{"id":4920,"date":"2024-10-20T21:58:21","date_gmt":"2024-10-20T13:58:21","guid":{"rendered":"https:\/\/nullthought.net\/?p=4920"},"modified":"2025-02-23T11:08:37","modified_gmt":"2025-02-23T03:08:37","slug":"%e6%b5%85%e8%b0%88%e8%a7%86%e8%a7%89%e8%af%ad%e8%a8%80%e6%a8%a1%e5%9e%8b%ef%bc%88vision-language-models-vlms%ef%bc%89","status":"publish","type":"post","link":"https:\/\/nullthought.net\/?p=4920","title":{"rendered":"\u6d45\u8c08\u89c6\u89c9\u8bed\u8a00\u6a21\u578b\uff08Vision-Language Models, VLMs\uff09"},"content":{"rendered":"\n<p>\u89c6\u89c9\u8bed\u8a00\u6a21\u578b\uff08Vision-Language Models, VLMs\uff09\u662f\u540c\u65f6\u5904\u7406\u89c6\u89c9\u4fe1\u606f\u548c\u6587\u672c\u4fe1\u606f\u7684\u6df1\u5ea6\u5b66\u4e60\u6a21\u578b\uff0c\u65e8\u5728\u5b9e\u73b0\u8ba1\u7b97\u673a\u5bf9\u591a\u6a21\u6001\u6570\u636e\u7684\u7406\u89e3\u548c\u63a8\u7406\u80fd\u529b\u3002VLMs\u80fd\u591f\u5c06\u89c6\u89c9\u548c\u6587\u672c\u4fe1\u606f\u8fdb\u884c\u878d\u5408\uff0c\u4f7f\u6a21\u578b\u5728\u5904\u7406\u56fe\u50cf\u63cf\u8ff0\u3001\u56fe\u50cf\u95ee\u7b54\u3001\u8de8\u6a21\u6001\u68c0\u7d22\u7b49\u4efb\u52a1\u4e2d\u8868\u73b0\u5f97\u66f4\u52a0\u667a\u80fd\u5316\u548c\u9ad8\u6548\u3002<\/p>\n\n\n\n<h4 class=\"wp-block-heading\">1. \u89c6\u89c9\u8bed\u8a00\u6a21\u578b\u7684\u80cc\u666f\u4e0e\u53d1\u5c55<\/h4>\n\n\n\n<h5 class=\"wp-block-heading\">1.1 \u878d\u5408\u89c6\u89c9\u548c\u8bed\u8a00\u7684\u9700\u6c42<\/h5>\n\n\n\n<p>\u968f\u7740\u6df1\u5ea6\u5b66\u4e60\u6280\u672f\u7684\u53d1\u5c55\uff0c\u89c6\u89c9\u7406\u89e3\uff08\u4f8b\u5982\u56fe\u50cf\u5206\u7c7b\u3001\u7269\u4f53\u68c0\u6d4b\uff09\u548c\u81ea\u7136\u8bed\u8a00\u5904\u7406\uff08NLP\uff0c\u4f8b\u5982\u8bed\u8a00\u751f\u6210\u3001\u7ffb\u8bd1\uff09\u9886\u57df\u5404\u81ea\u53d6\u5f97\u4e86\u663e\u8457\u8fdb\u5c55\u3002\u7136\u800c\uff0c\u5f88\u591a\u73b0\u5b9e\u4e16\u754c\u4e2d\u7684\u4efb\u52a1\uff08\u5982\u63cf\u8ff0\u56fe\u7247\u3001\u89c6\u89c9\u95ee\u7b54\u7b49\uff09\u9700\u8981\u540c\u65f6\u5904\u7406\u89c6\u89c9\u548c\u8bed\u8a00\u4fe1\u606f\u3002\u4f20\u7edf\u7684\u89c6\u89c9\u6a21\u578b\u548c\u6587\u672c\u6a21\u578b\u53ea\u80fd\u5206\u522b\u5904\u7406\u89c6\u89c9\u6216\u8bed\u8a00\u6570\u636e\uff0c\u65e0\u6cd5\u5728\u590d\u6742\u7684\u591a\u6a21\u6001\u4efb\u52a1\u4e2d\u83b7\u5f97\u6700\u4f73\u6548\u679c\u3002\u56e0\u6b64\uff0c\u89c6\u89c9\u8bed\u8a00\u6a21\u578b\u5e94\u8fd0\u800c\u751f\uff0c\u76ee\u7684\u662f\u4f7f\u6a21\u578b\u53ef\u4ee5\u9ad8\u6548\u5730\u7406\u89e3\u548c\u5904\u7406\u591a\u6a21\u6001\u4fe1\u606f\uff0c\u4ece\u800c\u66f4\u597d\u5730\u5e94\u5bf9\u9700\u8981\u7ed3\u5408\u56fe\u50cf\u548c\u6587\u672c\u7684\u590d\u6742\u4efb\u52a1\u3002<\/p>\n\n\n\n<h5 