{"id":3586,"date":"2023-09-10T09:50:03","date_gmt":"2023-09-10T01:50:03","guid":{"rendered":""},"modified":"2023-09-10T09:50:03","modified_gmt":"2023-09-10T01:50:03","slug":"","status":"publish","type":"post","link":"http:\/\/www.idea2003.com\/dongtai\/3586.html","title":{"rendered":"\u5c0f\u5374\u5f3a\u5927\uff01TinyLlama\uff1a\u4ec5\u752890\u5929\u30013\u4e07\u4ebftoken\u8bad\u7ec3\u7684 550MB AI\u6a21\u578b"},"content":{"rendered":"

\u6587\u7ae0\u6982\u8981:<\/strong><\/p>\n

1. TinyLlama<\/span><\/a>:\u4e00\u6b3e\u5c0f\u578b\u4f46\u5f3a\u5927\u7684AI\u6a21\u578b\uff0c\u7531\u65b0\u52a0\u5761\u79d1\u6280\u8bbe\u8ba1\u5927\u5b66\u7684\u7814\u7a76\u56e2\u961f\u5f00\u53d1\uff0c\u4ec5\u5360\u7528550MB\u5185\u5b58\u3002<\/p>\n

2. \u8be5\u6a21\u578b\u91c7\u7528\u4e86\u7a33\u5b9a\u6269\u6563XL\u6280\u672f\uff0c\u8ba1\u5212\u572890\u5929\u5185\u57283\u4e07\u4ebf\u4ee3\u5e01\u7684\u6570\u636e\u96c6\u4e0a\u8fdb\u884c\u8bad\u7ec3\uff0c\u65e8\u5728\u9002\u5e94\u5185\u5b58\u53d7\u9650\u7684\u8fb9\u7f18\u8bbe\u5907\u3002<\/p>\n

3. TinyLlama\u7684\u6210\u529f\u5c06\u4e3a\u5b9e\u65f6\u673a\u5668\u7ffb\u8bd1\u7b49\u591a\u79cd\u5e94\u7528\u63d0\u4f9b\u9ad8\u6027\u80fdAI\u89e3\u51b3\u65b9\u6848\u3002<\/p>\n

\u5fae\u65b0\u521b\u60f3\uff08idea2003.com\uff099\u67088\u65e5 \u6d88\u606f:<\/strong>\u65b0\u52a0\u5761\u79d1\u6280\u8bbe\u8ba1\u5927\u5b66\u7684\u7814\u7a76\u4eba\u5458\u6b63\u8fdb\u884c\u4e00\u9879\u5f15\u4eba\u6ce8\u76ee\u7684\u9879\u76ee\uff0c\u4ed6\u4eec\u6b63\u5728\u52aa\u529b\u521b\u5efa\u4e00\u6b3e\u5c0f\u578b\u4f46\u5f3a\u5927\u7684AI\u6a21\u578b\uff0c\u8fd9\u4e2a\u6a21\u578b\u88ab\u79f0\u4e3aTinyLlama\uff0c\u5b83\u91c7\u7528\u4e86\u4e00\u79cd\u7d27\u51d1\u7684\u8bbe\u8ba1\uff0c\u4ec5\u5360\u7528550MB\u7684\u5185\u5b58\u3002<\/p>\n

\u8fd9\u4e00\u6a21\u578b\u7684\u72ec\u7279\u4e4b\u5904\u5728\u4e8e\uff0c\u5b83\u5c06\u57283\u4e07\u4ebftoken\u7684\u6570\u636e\u96c6\u4e0a\u8fdb\u884c\u8bad\u7ec3\uff0c\u8fd9\u4e2a\u5e9e\u5927\u7684\u6570\u636e\u96c6\u5c06\u5728\u77ed\u77ed90\u5929\u5185\u5b8c\u6210\u8bad\u7ec3\u3002<\/p>\n

\n

\u8fd9\u4e2a\u9879\u76ee\u7684\u76ee\u6807\u662f\u4e3a\u5185\u5b58\u53d7\u9650\u7684\u8fb9\u7f18\u8bbe\u5907\u63d0\u4f9b\u9ad8\u6027\u80fd\u7684\u4eba\u5de5\u667a\u80fd\u89e3\u51b3\u65b9\u6848\u3002\u8d8a\u6765\u8d8a\u591a\u7684\u5f00\u53d1\u8005\u8981\u6c42\u521b\u5efa\u66f4\u5c0f\u7684AI\u6a21\u578b\uff0c\u56e0\u4e3a\u5bf9\u4e8e\u5185\u5b58\u548c\u8ba1\u7b97\u80fd\u529b\u53d7\u9650\u7684\u8fb9\u7f18\u8bbe\u5907\u6765\u8bf4\uff0c\u8f83\u5c11\u7684\u53c2\u6570\u66f4\u52a0\u4f18\u5316\u3002\u6b64\u5916\uff0c\u8f83\u5c0f\u7684\u6a21\u578b\u8fd8\u53ef\u4ee5\u7528\u6765\u534f\u52a9\u89e3\u7801\u66f4\u5927\u7684\u6a21\u578b\uff0c\u6b63\u5982\u524d\u7279\u65af\u62c9\u9ad8\u7ea7<\/span>AI\u603b\u76d1Andrej Karpathy\u6240\u8a00\u3002<\/p>\n

