{"id":59346,"date":"2024-02-23T18:46:33","date_gmt":"2024-02-23T10:46:33","guid":{"rendered":"https:\/\/www.ioiotimes.com\/?p=59346"},"modified":"2024-02-23T18:46:34","modified_gmt":"2024-02-23T10:46:34","slug":"%e5%85%b1%e7%b6%bb%e5%85%89%e8%8a%92%ef%bc%9a%e7%b6%93%e6%9c%80%e4%bd%b3%e5%8c%96%e8%aa%bf%e6%95%b4%e5%be%8c%e7%9a%84google-gemma%e6%a8%a1%e5%9e%8b%e5%8f%af%e5%9c%a8nvidia-gpu%e4%b8%8a%e9%81%8b","status":"publish","type":"post","link":"https:\/\/www.ioiotimes.com\/?p=59346","title":{"rendered":"\u5171\u7dbb\u5149\u8292\uff1a\u7d93\u6700\u4f73\u5316\u8abf\u6574\u5f8c\u7684Google Gemma\u6a21\u578b\u53ef\u5728NVIDIA GPU\u4e0a\u904b\u884c"},"content":{"rendered":"\n<h3 class=\"wp-block-heading\">Google\u65b0\u63a8\u51fa\u7684\u958b\u653e\u5f0f\u8a9e\u8a00\u6a21\u578b\u5728TensorRT-LLM\u7684\u52a0\u901f\u4e0b\uff0c\u53ef\u5728\u5305\u62ec\u672c\u5730\u7aefRTX AI PC\u7b49NVIDIA AI\u5e73\u53f0\u4e0a\u9ad8\u901f\u904b\u884c<\/h3>\n\n\n\n<p>NVIDIA \u8207 Google \u5408\u4f5c\u672c\u9031\u63a8\u51fa\u4e86\u5728\u6240\u6709 NVIDIA AI \u5e73\u53f0\u4e0a\u9069\u7528\u65bc <a href=\"https:\/\/blog.google\/technology\/developers\/gemma-open-models\/\" target=\"_blank\" rel=\"noreferrer noopener\">Gemma<\/a> \u6a21\u578b\u7684\u6700\u4f73\u5316\u529f\u80fd\u3002Gemma \u662f Google \u6700\u5148\u9032\u7684\u65b0\u6b3e\u8f15\u91cf\u7d1a\u958b\u653e\u5f0f\u8a9e\u8a00\u6a21\u578b\uff0c\u64c1\u6709 <a href=\"https:\/\/catalog.ngc.nvidia.com\/orgs\/nvidia\/teams\/ai-foundation\/models\/gemma-2b\" target=\"_blank\" rel=\"noreferrer noopener\">20 \u5104\u500b<\/a>\u548c <a href=\"https:\/\/catalog.ngc.nvidia.com\/orgs\/nvidia\/teams\/ai-foundation\/models\/gemma-7b\" target=\"_blank\" rel=\"noreferrer noopener\">70 \u5104\u500b<\/a>\u53c3\u6578\uff0c\u4e26\u53ef\u5728\u4efb\u4f55\u5730\u65b9\u904b\u884c\uff0c\u4e0d\u50c5\u53ef\u4ee5\u964d\u4f4e\u6210\u672c\uff0c\u4e5f\u80fd\u52a0\u5feb\u5728\u7279\u5b9a\u9818\u57df\u4f7f\u7528\u5834\u666f\u4e0a\u7684\u5275\u65b0\u3002<\/p>\n\n\n\n<figure class=\"wp-block-image size-full\"><img loading=\"lazy\" decoding=\"async\" width=\"1024\" height=\"576\" src=\"https:\/\/www.ioiotimes.com\/wordpress\/wp-content\/uploads\/2024\/02\/20240223nvidiagemma01.jpg\" alt=\"\" class=\"wp-image-59349\" title=\"\" srcset=\"https:\/\/www.ioiotimes.com\/wordpress\/wp-content\/uploads\/2024\/02\/20240223nvidiagemma01.jpg 1024w, https:\/\/www.ioiotimes.com\/wordpress\/wp-content\/uploads\/2024\/02\/20240223nvidiagemma01-300x169.jpg 300w, https:\/\/www.ioiotimes.com\/wordpress\/wp-content\/uploads\/2024\/02\/20240223nvidiagemma01-768x432.jpg 768w\" sizes=\"(max-width: 1024px) 100vw, 1024px\" \/><\/figure>\n\n\n\n<p><\/p>\n\n\n\n<p>NVIDIA \u8207 Google \u96d9\u65b9\u5718\u968a\u9032\u884c\u5bc6\u5207\u5408\u4f5c\uff0c\u900f\u904e\u9069\u7528\u65bc\u6700\u4f73\u5316\u5927\u578b\u8a9e\u8a00\u6a21\u578b\u63a8\u8ad6\u4f5c\u696d\u7684\u958b\u6e90\u51fd\u5f0f\u5eab <a href=\"https:\/\/github.com\/NVIDIA\/TensorRT-LLM\" target=\"_blank\" rel=\"noreferrer noopener\">NVIDIA TensorRT-LLM<\/a>\uff0c\u5728\u8cc7\u6599\u4e2d\u5fc3\u6216\u96f2\u7aef\u74b0\u5883\u904b\u884c\u7684 NVIDIA GPU\uff0c\u4ee5\u53ca\u642d\u8f09 <a href=\"https:\/\/www.nvidia.com\/zh-tw\/geforce\/rtx\/\" target=\"_blank\" rel=\"noreferrer noopener\">NVIDIA RTX<\/a> GPU \u7684 PC \u4e0a\uff0c\u52a0\u901f\u4e86 Gemma \u7684\u904b\u884c\u6548\u80fd\u3002\u503c\u5f97\u4e00\u63d0\u7684\u662f\uff0cGemma \u4f7f\u7528\u8207\u958b\u767c Gemini \u6a21\u578b\u76f8\u540c\u7684\u7814\u7a76\u6210\u679c\u548c\u6280\u8853\u3002<\/p>\n\n\n\n<p>\u5982\u6b64\u4e00\u4f86\uff0c\u958b\u767c\u8005\u4fbf\u80fd\u9396\u5b9a\u5168\u7403\u9ad8\u6548\u80fd AI PC \u4e0a\u53ef\u7528\u7684\u8d85\u904e\u4e00\u5104\u9846 NVIDIA RTX GPU \u7684\u5b89\u88dd\u57fa\u790e\u9032\u884c\u958b\u767c\u3002<\/p>\n\n\n\n<p>\u958b\u767c\u8005\u9084\u80fd\u4ee5\u96f2\u7aef\u74b0\u5883\u88e1\u7684 NVIDIA GPU \u904b\u884c Gemma \u6a21\u578b\uff0c\u5305\u62ec\u5728\u642d\u8f09 H100 Tensor \u6838\u5fc3 GPU \u7684 Google Cloud A3 \u5be6\u9ad4\u4e0a\u904b\u884c\uff0c\u4ee5\u53ca Google \u672a\u4f86\u5c07\u5f15\u5165\u7684 NVIDIA <a href=\"https:\/\/nvidianews.nvidia.com\/news\/nvidia-supercharges-hopper-the-worlds-leading-ai-computing-platform\" target=\"_blank\" rel=\"noreferrer noopener\">H200 Tensor \u6838\u5fc3 GPU<\/a>\uff0c\u8a72 GPU \u64c1\u6709  141GB HBM3e \u8a18\u61b6\u9ad4\uff0c\u6bcf\u79d2\u57f7\u884c\u901f\u5ea6\u70ba 4.8 TB\u3002<\/p>\n\n\n\n<p>\u4f01\u696d\u958b\u767c\u4eba\u54e1\u4e5f\u53ef\u4ee5\u904b\u7528 NVIDIA \u8c50\u5bcc\u7684\u5de5\u5177\u751f\u614b\u7cfb\u7d71\u4f86\u5fae\u8abf Gemma\uff0c\u5305\u62ec\u914d\u5099 <a href=\"https:\/\/github.com\/NVIDIA\/NeMo\" target=\"_blank\" rel=\"noreferrer noopener\">NeMo \u6846\u67b6<\/a>\u548c <a href=\"https:\/\/github.com\/NVIDIA\/TensorRT-LLM\" target=\"_blank\" rel=\"noreferrer noopener\">TensorRT-LLM<\/a> \u7684 <a href=\"https:\/\/www.nvidia.com\/zh-tw\/data-center\/products\/ai-enterprise\/\" target=\"_blank\" rel=\"noreferrer noopener\">NVIDIA AI Enterprise<\/a>\uff0c\u4e26\u4e14\u5728\u5176\u751f\u7522\u61c9\u7528\u7a0b\u5f0f\u4e2d\u90e8\u7f72\u7d93\u904e\u6700\u4f73\u5316\u8abf\u6574\u7684\u6a21\u578b\u3002<\/p>\n\n\n\n<p>\u6df1\u5165\u4e86\u89e3 <a href=\"https:\/\/developer.nvidia.com\/blog\/nvidia-tensorrt-llm-revs-up-inference-for-google-gemma\/\" target=\"_blank\" rel=\"noreferrer noopener\">TensorRT-LLM \u5982\u4f55\u52a0\u5feb Gemma \u7684\u63a8\u8ad6\u901f\u5ea6<\/a>\uff0c\u4ee5\u53ca\u66f4\u591a\u63d0\u4f9b\u7d66\u958b\u767c\u4eba\u54e1\u7684\u8cc7\u8a0a\u3002\u9019\u5305\u62ec Gemma \u7684\u591a\u500b\u6a21\u578b\u6aa2\u67e5\u9ede\u53ca\u6a21\u578b\u7684 FP8 \u91cf\u5316\u7248\u672c\uff0c\u5168\u90fd\u4f7f\u7528 TensorRT-LLM \u5b8c\u6210\u6700\u4f73\u5316\u8abf\u6574\u3002<\/p>\n\n\n\n<p>\u656c\u8acb\u4f7f\u7528\u7db2\u9801\u700f\u89bd\u5668\u958b\u555f NVIDIA AI Playground\uff0c\u4fbf\u80fd\u76f4\u63a5\u9ad4\u9a57 <a href=\"https:\/\/catalog.ngc.nvidia.com\/orgs\/nvidia\/teams\/ai-foundation\/models\/gemma-2b\" target=\"_blank\" rel=\"noreferrer noopener\">Gemma 2B<\/a>&nbsp;\u53ca&nbsp;<a href=\"https:\/\/catalog.ngc.nvidia.com\/orgs\/nvidia\/teams\/ai-foundation\/models\/gemma-7b\" target=\"_blank\" rel=\"noreferrer noopener\">Gemma 7B<\/a> \u7684\u5f37\u5927\u5a01\u529b\u3002<\/p>\n\n\n\n<p><\/p>\n\n\n\n<h3 class=\"wp-block-heading\"><strong>Chat With RTX\u5373\u5c07\u652f\u63f4Gemma<\/strong><\/h3>\n\n\n\n<p><a href=\"https:\/\/blogs.nvidia.com.tw\/2024\/02\/16\/chat-with-rtx-available-now\/\" target=\"_blank\" rel=\"noreferrer noopener\">Chat with RTX<\/a> \u662f\u4e00\u9805\u4f7f\u7528<a href=\"https:\/\/blogs.nvidia.com\/blog\/what-is-retrieval-augmented-generation\/\" target=\"_blank\" rel=\"noreferrer noopener\">\u6aa2\u7d22\u589e\u5f37\u751f\u6210<\/a>\u548c NVIDIA TensorRT-LLM \u8edf\u9ad4\u7684 NVIDIA \u6280\u8853\u5c55\u793a\u5167\u5bb9\uff0c\u8b93\u7528\u6236\u5728\u81ea\u5df1\u672c\u5730\u7aef\u6709\u642d\u8f09 RTX \u652f\u63f4\u7684 Windows PC \u4e0a\u5c31\u53ef\u4ee5\u4f7f\u7528\u751f\u6210\u5f0f\u4eba\u5de5\u667a\u6167\uff08AI\uff09\u529f\u80fd\u3002\u9019\u9805\u5de5\u5177\u4e5f\u5c07\u52a0\u5165\u652f\u63f4 Gemma\u3002<\/p>\n\n\n\n<p>Chat with RTX \u8b93\u7528\u6236\u53ef\u4ee5\u8f15\u9b06\u5c07 PC \u4e0a\u7684\u672c\u6a5f\u7aef\u6a94\u6848\u9023\u63a5\u5230\u5927\u578b\u8a9e\u8a00\u6a21\u578b\uff0c\u4f7f\u7528\u81ea\u5df1\u7684\u8cc7\u6599\u6253\u9020\u500b\u4eba\u5c08\u5c6c\u7684\u804a\u5929\u6a5f\u5668\u4eba\u3002<\/p>\n\n\n\n<p>\u7531\u65bc\u6a21\u578b\u4ee5\u672c\u6a5f\u7aef\u7684\u65b9\u5f0f\u904b\u884c\uff0c\u53ef\u4ee5\u5feb\u901f\u63d0\u4f9b\u904b\u884c\u7d50\u679c\uff0c\u4e26\u80fd\u5920\u8b93\u4f7f\u7528\u8005\u8cc7\u6599\u7559\u5728\u88dd\u7f6e\u4e0a\u3002Chat with RTX \u8207\u4f9d\u8cf4\u96f2\u7aef\u74b0\u5883\u7684 LLM \u670d\u52d9\u4e0d\u540c\uff0c\u8b93\u7528\u6236\u53ef\u4ee5\u5728\u672c\u5730\u7aef\u7684 PC \u4e0a\u8655\u7406\u654f\u611f\u8cc7\u6599\uff0c\u7121\u9700\u5c07\u8cc7\u6599\u5206\u4eab\u7d66\u7b2c\u4e09\u65b9\u6216\u662f\u9023\u63a5\u5230\u7db2\u8def\u3002<\/p>\n\n\n\n<p><\/p>\n\n\n\n<p><\/p>\n\n\n\n<h4 class=\"wp-block-heading has-text-align-right has-very-light-gray-to-cyan-bluish-gray-gradient-background has-background\">\ud83d\udfe6<a href=\"https:\/\/reurl.cc\/orq4Mg\" target=\"_blank\" rel=\"noopener\">\u770bPREDATOR Pallas II DDR5 \u6587\u7ae0 \u6436\u62ffSSD <\/a><br><strong>\ud83d\udfe6<\/strong>\u5feb\u4f86<a href=\"https:\/\/reurl.cc\/eLa7gM\" target=\"_blank\" rel=\"noopener\">\u53c3\u52a0\u6d3b\u52d5<\/a> \u9001\u60a890\u79d2\u7acb\u5373\u98df\u7528\u7684\u5fae\u6ce2\u767d\u98ef\u597d\u6ecb\u5473<br><strong>\ud83d\udfe6<\/strong>ioioTIMES\u300c\u5e74\u5ea6\u91d1\u8cde2023\u300d <a href=\"https:\/\/reurl.cc\/qrppny\" data-type=\"link\" data-id=\"https:\/\/reurl.cc\/qrppny\" target=\"_blank\" rel=\"noreferrer noopener\">\u770b\u6587\u7ae0\u3001\u5206\u4eab  \u62ff\u597d\u79ae!!