{"id":61585,"date":"2022-02-25T17:19:17","date_gmt":"2022-02-25T08:19:17","guid":{"rendered":"https:\/\/smilegate.ai\/?p=61585"},"modified":"2022-02-25T17:19:19","modified_gmt":"2022-02-25T08:19:19","slug":"an-open-source-framework-for-conversational-ai-nvidia-nemo","status":"publish","type":"post","link":"https:\/\/smilegate.ai\/cn\/2022\/02\/25\/an-open-source-framework-for-conversational-ai-nvidia-nemo\/","title":{"rendered":"An Open Source Framework for Conversational AI: NVIDIA NeMo"},"content":{"rendered":"<p>[\uac00\uc0c1\uc778\uac04\uc5f0\uad6c\ud300 \ud669\uc900\uc120]<\/p>\n\n\n\n<p><a href=\"https:\/\/developer.nvidia.com\/nvidia-nemo\" class=\"rank-math-link\" target=\"_blank\" rel=\"noopener\">NVIDIA NeMo<\/a>\ub294 \uac04\ub2e8\ud55c Python \uc778\ud130\ud398\uc774\uc2a4\ub97c \uc0ac\uc6a9\ud558\uc5ec GPU \uac00\uc18d \uc74c\uc131 \ubc0f \uc790\uc5f0\uc5b4 \uc774\ud574 \ubaa8\ub378\uc744 \uad6c\ucd95, training \ubc0f fine-tuning\ud558\uae30 \uc704\ud55c <a href=\"https:\/\/github.com\/NVIDIA\/NeMo\" class=\"rank-math-link\" target=\"_blank\" rel=\"noopener\">\uc624\ud508\uc18c\uc2a4 \ud504\ub808\uc784\uc6cc\ud06c<\/a>\uc785\ub2c8\ub2e4. NeMo\ub97c \ud65c\uc6a9\ud558\uba74 \uc2e4\uc2dc\uac04 \uc790\ub3d9 \uc74c\uc131 \uc778\uc2dd, \uc790\uc5f0\uc5b4 \ucc98\ub9ac, \ud14d\uc2a4\ud2b8 \uc74c\uc131 \ubcc0\ud658 \ub4f1 \uc560\ud50c\ub9ac\ucf00\uc774\uc158\uc6a9 \ubaa8\ub378\uc744 \uad6c\ucd95\ud560 \uc218 \uc788\ub2e4\uace0 \ud569\ub2c8\ub2e4.<\/p>\n\n\n\n<figure class=\"wp-block-embed-youtube wp-block-embed is-type-video is-provider-youtube wp-embed-aspect-16-9 wp-has-aspect-ratio\"><div class=\"wp-block-embed__wrapper\">\n<iframe title=\"NVIDIA NeMo: Developing State-Of-The-Art Conversational AI Models in Three Lines Of Code\" width=\"1200\" height=\"675\" src=\"https:\/\/www.youtube.com\/embed\/wBgpMf_KQVw?feature=oembed\" frameborder=\"0\" allow=\"accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture\" allowfullscreen><\/iframe>\n<\/div><\/figure>\n\n\n\n<p>NeMo\uc5d0\ub294 ASR, NLP \ubc0f TTS\uc6a9 \ub3c4\uba54\uc778 \ubcc4\ub85c \uc5ec\ub7ec \uae30\ub2a5\uc744 \ud3ec\ud568\ud558\uba70 Citrinet, Jasper, BERT, Fastpitch, HiFiGAN\uacfc \uac19\uc740 \ucd5c\ucca8\ub2e8 \ubaa8\ub378\uc744 \uac1c\ubc1c\ud560 \uc218 \uc788\uc2b5\ub2c8\ub2e4. NeMo\ub294 Neural Module\ub4e4\ub85c \uad6c\uc131\ub418\uba70, \uc774\ub7ec\ud55c \ubaa8\ub4c8\uc758 \uc785\ub825 \ubc0f \ucd9c\ub825\uc740 \ubaa8\ub4c8 \uac04\uc758 semantic check\ub97c \uc790\ub3d9\uc73c\ub85c \uc218\ud589\ud569\ub2c8\ub2e4. \uc774\ub97c \ub2e4\uc774\uc5b4\uadf8\ub7a8\uc73c\ub85c \ud45c\ud604\ud558\uba74 \uc544\ub798\uc758 \uadf8\ub9bc\uacfc \uac19\uc73c\uba70 <a href=\"https:\/\/github.com\/facebookresearch\/hydra\" class=\"rank-math-link\" target=\"_blank\" rel=\"noopener\">Hydra framework<\/a> \uae30\ubc18\uc73c\ub85c \uad6c\uc131\ub418\uc5b4 \uc788\uc2b5\ub2c8\ub2e4.<\/p>\n\n\n\n<figure class=\"wp-block-image\"><img decoding=\"async\" src=\"https:\/\/developer.nvidia.com\/sites\/default\/files\/akamai\/nemo\/nemo-architecture-2c50-d@2x.png?v2\" alt=\"Nemo New Model Architectures\"\/><\/figure>\n\n\n\n<p>\ub300\uaddc\ubaa8\uc758 Transformer \uae30\ubc18 \uc5b8\uc5b4 \ubaa8\ub378\uc740 \ud14d\uc2a4\ud2b8 \uc0dd\uc131, \uc694\uc57d \ubc0f \ucc57\ubd07\uacfc \uac19\uc740 \uad11\ubc94\uc704\ud55c \uc790\uc5f0\uc5b4 \uc791\uc5c5\uc5d0 \uc801\uc6a9\ub418\uace0 \uc788\uc2b5\ub2c8\ub2e4. NeMo Megatron(530B)\uc740 \ud6c8\ub828 \ub370\uc774\ud130\uc5d0 \ub300\ud574 \uc804\ucc98\ub9ac \uae30\ub2a5 \uc218\ud589 \ubc0f \ub370\uc774\ud130\ub97c \uc120\ubcc4\ud558\uace0 \ucd5c\ub300 \uc218\uc870 \uac1c\uc758 \ud30c\ub77c\ubbf8\ud130\ub97c \uac00\uc9c4 \ub300\uaddc\ubaa8 \ubaa8\ub378\uc744 \ud6c8\ub828\ud560 \uc218 \uc788\ub294 \uae30\ub2a5\uc744 \uc81c\uacf5\ud569\ub2c8\ub2e4. NeMo Megatron\uc758 \uc544\ud0a4\ud14d\uccd0\ub294 \uc544\ub798 \uadf8\ub9bc\uacfc \uac19\uc73c\uba70, \ud604\uc7ac <a href=\"https:\/\/developer.nvidia.com\/nemo-megatron-early-access\" class=\"rank-math-link\" target=\"_blank\" rel=\"noopener\">\uc5bc\ub9ac\uc5b5\uc138\uc2a4<\/a>\uac00 \uac00\ub2a5\ud569\ub2c8\ub2e4.<\/p>\n\n\n\n<figure class=\"wp-block-image\"><img decoding=\"async\" src=\"https:\/\/developer.nvidia.com\/sites\/default\/files\/akamai\/nemo\/nemo-megatron-1cn-d%402x.png\" alt=\"\"><\/figure>\n\n\n\n<p>\ub9c8\uc9c0\ub9c9\uc73c\ub85c, NeMo\uc758 \uacf5\uc2dd GitHub\uc744 \ud1b5\ud574 \uac04\ub2e8\ud55c <a href=\"https:\/\/docs.nvidia.com\/deeplearning\/nemo\/user-guide\/docs\/en\/stable\/starthere\/tutorials.html\" class=\"rank-math-link\" target=\"_blank\" rel=\"noopener\">\ud29c\ud1a0\ub9ac\uc5bc<\/a>\uc744 \uc81c\uacf5\ud558\uace0 \uc788\uc2b5\ub2c8\ub2e4. \uadf8 \uc911\uc5d0\uc11c, <a href=\"https:\/\/colab.research.google.com\/github\/NVIDIA\/NeMo\/blob\/stable\/tutorials\/AudioTranslationSample.ipynb\" class=\"rank-math-link\" target=\"_blank\" rel=\"noopener\">Audio Translation<\/a> \ud29c\ud1a0\ub9ac\uc5bc\uc774 \uc27d\uac8c \uc2e4\ud589\uc2dc\ud0ac \uc218 \uc788\uace0, \uc9c1\uc811 \ub4e4\uc5b4\ubcfc \uc218\ub3c4 \uc788\uae30 \ub54c\ubb38\uc5d0 \ud55c\ubc88 \uc2e4\ud589\uc2dc\ucf1c\ubcf4\ub294 \uac83\uc744 \ucd94\ucc9c\ub4dc\ub9bd\ub2c8\ub2e4.<\/p>\n\n\n\n<p>\u53c2\u8003<\/p>\n\n\n\n<p>[1] https:\/\/developer.nvidia.com\/nvidia-nemo<br>[2] https:\/\/github.com\/NVIDIA\/NeMo<br>[3] https:\/\/github.com\/facebookresearch\/hydra<\/p>\n<div class=\"likebtn_container\" style=\"\"><!-- LikeBtn.com BEGIN --><span class=\"likebtn-wrapper\"  data-identifier=\"post_61585\"  data-site_id=\"5f255b28943ec9a60bed08fd\"  data-lang=\"ko\"  data-dislike_enabled=\"false\"  data-icon_dislike_show=\"false\"  data-tooltip_enabled=\"false\"  data-white_label=\"true\"  data-rich_snippet=\"true\"  data-popup_disabled=\"true\"  data-i18n_like=\"Like\"  data-style=\"\"  data-unlike_allowed=\"\"  data-show_copyright=\"\"  data-item_url=\"https:\/\/smilegate.ai\/cn\/2022\/02\/25\/an-open-source-framework-for-conversational-ai-nvidia-nemo\/\"  data-item_title=\"An Open Source Framework for Conversational AI: NVIDIA NeMo\"  data-item_image=\"https:\/\/smilegate.ai\/wp-content\/uploads\/2022\/02\/state-of-the-art-convai-models-2c50-d@2x-1024x576.png\"  data-item_date=\"2022-02-25T17:19:17+09:00\"  data-engine=\"WordPress\"  data-plugin_v=\"2.6.55\"  data-prx=\"https:\/\/smilegate.ai\/wp-admin\/admin-ajax.php?action=likebtn_prx\"  data-event_handler=\"likebtn_eh\" ><\/span><!