TinyLlama\u9879\u76ee\u7531\u65b0\u52a0\u5761\u79d1\u6280\u8bbe\u8ba1\u5927\u5b66\u7684\u7814\u7a76\u52a9\u7406\u9886\u5bfc\uff0c\u4ed6\u4eec\u8bd5\u56fe\u5728\u4e09\u4e07\u4ebftoken\u7684\u6570\u636e\u96c6\u4e0a\u9884\u8bad\u7ec3\u4e00\u4e2a11\u4ebftoken\u7684Llama\u6a21\u578b\u3002\u8fd9\u4e2a\u6a21\u578b\u4ec5\u5360\u7528550MB\u7684\u5185\u5b58\uff0c\u56e2\u961f\u8ba4\u4e3a\u5b83\u7684\u7d27\u51d1\u6027\u5c06\u4f7f\u5b83\u80fd\u591f\u6ee1\u8db3\u591a\u79cd\u5e94\u7528\u7684\u9700\u6c42\uff0c\u8fd9\u4e9b\u5e94\u7528\u9700\u8981\u53d7\u9650\u7684\u8ba1\u7b97\u548c\u5185\u5b58\u5360\u7528\uff0c\u4ee5\u5b9e\u73b0\u8bf8\u5982\u65e0\u9700\u4e92\u8054\u7f51\u8fde\u63a5\u7684\u5b9e\u65f6\u673a\u5668\u7ffb\u8bd1\u7b49\u529f\u80fd\u3002<\/p>\n

TinyLlama\u7684\u8bad\u7ec3\u4e8e9\u67081\u65e5\u5f00\u59cb\uff0c\u4f7f\u7528\u4e8616\u4e2aA100-40G GPU\uff0c\u56e2\u961f\u8ba1\u5212\u5728\u77ed\u77ed90\u5929\u5185\u5b8c\u6210\u8bad\u7ec3\u3002\u622a\u6b62\u5230\u76ee\u524d\uff0c\u56e2\u961f\u5df2\u5b8c\u6210\u4e861050\u4ebftoken\u7684\u8bad\u7ec3\u3002<\/p>\n

\u6a21\u578b\u7684\u6784\u5efa\u8005\u8868\u793a\uff0c\u4ed6\u4eec\u6b63\u5728\u4f7f\u7528\u4e0eMeta\u7528\u4e8e\u8bad\u7ec3Llama2\u7684\u201c\u5b8c\u5168\u76f8\u540c\u7684\u67b6\u6784\u548c\u5206\u8bcd\u5668\u201d\uff0c\u56e0\u6b64\u5b83\u53ef\u4ee5\u8f7b\u677e\u5e94\u7528\u4e8e\u57fa\u4e8eLlama\u6784\u5efa\u7684\u5f00\u6e90\u9879\u76ee\u3002<\/p>\n

TinyLlama\u56e2\u961f\u6b63\u5728\u4f7f\u7528\u5305\u62ecCerebras Systems\u7684Slimpajama\u548cStarCoder\u6570\u636e\u5728\u5185\u7684\u4e09\u4e07\u4ebftoken\u6570\u636e\u96c6\uff0c\u8fd9\u4e9b\u6570\u636e\u96c6\u66fe\u7528\u4e8e\u8bad\u7ec3StarCoder\uff0c\u8fd9\u662f\u4e00\u4e2a\u4ee3\u7801\u751f\u6210\u6a21\u578b\u3002<\/p>\n

\u4e00\u65e6\u5b8c\u6210\uff0cTinyLlama\u5c06\u52a0\u5165\u8d8a\u6765\u8d8a\u591a\u7684\u8f83\u5c0f\u8bed\u8a00\u6a21\u578b\u7684\u884c\u5217\uff0c\u8fd9\u4e9b\u6a21\u578b\u7531\u5f00\u53d1\u8005\u7528\u4e8e\u6784\u5efa\u5404\u79cd\u5e94\u7528\u3002\u540c\u65f6\uff0cEleutherAI\u7684Pythia-1b\u548cDatabricks\u65d7\u4e0bMosaicML\u7684MPT-1b\u4e5f\u5728\u53d6\u5f97\u8fdb\u5c55\u3002<\/p>\n

\u9879\u76ee\u7f51\u5740:https:\/\/github.com\/eivindbohler\/tinyllama<\/p>\n","protected":false},"excerpt":{"rendered":"

\u6587\u7ae0\u6982\u8981: 1. TinyLlama:\u4e00\u6b3e\u5c0f\u578b\u4f46\u5f3a\u5927\u7684AI\u6a21\u578b\uff0c\u7531\u65b0...<\/p>\n","protected":false},"author":1,"featured_media":831,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[19],"tags":[],"class_list":["post-3586","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-dongtai"],"_links":{"self":[{"href":"http:\/\/www.idea2003.com\/wp-json\/wp\/v2\/posts\/3586","targetHints":{"allow":["GET"]}}],"collection":[{"href":"http:\/\/www.idea2003.com\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"http:\/\/www.idea2003.com\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"http:\/\/www.idea2003.com\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"http:\/\/www.idea2003.com\/wp-json\/wp\/v2\/comments?post=3586"}],"version-history":[{"count":0,"href":"http:\/\/www.idea2003.com\/wp-json\/wp\/v2\/posts\/3586\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"http:\/\/www.idea2003.com\/wp-json\/wp\/v2\/media\/831"}],"wp:attachment":[{"href":"http:\/\/www.idea2003.com\/wp-json\/wp\/v2\/media?parent=3586"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"http:\/\/www.idea2003.com\/wp-json\/wp\/v2\/categories?post=3586"},{"taxonomy":"post_tag","embeddable":true,"href":"http:\/\/www.idea2003.com\/wp-json\/wp\/v2\/tags?post=3586"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}