{"id":15235,"date":"2024-12-11T10:12:00","date_gmt":"2024-12-11T02:12:00","guid":{"rendered":"https:\/\/www.aisharenet.com\/?p=15235"},"modified":"2025-08-26T00:23:30","modified_gmt":"2025-08-25T16:23:30","slug":"llama-ocr","status":"publish","type":"post","link":"https:\/\/www.kdjingpai.com\/ja\/llama-ocr\/","title":{"rendered":"Llama OCR\uff1a\u5229\u7528\u514d\u8d39Llama 3.2 Vision\u63a5\u53e3\uff0c\u4e09\u884c\u4ee3\u7801\u5c06\u56fe\u50cf\u8f6c\u6362\u4e3aMarkdown\u7684OCR\u5e93"},"content":{"rendered":"

Llama OCR \u662f\u4e00\u4e2a\u57fa\u4e8e Llama 3.2 Vision \u7684 OCR\uff08\u5149\u5b66\u5b57\u7b26\u8bc6\u522b\uff09\u5e93\uff0c\u80fd\u591f\u5c06\u6587\u6863\u8f6c\u6362\u4e3a Markdown \u683c\u5f0f\u3002\u8be5\u5e93\u7531 Nutlope \u5f00\u53d1\uff0c\u4f7f\u7528 Together<\/a> AI \u63d0\u4f9b\u7684\u514d\u8d39 Llama 3.2 \u63a5\u53e3\u8fdb\u884c\u56fe\u50cf\u89e3\u6790\uff0c\u5e76\u8fd4\u56de Markdown \u6587\u672c\u3002Llama OCR \u652f\u6301\u672c\u5730\u548c\u8fdc\u7a0b\u56fe\u50cf\u7684 OCR \u5904\u7406\uff0c\u672a\u6765\u8fd8\u8ba1\u5212\u652f\u6301 PDF \u6587\u4ef6\u7684 OCR\u3002\u7528\u6237\u53ef\u4ee5\u901a\u8fc7 npm \u5b89\u88c5\u8be5\u5e93\uff0c\u5e76\u5728\u9879\u76ee\u4e2d\u8f7b\u677e\u8c03\u7528\u5176\u529f\u80fd\u3002<\/p>\n

\u53c2\u8003\u9879\u76ee\uff1a\u00a0Zerox<\/a><\/p>\n

\"Llama

\u6f14\u793a\uff1ahttps:\/\/llamaocr.com\/<\/p><\/div>\n

 <\/p>\n

\"Llama

\u4f9d\u8d56together\u63d0\u4f9b\u7684Meta Llama Vision\u514d\u8d39\u63a5\u53e3\uff1ahttps:\/\/api.together.ai\/models\/meta-llama\/Llama-Vision-Free<\/p><\/div>\n

 <\/p>\n

\u66f4\u591a\u514d\u8d39\u89c6\u89c9\u6a21\u578b\uff1a\u667a\u8c31\u5f00\u653e\u5e73\u53f0\uff0c\u4e0a\u7ebf\u9996\u6b3e\u514d\u8d39\u591a\u6a21\u6001\u89c6\u89c9\u6a21\u578bGLM-4V-Flash\uff0c\u4e0d\u9650\u91cf\u4f7f\u7528\uff01<\/a><\/p>\n

