™ã‚‹ã‚³ãƒ¡ãƒ³ãƒˆãŒè±Šå¯Œã§å‚è€ƒã«ãªã‚‹è¨˜äº‹ æ–°ãŸãªãƒ©ã‚¤ãƒ–ãƒ©ãƒªsonnet sonnetã¨ã¯ DeepMindç¤¾ãŒç¤¾å†…ã§åˆ©ç”¨ã—ã¦ã„ãŸæ·±å±¤å¦ç¿’ãƒ©ã‚¤ãƒ–ãƒ©ãƒªã§ã™ã€‚ æ˜¨æ—¥ã“ã‚ŒãŒã‚ªãƒ¼ãƒ—ãƒ³ã‚½ãƒ¼ã‚¹åŒ–ã•ã‚Œã¾ã—ãŸã€‚ DeepMindç¤¾è£½ã§ã‚ã‚‹ã“ã¨ DeepMindã¯ç¾åœ¨Googleå‚˜ä¸‹ã®äººå·¥çŸ¥èƒ½é–‹ç™ºä¼æ¥ã§ã™ã€‚ åœ§å€’çš„ãªå°‚é–€å®¶é›†å›£ã®é›†ã¾ã‚Šã§ã‚ã‚Šã€æ·±å±¤å¦ç¿’ã®ç ”ç©¶ã«GoogleãŒä¹—ã‚Šåˆ‡ã‚‹éš›ã«ã€DeepMindç¤¾ã¯è²·åŽã•ã‚Œã¾ã—ãŸã€‚ãã®å¾Œã€è³‡é‡‘åŠ›ã¨ç ”ç©¶åŠ›ã‚’
ishideo 2017/06/09
deepmind

python

sonnet

tensorflow

tensorflow-fold
ãƒªãƒ³ã‚¯
1