Tensorflowã§GPUを制é™ãƒ»ç„¡åŠ¹åŒ–ã™ã‚‹ - Qiita
Fri Oct 28 15:21:14 2016 +-----------------------------------------------------------------------------+ | NVIDIA-SMI 367.44 Driver Version: 367.44 | |-------------------------------+----------------------+----------------------+ | GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC | | Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. | |===============================+====
WebAssembly版ã®TensorFlow.jsãŒç™»å ´ã€Webブラウザã§ã®æŽ¨è«–処ç†ã‚’10å€ä»¥ä¸Šé«˜é€Ÿã«å®Ÿè¡Œ
WebAssembly版ã®TensorFlow.jsãŒç™»å ´ã€Webブラウザã§ã®æŽ¨è«–処ç†ã‚’10å€ä»¥ä¸Šé«˜é€Ÿã«å®Ÿè¡Œ GoogleãŒã‚ªãƒ¼ãƒ—ンソースã§å…¬é–‹ã—ã¦ã„る機械å¦ç¿’ライブラリTensorFlowã®1ã¤ã¨ã—ã¦ã€Webブラウザ上ã§æ©Ÿæ¢°å¦ç¿’ã®ãƒ¢ãƒ‡ãƒ«ã®æ§‹ç¯‰ã€å¦ç¿’ã€å¦ç¿’済ã¿ãƒ¢ãƒ‡ãƒ«ã«ã‚ˆã‚‹æŽ¨è«–ã®å®Ÿè¡Œãªã©ãŒå¯èƒ½ã«ãªã‚‹JavaScriptライブラリ「TensorFlow.jsã€ãŒã‚ã‚Šã¾ã™ã€‚ TensorFlow開発ãƒãƒ¼ãƒ ã¯ã€ã“ã®TensorFlow.jsã®å†…部ã§è¡Œã‚ã‚Œã¦ã„る処ç†ã‚’WebAssemblyã§å®Ÿè£…ã—ãŸã€ŒWebAssembly backend for TensorFlow.jsã€ã®ã‚¢ãƒ«ãƒ•ã‚¡ç‰ˆã‚’発表ã—ã¾ã—ãŸã€‚ We’re excited to release the Alpha of our WebAssembly backend for TensorFlow.js! WASM has wider dev
[Python]Kerasã‚’TensorFlowã‹ã‚‰ï¼ŒTensorFlowã‚’c++ã‹ã‚‰å©ã„ã¦å®Ÿè¡Œé€Ÿåº¦ã‚’上ã’ã‚‹ - Qiita
ã¯ã˜ã‚㫠深層å¦ç¿’ã¯å¦ç¿’自体ã«æ™‚é–“ãŒã‹ã‹ã‚‹ã®ã¯ã‚‚ã¡ã‚ã‚“ã®ã“ã¨ï¼Œè¨“練済ã¿å¦ç¿’モデルを走らã›ã‚‹éš›ã«ã‚‚ãã‚Œãªã‚Šã«æ™‚é–“ãŒã‹ã‹ã‚Šã¾ã™ï¼Žã—ã‹ã—，SSD（Single Shot MultiBox Detector）ç‰ã§ãƒªã‚¢ãƒ«ã‚¿ã‚¤ãƒ 物体èªè˜ã‚’ã•ã›ãŸã„ï¼ã¨ã‹ï¼ŒDQN(DeepQNetwork)ç‰ã§å¼·åŒ–å¦ç¿’ã—ãŸAIã¨äººã¨ã§ãƒªã‚¢ãƒ«ã‚¿ã‚¤ãƒ アクション対戦ゲームをã•ã›ãŸã„ã¨ã„ã£ãŸéš›ã«ã¯ï¼Œãƒ¢ãƒ‡ãƒ«å®Ÿè¡Œã®ãƒªã‚¢ãƒ«ã‚¿ã‚¤ãƒ 性ãŒã‹ãªã‚Šé‡è¦ã«ãªã£ã¦ãã¾ã™ï¼Ž