 <\/p>\n

\u529f\u80fd\u5217\u8868<\/h2>\n
    \n
  • \u56fe\u50cfOCR<\/strong>\uff1a\u652f\u6301\u672c\u5730\u548c\u8fdc\u7a0b\u56fe\u50cf\u7684\u5149\u5b66\u5b57\u7b26\u8bc6\u522b\u3002<\/li>\n
  • Markdown\u8f93\u51fa<\/strong>\uff1a\u5c06\u8bc6\u522b\u7684\u6587\u672c\u8f6c\u6362\u4e3a Markdown \u683c\u5f0f\u3002<\/li>\n
  • \u591a\u6a21\u578b\u652f\u6301<\/strong>\uff1a\u63d0\u4f9b\u514d\u8d39\u548c\u4ed8\u8d39\u7684 Llama 3.2 \u6a21\u578b\u63a5\u53e3\uff0c\u6ee1\u8db3\u4e0d\u540c\u6027\u80fd\u9700\u6c42\u3002<\/li>\n
  • API\u96c6\u6210<\/strong>\uff1a\u901a\u8fc7 Together AI \u7684 API \u8fdb\u884c\u56fe\u50cf\u89e3\u6790\u3002<\/li>\n
  • \u672a\u6765\u529f\u80fd<\/strong>\uff1a\u8ba1\u5212\u652f\u6301\u5355\u9875\u548c\u591a\u9875 PDF \u7684 OCR \u5904\u7406\uff0c\u4ee5\u53ca JSON \u683c\u5f0f\u8f93\u51fa\u3002<\/li>\n<\/ul>\n