-- LikeBtn.com END --><\/div>","protected":false},"excerpt":{"rendered":"<p>[\uac00\uc0c1\uc778\uac04\uc5f0\uad6c\ud300 \ud669\uc900\uc120] NVIDIA NeMo\ub294 \uac04\ub2e8\ud55c Python \uc778\ud130\ud398\uc774\uc2a4\ub97c \uc0ac\uc6a9\ud558\uc5ec GPU \uac00\uc18d \uc74c\uc131 \ubc0f \uc790\uc5f0\uc5b4 \uc774\ud574 \ubaa8\ub378\uc744 \uad6c\ucd95, training \ubc0f fine-tuning\ud558\uae30 \uc704\ud55c \uc624\ud508\uc18c\uc2a4 \ud504\ub808\uc784\uc6cc\ud06c\uc785\ub2c8\ub2e4. NeMo\ub97c \ud65c\uc6a9\ud558\uba74 \uc2e4\uc2dc\uac04 \uc790\ub3d9 \uc74c\uc131 \uc778\uc2dd, \uc790\uc5f0\uc5b4 \ucc98\ub9ac, \ud14d\uc2a4\ud2b8 \uc74c\uc131 \ubcc0\ud658 \ub4f1 \uc560\ud50c\ub9ac\ucf00\uc774\uc158\uc6a9 \ubaa8\ub378\uc744 \uad6c\ucd95\ud560 \uc218 \uc788\ub2e4\uace0 \ud569\ub2c8\ub2e4. NeMo\uc5d0\ub294 ASR, NLP \ubc0f TTS\uc6a9 \ub3c4\uba54\uc778 \ubcc4\ub85c \uc5ec\ub7ec \uae30\ub2a5\uc744 \ud3ec\ud568\ud558\uba70 Citrinet, Jasper, BERT, Fastpitch, HiFiGAN\uacfc&hellip;<\/p>\n<div class=\"likebtn_container\" style=\"\"><!-- LikeBtn.com BEGIN --><span class=\"likebtn-wrapper\"  data-identifier=\"post_61585\"  data-site_id=\"5f255b28943ec9a60bed08fd\"  data-lang=\"ko\"  data-dislike_enabled=\"false\"  data-icon_dislike_show=\"false\"  data-tooltip_enabled=\"false\"  data-white_label=\"true\"  data-rich_snippet=\"true\"  data-popup_disabled=\"true\"  data-i18n_like=\"Like\"  data-style=\"\"  data-unlike_allowed=\"\"  data-show_copyright=\"\"  data-item_url=\"https:\/\/smilegate.ai\/cn\/2022\/02\/25\/an-open-source-framework-for-conversational-ai-nvidia-nemo\/\"  data-item_title=\"An Open Source Framework for Conversational AI: NVIDIA NeMo\"  data-item_image=\"https:\/\/smilegate.ai\/wp-content\/uploads\/2022\/02\/state-of-the-art-convai-models-2c50-d@2x-1024x576.png\"  data-item_date=\"2022-02-25T17:19:17+09:00\"  data-engine=\"WordPress\"  data-plugin_v=\"2.6.55\"  data-prx=\"https:\/\/smilegate.ai\/wp-admin\/admin-ajax.php?action=likebtn_prx\"  data-event_handler=\"likebtn_eh\" ><\/span><!-- LikeBtn.com END --><\/div>","protected":false},"author":1,"featured_media":61586,"comment_status":"closed","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"rank_math_lock_modified_date":false,"footnotes":""},"categories":[532,17,196],"tags":[],"class_list":["post-61585","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-nlp","category-tech02","category-code","category-532","category-17","category-196","description-off"],"_links":{"self":[{"href":"https:\/\/smilegate.ai\/cn\/wp-json\/wp\/v2\/posts\/61585","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/smilegate.ai\/cn\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/smilegate.ai\/cn\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/smilegate.ai\/cn\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/smilegate.ai\/cn\/wp-json\/wp\/v2\/comments?post=61585"}],"version-history":[{"count":2,"href":"https:\/\/smilegate.ai\/cn\/wp-json\/wp\/v2\/posts\/61585\/revisions"}],"predecessor-version":[{"id":61588,"href":"https:\/\/smilegate.ai\/cn\/wp-json\/wp\/v2\/posts\/61585\/revisions\/61588"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/smilegate.ai\/cn\/wp-json\/wp\/v2\/media\/61586"}],"wp:attachment":[{"href":"https:\/\/smilegate.ai\/cn\/wp-json\/wp\/v2\/media?parent=61585"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/smilegate.ai\/cn\/wp-json\/wp\/v2\/categories?post=61585"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/smilegate.ai\/cn\/wp-json\/wp\/v2\/tags?post=61585"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}