ã„ã„PCè²·ãˆï¼ã¨ã„ãˆã°ãã‚Œã¾ã§ã§ã™ãŒãã‚“ãªãŠé‡‘ã‚‚ãªã„ã§ã™ã—，ノートPCç‰ã§ãƒãƒ¼ã‚¿ãƒ–ルã«å®Ÿè¡Œã—ãŸã„ã¨ã„ã†ã“ã¨ã‚‚ã‚ã‚‹ã§ã—ょã†ï¼Žãã“ã§ä»Šå›žã¯Keras(TensorFlow)ã‚’ã„ã‹ã«é«˜é€Ÿã«å®Ÿè¡Œã™ã‚‹ã‹ã«ã¤ã„ã¦è€ƒãˆã¦ã„ãã¾ã™ï¼Ž 実行環境 Python3.5.2 Keras 1.2.1 tensorflow 1.0.0 MacBookPro(Late 2013) 高速化 ã§ã¯ã‚„ã£ã¦ã„ãã¾ã—ょ
![[Python]Kerasã‚’TensorFlowã‹ã‚‰ï¼ŒTensorFlowã‚’c++ã‹ã‚‰å©ã„ã¦å®Ÿè¡Œé€Ÿåº¦ã‚’上ã’ã‚‹ - Qiita](https://cdn-ak-scissors.b.st-hatena.com/image/square/82869e72c0767fc0a8421bf587458eb79ffe4146/height=288;version=1;width=512/https%3A%2F%2Fqiita-user-contents.imgix.net%2Fhttps%253A%252F%252Fqiita-user-contents.imgix.net%252Fhttps%25253A%25252F%25252Fcdn.qiita.com%25252Fassets%25252Fpublic%25252Farticle-ogp-background-afbab5eb44e0b055cce1258705637a91.png%253Fixlib%253Drb-4.0.0%2526w%253D1200%2526blend64%253DaHR0cHM6Ly9xaWl0YS11c2VyLXByb2ZpbGUtaW1hZ2VzLmltZ2l4Lm5ldC9odHRwcyUzQSUyRiUyRnFpaXRhLWltYWdlLXN0b3JlLnMzLmFtYXpvbmF3cy5jb20lMkYwJTJGNTk4NjQlMkZwcm9maWxlLWltYWdlcyUyRjE0NzM2OTUwNTg_aXhsaWI9cmItNC4wLjAmYXI9MSUzQTEmZml0PWNyb3AmbWFzaz1lbGxpcHNlJmZtPXBuZzMyJnM9YTE3N2MyNGEwNmJhYmVhMWE0ZDUwMmNmNWM5MGVjZTI%2526blend-x%253D120%2526blend-y%253D467%2526blend-w%253D82%2526blend-h%253D82%2526blend-mode%253Dnormal%2526s%253Dc8dd32b31fae03c70963421bf3671e4f%3Fixlib%3Drb-4.0.0%26w%3D1200%26fm%3Djpg%26mark64%3DaHR0cHM6Ly9xaWl0YS11c2VyLWNvbnRlbnRzLmltZ2l4Lm5ldC9-dGV4dD9peGxpYj1yYi00LjAuMCZ3PTk2MCZoPTMyNCZ0eHQ9JTVCUHl0aG9uJTVES2VyYXMlRTMlODIlOTJUZW5zb3JGbG93JUUzJTgxJThCJUUzJTgyJTg5JUVGJUJDJThDVGVuc29yRmxvdyVFMyU4MiU5MmMlMkIlMkIlRTMlODElOEIlRTMlODIlODklRTUlOEYlQTklRTMlODElODQlRTMlODElQTYlRTUlQUUlOUYlRTglQTElOEMlRTklODAlOUYlRTUlQkElQTYlRTMlODIlOTIlRTQlQjglOEElRTMlODElOTIlRTMlODIlOEImdHh0LWFsaWduPWxlZnQlMkN0b3AmdHh0LWNvbG9yPSUyMzFFMjEyMSZ0eHQtZm9udD1IaXJhZ2lubyUyMFNhbnMlMjBXNiZ0eHQtc2l6ZT01NiZ0eHQtcGFkPTAmcz0zMjNkYzdhZDM1YzQ3YzkwZGY4OWY0NjdjM2ZlNjk5YQ%26mark-x%3D120%26mark-y%3D112%26blend64%3DaHR0cHM6Ly9xaWl0YS11c2VyLWNvbnRlbnRzLmltZ2l4Lm5ldC9-dGV4dD9peGxpYj1yYi00LjAuMCZ3PTgzOCZoPTU4JnR4dD0lNDB5dWtpQiZ0eHQtY29sb3I9JTIzMUUyMTIxJnR4dC1mb250PUhpcmFnaW5vJTIwU2FucyUyMFc2JnR4dC1zaXplPTM2JnR4dC1wYWQ9MCZzPTQwNDMzNzY3OWM3ZGFlZDg1NWQ2ZmFkZmQ5NTQ2NTYy%26blend-x%3D242%26blend-y%3D454%26blend-w%3D838%26blend-h%3D46%26blend-fit%3Dcrop%26blend-crop%3Dleft%252Cbottom%26blend-mode%3Dnormal%26txt64%3DaW4g5qCq5byP5Lya56S-44OG44Oz44Kv44O8%26txt-x%3D242%26txt-y%3D539%26txt-width%3D838%26txt-clip%3Dend%252Cellipsis%26txt-color%3D%25231E2121%26txt-font%3DHiragino%2520Sans%2520W6%26txt-size%3D28%26s%3De1ec73c44c1fdf3971a8526efbc9c9e4)
TensorFlowã®ã‚¯ãƒã‚¹ã‚¨ãƒ³ãƒˆãƒãƒ”ー関数ã®å‹•ä½œ - Qiita
ディープラーニングã®å¤šã‚¯ãƒ©ã‚¹åˆ†é¡žã§ã¯$x_{d,i}$ã«ã‚½ãƒ•ãƒˆãƒžãƒƒã‚¯ã‚¹é–¢æ•°ã®å‡ºåŠ›ã‚’ã€$y_{d,i}$ã«æ•™å¸«ãƒ‡ãƒ¼ã‚¿ã‚’ãã‚Œãžã‚Œå…¥åŠ›ã—ã¾ã™ã€‚ TensorFlowã§ã¯ ソフトマックス関数ã¯tf.nn.softmaxã¨ã„ã†é–¢æ•°ã§è¨ˆç®—ã§ãã¾ã™ã€‚ クãƒã‚¹ã‚¨ãƒ³ãƒˆãƒãƒ”ーã«ã¤ã„ã¦ã¯TensorFlowã§ã¯å˜ä½“ã®é–¢æ•°ã¯ç”¨æ„ã•ã‚Œã¦ã„ãªã„よã†ãªã®ã§ã€ã‚½ãƒ•ãƒˆãƒžãƒƒã‚¯ã‚¹é–¢æ•°ã¨ã‚¯ãƒã‚¹ã‚¨ãƒ³ãƒˆãƒãƒ”ーをセットã«ã—ãŸã€tf.nn.softmax_cross_entropy_with_logits_v2ã‚„tf.nn.sparse_softmax_cross_entropy_with_logitsを使ã„ã¾ã™ã€‚"with_logits"ã¨ã„ã†ã®ã¯ã€ã“ã®é–¢æ•°ã¯å†…部ã§ã‚½ãƒ•ãƒˆãƒžãƒƒã‚¯ã‚¹ã‚‚計算ã™ã‚‹ã‹ã‚‰ã€ãƒ‹ãƒ¥ãƒ¼ãƒ©ãƒ«ãƒãƒƒãƒˆãƒ¯ãƒ¼ã‚¯ã®å‡ºåŠ›ã‚’ãã®ã¾ã¾å…¥åŠ›ã—ã¦ãã€ã¨ã„ã†æ„味ã§ã™ã€‚ （å¤ã„ãƒãƒ¼ã‚¸ãƒ§ãƒ³ã§ã¯_v2ã¨ã„ã†åå‰ã§ã¯ãªã‹ã£ãŸã®ã§ã™ãŒã€æœ€æ–°ç‰ˆã§ã¯_v2