     <\/p>\n

    \u4f7f\u7528\u5e2e\u52a9<\/h2>\n

    \u5b89\u88c5\u6d41\u7a0b<\/h3>\n
      \n
    1. \u786e\u4fdd\u5df2\u5b89\u88c5 Node.js \u73af\u5883\u3002<\/li>\n
    2. \u4f7f\u7528 npm \u5b89\u88c5 Llama OCR \u5e93\uff1a<\/li>\n<\/ol>\n
         npm i llama-ocr\r\n<\/code><\/pre>\n
      \u4f7f\u7528\u65b9\u6cd5<\/h3>\n
        \n
      1. \u5bfc\u5165 Llama OCR \u5e93\uff1a<\/li>\n<\/ol>\n
           import { ocr } from \"llama-ocr\";\r\n<\/code><\/pre>\n
          \n
        1. \u8c03\u7528 ocr<\/code> \u51fd\u6570\u8fdb\u884c\u56fe\u50cf\u89e3\u6790\uff1a<\/li>\n<\/ol>\n
             const markdown = await ocr({\r\nfilePath: \".\/trader-joes-receipt.jpg\", \/\/ \u56fe\u50cf\u6587\u4ef6\u8def\u5f84\r\napiKey: process.env.TOGETHER_API_KEY, \/\/ Together AI API \u5bc6\u94a5\r\n});\r\n<\/code><\/pre>\n
            \n
          1. \u5904\u7406\u8fd4\u56de\u7684 Markdown \u6587\u672c\uff1a<\/li>\n<\/ol>\n
               console.log(markdown);\r\n<\/code><\/pre>\n
            \u8be6\u7ec6\u529f\u80fd\u64cd\u4f5c<\/h3>\n
              \n
            • \u56fe\u50cfOCR<\/strong>\uff1a\u5c06\u56fe\u50cf\u6587\u4ef6\u8def\u5f84\u4f20\u9012\u7ed9 ocr<\/code> \u51fd\u6570\uff0c\u5373\u53ef\u83b7\u53d6\u56fe\u50cf\u4e2d\u7684\u6587\u672c\u5185\u5bb9\u3002<\/li>\n
            • Markdown\u8f93\u51fa<\/strong>\uff1a\u89e3\u6790\u540e\u7684\u6587\u672c\u81ea\u52a8\u8f6c\u6362\u4e3a Markdown \u683c\u5f0f\uff0c\u65b9\u4fbf\u5728\u6587\u6863\u4e2d\u4f7f\u7528\u3002<\/li>\n
            • \u591a\u6a21\u578b\u652f\u6301<\/strong>\uff1a\u901a\u8fc7\u8bbe\u7f6e model<\/code> \u53c2\u6570\uff0c\u53ef\u4ee5\u9009\u62e9\u4e0d\u540c\u7684 Llama 3.2 \u6a21\u578b\uff08\u5982 Llama-3.2-90B-Vision<\/code> \u6216 Llama-3.2-11B-Vision<\/code>\uff09\u4ee5\u6ee1\u8db3\u4e0d\u540c\u7684\u6027\u80fd\u9700\u6c42\u3002<\/li>\n
            • API\u96c6\u6210<\/strong>\uff1a\u9700\u8981\u5728\u73af\u5883\u53d8\u91cf\u4e2d\u8bbe\u7f6e Together AI \u7684 API \u5bc6\u94a5\uff0c\u4ee5\u4fbf\u8c03\u7528\u5176\u63a5\u53e3\u8fdb\u884c\u56fe\u50cf\u89e3\u6790\u3002<\/li>\n<\/ul>\n
              \u793a\u4f8b\u4ee3\u7801<\/h3>\n
              import { ocr } from \"llama-ocr\";\r\nasync function runOCR() {\r\nconst markdown = await ocr({\r\nfilePath: \".\/example-image.jpg\",\r\napiKey: \"your-together-ai-api-key\",\r\n});\r\nconsole.log(markdown);\r\n}\r\nrunOCR();\r\n<\/code><\/pre>\n
              \u672a\u6765\u529f\u80fd<\/h3>\n
                \n
              • PDF\u652f\u6301<\/strong>\uff1a\u672a\u6765\u7248\u672c\u5c06\u652f\u6301\u5355\u9875\u548c\u591a\u9875 PDF \u6587\u4ef6\u7684 OCR \u5904\u7406\u3002<\/li>\n
              • JSON\u8f93\u51fa<\/strong>\uff1a\u9664\u4e86 Markdown \u683c\u5f0f\u5916\uff0c\u8fd8\u5c06\u652f\u6301 JSON \u683c\u5f0f\u7684\u8f93\u51fa\uff0c\u4fbf\u4e8e\u6570\u636e\u5904\u7406\u548c\u96c6\u6210\u3002<\/li>\n<\/ul>\n
                \u901a\u8fc7\u4ee5\u4e0a\u6b65\u9aa4\uff0c\u7528\u6237\u53ef\u4ee5\u8f7b\u677e\u5b89\u88c5\u548c\u4f7f\u7528 Llama OCR \u5e93\uff0c\u5c06\u56fe\u50cf\u4e2d\u7684\u6587\u672c\u5185\u5bb9\u8f6c\u6362\u4e3a Markdown \u683c\u5f0f\uff0c\u63d0\u5347\u6587\u6863\u5904\u7406\u6548\u7387\u3002<\/p>\n","protected":false},"excerpt":{"rendered":"
                Llama OCR \u662f\u4e00\u4e2a\u57fa\u4e8e Llama 3.2 Vision \u7684 OCR\uff08\u5149\u5b66\u5b57\u7b26\u8bc6\u522b\uff09\u5e93\uff0c\u80fd\u591f\u5c06\u6587\u6863\u8f6c\u6362\u4e3a Markdown \u683c\u5f0f\u3002\u8be5\u5e93\u7531 Nutlope \u5f00\u53d1\uff0c\u4f7f\u7528 Together AI \u63d0\u4f9b\u7684\u514d\u8d39 Llama 3.2 \u63a5\u53e3\u8fdb\u884c\u56fe…<\/p>\n","protected":false},"author":1,"featured_media":61396,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[20,502],"tags":[230,248],"class_list":["post-15235","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-tool","category-text-recognition","tag-aikaiyuanxiangmu","tag-ocr"],"_links":{"self":[{"href":"https:\/\/www.kdjingpai.com\/ja\/wp-json\/wp\/v2\/posts\/15235","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.kdjingpai.com\/ja\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.kdjingpai.com\/ja\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.kdjingpai.com\/ja\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.kdjingpai.com\/ja\/wp-json\/wp\/v2\/comments?post=15235"}],"version-history":[{"count":0,"href":"https:\/\/www.kdjingpai.com\/ja\/wp-json\/wp\/v2\/posts\/15235\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.kdjingpai.com\/ja\/wp-json\/wp\/v2\/media\/61396"}],"wp:attachment":[{"href":"https:\/\/www.kdjingpai.com\/ja\/wp-json\/wp\/v2\/media?parent=15235"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.kdjingpai.com\/ja\/wp-json\/wp\/v2\/categories?post=15235"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.kdjingpai.com\/ja\/wp-json\/wp\/v2\/tags?post=15235"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}