<\/a>  <strong>~\u5feb\u4f86\u53c3\u52a0~<\/strong><br>\ud83d\udfe6<strong>\u73fe\u5728\u5c31\u52a0\u5165&nbsp;<a href=\"https:\/\/www.facebook.com\/profile.php?id=100086628162118\" target=\"_blank\" rel=\"noreferrer noopener\">ioioTIMES \u81c9\u66f8\u7c89\u7d72\u5718<\/a>&nbsp;\u66f4\u591a\u4e92\u52d5\u3001\u66f4\u591a\u597d\u5eb7\u650f\u62b5\u52a0!!<\/strong><br>\ud83d\udfe6<strong>\u6211\u5011\u6709<a href=\"https:\/\/today.line.me\/tw\/v2\/publisher\/103117\" target=\"_blank\" rel=\"noreferrer noopener nofollow\">LINE TODAY<\/a>\u983b\u9053\u4e86\uff0c\u5feb\u4f86\u8ffd\u8e2a\u6211\u5011\u5427!!&#8211;\u6700\u65b0\u79d1\u6280\u65b0\u805e \u76e1\u5728\u4f60\u624b<\/strong><\/h4>\n","protected":false},"excerpt":{"rendered":"<p>Google\u65b0\u63a8\u51fa\u7684\u958b\u653e\u5f0f\u8a9e\u8a00\u6a21\u578b\u5728<\/p>\n","protected":false},"author":3,"featured_media":59349,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"rank_math_lock_modified_date":false,"footnotes":""},"categories":[13],"tags":[8798,71,8799,7182],"class_list":["post-59346","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-news","tag-google-gemma","tag-nvidia","tag-rtx-ai-pc","tag-tensorrt-llm"],"_links":{"self":[{"href":"https:\/\/www.ioiotimes.com\/index.php?rest_route=\/wp\/v2\/posts\/59346"}],"collection":[{"href":"https:\/\/www.ioiotimes.com\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.ioiotimes.com\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.ioiotimes.com\/index.php?rest_route=\/wp\/v2\/users\/3"}],"replies":[{"embeddable":true,"href":"https:\/\/www.ioiotimes.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=59346"}],"version-history":[{"count":3,"href":"https:\/\/www.ioiotimes.com\/index.php?rest_route=\/wp\/v2\/posts\/59346\/revisions"}],"predecessor-version":[{"id":59350,"href":"https:\/\/www.ioiotimes.com\/index.php?rest_route=\/wp\/v2\/posts\/59346\/revisions\/59350"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.ioiotimes.com\/index.php?rest_route=\/wp\/v2\/media\/59349"}],"wp:attachment":[{"href":"https:\/\/www.ioiotimes.com\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=59346"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.ioiotimes.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=59346"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.ioiotimes.com\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=59346"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}