Tensorflowã§CNNを作る際ã«ä½¿ã„ãã†ãªé–¢æ•°ã‚’列挙ã—ã¦ã¿ãŸ - Qiita
ã¯ã˜ã‚ã« GoogleãŒä½œæˆã—ãŸDeepLearningフレームワークã®Tensorflow ã„ã‚ã„ã‚記事ãŒä¸ŠãŒã£ã¦ã„ã¦éžå¸¸ã«é¢ç™½ã„ã§ã™ãŒã€å®Ÿéš›ã«NNを組む際ã«ä½¿ç”¨ã™ã‚‹é–¢æ•°ã¯ã©ã‚Œï¼Ÿã¨ã„ã†ã®ã‚’備忘ã¨ã—ã¦ã¾ã¨ã‚ã¦ã¿ã¾ã—㟠ãªãŠç†è€…ã¯Deeplearningを大å¦ã§å¦ã‚“ã ã‚ã‘ã§ã¯ãªã自己å¦ç¿’ã—ãŸã ã‘ãªã®ã§ã€é–“é•ã„も多々ã‚ã‚‹ã‹ã¨æ€ã„ã¾ã™ãŒã€å¾Œå¦ã®ãŸã‚指摘ã„ãŸã ã‘ã‚‹ã¨å¹¸ã„ã§ã™ 畳込ã¿å±¤ã§ä½¿ç”¨ tf.nn.conv2d(input, filter, strides, padding, use_cudnn_on_gpu=None, name=None) 畳込ã¿ã‚’è¡Œã†é–¢æ•° ä»–ã«ã¯depthwise_conv2d, separable_conv2dãŒå˜åœ¨ã™ã‚‹ 第1引数: input インプットデータをã€4次元([batch, in_height, in_width, in_channels])ã®ãƒ†ãƒ³ã‚½ãƒ«ã‚’渡
Tensorflowã®ä¾¿åˆ©ãƒ©ã‚¤ãƒ–ラリTF-Slim - Qiita
with slim.arg_scope([slim.ops.conv2d, slim.ops.fc], stddev=0.01, weight_decay=0.0005): net = slim.ops.repeat_op(2, inputs, slim.ops.conv2d, 64, [3, 3], scope='conv1') net = slim.ops.max_pool(net, [2, 2], scope='pool1') net = slim.ops.repeat_op(2, net, slim.ops.conv2d, 128, [3, 3], scope='conv2') net = slim.ops.max_pool(net, [2, 2], scope='pool2') net = slim.ops.repeat_op(3, net, slim.ops.conv2d, 2
ã‚らゆるデータをå¯è¦–化ã™ã‚‹TensorBoard徹底入門
TensorBoardã®ä¸»è¦æ©Ÿèƒ½ 折れ線グラフ ç”»åƒ éŸ³å£° ヒストグラム計算グラフ 次元削減ã®ãƒ—ãƒãƒƒãƒˆ TensorBoardã®èªã¿æ–¹ シンボルã®æ„味 name scopeã¨node グラフã®è‰² Structure View Device View 計算時間・メモリ テンソルã®æ¬¡å…ƒæ•° Summary Operation scalar histogram image audio å¯è¦–化ã—ã¦ã¿ã‚‹ ãƒã‚¤ãƒ‘ーパラメータã®æŽ¢ç´¢ Embedding Visualization メタデータファイルã®ä½œã‚Šæ–¹ スプライト画åƒã®ä½œã‚Šæ–¹ ラベルã«ã‚ˆã‚‹è‰²åˆ†ã‘ t-SNEã¨PCA ã¾ã¨ã‚ TensorFlowã®å„ªã‚ŒãŸæ©Ÿèƒ½ã¨ã—ã¦ã€TensorBoardã«ã‚ˆã‚‹å……実ã—ãŸå¯è¦–化環境ãŒæŒ™ã’られã¾ã™ã€‚TensorBoardãŒã‚ã‚Œã°ã€ãƒ‹ãƒ¥ãƒ¼ãƒ©ãƒ«ãƒãƒƒãƒˆãƒ¯ãƒ¼ã‚¯ã®å¦ç¿’ãŒä¸Šæ‰‹ãã„ã‹ãªã„ã¨ãã«ã€ä¿¯çž°ã—ã¦ãƒãƒƒãƒˆãƒ¯ãƒ¼ã‚¯ã‚’表示ã—ãŸã‚Šã€æ§˜ã€…ãªãƒ‡
TensorFlowã®ãƒãƒ¥ãƒ¼ãƒˆãƒªã‚¢ãƒ«ã‚’通ã—ã¦ã€äººå·¥çŸ¥èƒ½ã®åŽŸç†ã«ã¤ã„ã¦å¦ç¿’ã™ã‚‹ - Qiita
Deleted articles cannot be recovered. Draft of this article would be also deleted. Are you sure you want to delete this article? 以下ã®ã“ã¨ã«ã¤ã„ã¦çŸ¥ã‚ŠãŸã„人å‘ã‘ã®è¨˜äº‹ã€‚ ①機械å¦ç¿’ã®åˆ©ç‚¹ã¨æ¬ 点ã€TensorFlowã®åˆ©ç‚¹ã¨æ¬ 点 ②ニューラルãƒãƒƒãƒˆãƒ¯ãƒ¼ã‚¯ã®åŽŸç†ã«ã¤ã„ã¦ã®ç°¡æ½”ãªèª¬æ˜Ž (オンライン書ç±ï¼šhttp://nnadl-ja.github.io/nnadl_site_ja/index.html ã‚’ã¾ã¨ã‚ãŸã‚‚ã®) â‘¢TensorFlowã®ãƒãƒ¥ãƒ¼ãƒˆãƒªã‚¢ãƒ«ã«æ²¿ã£ã¦ãƒ—ãƒã‚°ãƒ©ãƒ 作æˆã€ãã®å®Ÿè¡Œçµæžœ 以下ã€å‚考ã«ã—ãŸã‚µã‚¤ãƒˆ 1次関数ã®è¿‘似： https://www.tensorflow.org/versions/master/get_started/index.html h
Googleã®å…¬é–‹ã—ãŸäººå·¥çŸ¥èƒ½ãƒ©ã‚¤ãƒ–ラリTensorFlowを触ã£ã¦ã¿ãŸ - LIFULL Creators Blog
ã“ã‚“ã«ã¡ã¯ã€‚ãŠã†ã¡ãƒãƒƒã‚«ãƒ¼ã®çŸ³ç”°ã§ã™ã€‚ ã„ã¤ã‚‚ã¯ãŠã†ã¡ãƒãƒƒã‚¯ãƒã‚¿ã°ã‹ã‚Šã§ã™ãŒã€ä»Šæ—¥ã¯äººå·¥çŸ¥èƒ½é–¢é€£ã®è©±é¡Œã§ã™ã€‚ 今日2015/11/10ã€GoogleãŒè‡ªç¤¾ã‚µãƒ¼ãƒ“スã§ä½¿ã£ã¦ã„ã‚‹DeepLearningを始ã‚ã¨ã™ã‚‹æ©Ÿæ¢°å¦ç¿’技術ã®ãƒ©ã‚¤ãƒ–ラリを公開ã—ã¾ã—ãŸã€‚ TensorFlowã¨ã„ã†åå‰ã§ã€ãŠãらãテンソルフãƒãƒ¼ã¨å‘¼ã³ã¾ã™ã€‚ テンソルã¯ã€æ•°å¦ã®ç·šå½¢ã®é‡ã‚’表ã™æ¦‚念ã§ã€ãƒ™ã‚¯ãƒˆãƒ«ã®è¦ªæˆšã¿ãŸã„ãªã‚‚ã®ã§ã™ã€‚ãã‚Œã«ãƒ•ãƒãƒ¼ã‚’ã¤ã‘ã‚‹ã¨ã„ã†ã“ã¨ã¯ã€ãã†ã„ã£ãŸè¤‡é›‘ãªå¤šæ¬¡å…ƒãƒ™ã‚¯ãƒˆãƒ«é‡ã‚’æµã‚Œã‚‹ã‚ˆã†ã«å‡¦ç†ã§ãã‚‹ã€ã¨ã„ã†æ„味ãŒè¾¼ã‚られã¦ã„ã‚‹ã®ã ã¨æ€ã„ã¾ã™ã€‚ ã“ã¡ã‚‰ã‚’ã•ã£ãã触ã£ã¦ã¿ãŸã®ã§ã€ç´¹ä»‹ã—ãŸã„ã¨æ€ã„ã¾ã™ã€‚ TensorFlowã®ç‰¹å¾´ å…¬å¼ç´¹ä»‹ãƒšãƒ¼ã‚¸ã‹ã‚‰ç‰¹å¾´ã‚’ã„ãã¤ã‹ãƒ”ックアップã—ã¾ã™ã€‚ Deep Flexibility ~æ·±ã„柔軟性~ è¦æœ›ã«å¿œã˜ã¦ã€æŸ”軟ã«ãƒ‹ãƒ¥ãƒ¼ãƒ©ãƒ«ãƒãƒƒãƒˆãƒ¯ãƒ¼ã‚¯ã‚’構築ã§ãã¾ã™ã€‚ニューラルãƒãƒƒãƒˆãƒ¯ãƒ¼ã‚¯ã®
TensorFlow Tutorial ã§ãƒ‹ãƒ¥ãƒ¼ãƒ©ãƒ«ãƒãƒƒãƒˆãƒ¯ãƒ¼ã‚¯ã‚’使ã†[Python] - よーãã‚‹ã¨ã®ãƒ–ãƒã‚°
GoogleãŒç™ºè¡¨ã—ãŸäººå·¥çŸ¥èƒ½ãƒ©ã‚¤ãƒ–ラリTensorFlowã‚’å‹•ã‹ã—ã¦ã¿ã¾ã—ãŸã€‚ ãƒãƒ¥ãƒ¼ãƒˆãƒªã‚¢ãƒ«ã‚’追ã„ãªãŒã‚‰ã€ãƒ‹ãƒ¥ãƒ¼ãƒ©ãƒ«ãƒãƒƒãƒˆãƒ¯ãƒ¼ã‚¯ã®å‹‰å¼·(復習)ã‚’ã—ãŸã®ã§ç†è§£ã¯æ·±ã¾ã‚Šã¾ã—ãŸãŒã€èª¤ã‚Šã‚„é–“é•ã„ã‚‚ã‚ã‚‹ã¨æ€ã„ã¾ã™ã€‚見ã¤ã‘ãŸå ´åˆã¯ã”指摘ãã ã•ã„。 Download & Install MacOSã§ã®ãƒ€ã‚¦ãƒ³ãƒãƒ¼ãƒ‰ã¯æ¬¡ã®ã‚ˆã†ã«ãªã‚Šã¾ã™ã€‚Ubuntu/Linuxã ã¨GPU使ã†ãƒãƒ¼ã‚¸ãƒ§ãƒ³ã‚‚ダウンãƒãƒ¼ãƒ‰ã§ãる模様 ダウンãƒãƒ¼ãƒ‰ãƒšãƒ¼ã‚¸ # Only CPU-version is available at the moment. $ pip install https://storage.googleapis.com/tensorflow/mac/tensorflow-0.5.0-py2-none-any.whl å‹•ä½œç¢ºèª $ python >>> import tensorflow as tf >>> hel
![TensorFlow Tutorial ã§ãƒ‹ãƒ¥ãƒ¼ãƒ©ãƒ«ãƒãƒƒãƒˆãƒ¯ãƒ¼ã‚¯ã‚’使ã†[Python] - よーãã‚‹ã¨ã®ãƒ–ãƒã‚°](https://cdn-ak-scissors.b.st-hatena.com/image/square/de77eed1d6343af5111ac5c51f4a1d6f85852932/height=288;version=1;width=512/https%3A%2F%2Fcdn.image.st-hatena.com%2Fimage%2Fscale%2Ff2e66a418f35c03bc5d45eabd630cebcb62bec8e%2Fbackend%3Dimagemagick%3Bversion%3D1%3Bwidth%3D1300%2Fhttp%253A%252F%252Fcdn-ak.f.st-hatena.com%252Fimages%252Ffotolife%252Fy%252Fyoghurt1131%252F20151112%252F20151112021353.jpg)
1
リリースã€éšœå®³æƒ…å ±ãªã©ã®ã‚µãƒ¼ãƒ“スã®ãŠçŸ¥ã‚‰ã›
最新ã®äººæ°—エントリーã®é…ä¿¡
j次ã®ãƒ–ックマーク
kå‰ã®ãƒ–ックマーク
lã‚ã¨ã§èªã‚€
eコメント一覧を開ã
oページを開ã
{{#tags}}- {{label}}
{{/tags}}