class=\"wp-block-heading\">1.2 \u65e9\u671f\u89c6\u89c9\u8bed\u8a00\u6a21\u578b\u7684\u63a2\u7d22<\/h5>\n\n\n\n<p>\u65e9\u671f\u7684\u89c6\u89c9\u8bed\u8a00\u6a21\u578b\u901a\u8fc7\u7b80\u5355\u5730\u5c06\u89c6\u89c9\u7279\u5f81\u4e0e\u8bed\u8a00\u7279\u5f81\u62fc\u63a5\u6765\u8fdb\u884c\u591a\u6a21\u6001\u4efb\u52a1\u7684\u5904\u7406\u3002\u6b64\u7c7b\u6a21\u578b\u4f9d\u8d56\u4e8e\u4e24\u4e2a\u6a21\u5757\uff1a\u4e00\u4e2a\u89c6\u89c9\u7f16\u7801\u5668\uff08\u901a\u5e38\u662f\u5377\u79ef\u795e\u7ecf\u7f51\u7edc\uff0c\u5982ResNet\uff09\u7528\u4e8e\u63d0\u53d6\u56fe\u50cf\u7279\u5f81\uff0c\u548c\u4e00\u4e2a\u8bed\u8a00\u7f16\u7801\u5668\uff08\u5982RNN\u6216Transformer\uff09\u7528\u4e8e\u63d0\u53d6\u6587\u672c\u7279\u5f81\u3002\u968f\u540e\uff0c\u6a21\u578b\u5c06\u8fd9\u4e24\u79cd\u7279\u5f81\u7ed3\u5408\uff0c\u9001\u5165\u4e0b\u6e38\u4efb\u52a1\u7684\u5206\u7c7b\u5668\u6216\u751f\u6210\u5668\u3002<\/p>\n\n\n\n<p>\u968f\u7740\u89c6\u89c9-\u8bed\u8a00\u5bf9\u6bd4\u5b66\u4e60\uff08\u5982CLIP\uff09\u548c\u8de8\u6a21\u6001\u6ce8\u610f\u529b\u673a\u5236\uff08\u5982ViLT\uff09\u7684\u51fa\u73b0\uff0c\u89c6\u89c9\u8bed\u8a00\u6a21\u578b\u7684\u6027\u80fd\u5f97\u5230\u4e86\u663e\u8457\u63d0\u5347\u3002\u8fd9\u4e9b\u6a21\u578b\u53ef\u4ee5\u66f4\u597d\u5730\u6355\u6349\u5230\u89c6\u89c9\u548c\u8bed\u8a00\u4e4b\u95f4\u7684\u5173\u8054\uff0c\u5b9e\u73b0\u5bf9\u591a\u6a21\u6001\u6570\u636e\u7684\u7efc\u5408\u7406\u89e3\u3002<\/p>\n\n\n\n<h4 class=\"wp-block-heading\">2. \u89c6\u89c9\u8bed\u8a00\u6a21\u578b\u7684\u5de5\u4f5c\u539f\u7406<\/h4>\n\n\n\n<p>\u89c6\u89c9\u8bed\u8a00\u6a21\u578b\u7684\u5173\u952e\u5728\u4e8e\u5982\u4f55\u6709\u6548\u5730\u7f16\u7801\u56fe\u50cf\u548c\u6587\u672c\u5e76\u878d\u5408\u8fd9\u4e9b\u591a\u6a21\u6001\u7279\u5f81\u3002\u901a\u5e38\uff0cVLM\u7684\u5de5\u4f5c\u6d41\u7a0b\u53ef\u4ee5\u5206\u4e3a\u4e09\u4e2a\u9636\u6bb5\uff1a<strong>\u89c6\u89c9\u7279\u5f81\u63d0\u53d6<\/strong>\u3001<strong>\u8bed\u8a00\u7279\u5f81\u63d0\u53d6<\/strong>\u548c<strong>\u591a\u6a21\u6001\u878d\u5408<\/strong>\u3002<\/p>\n\n\n\n<h5 class=\"wp-block-heading\">2.1 \u89c6\u89c9\u7279\u5f81\u63d0\u53d6<\/h5>\n\n\n\n<p>\u89c6\u89c9\u7279\u5f81\u63d0\u53d6\u662f\u7531\u89c6\u89c9\u7f16\u7801\u5668\u5b8c\u6210\u7684\u3002\u73b0\u4ee3VLM\u4e2d\u5e38\u4f7f\u7528\u7c7b\u4f3c\u4e8eCLIP\u7684\u89c6\u89c9\u7f16\u7801\u5668\uff0c\u5176\u901a\u5e38\u57fa\u4e8e\u5377\u79ef\u795e\u7ecf\u7f51\u7edc\uff08CNN\uff09\u6216\u89c6\u89c9Transformer\uff08\u5982ViT\uff09\u6765\u63d0\u53d6\u56fe\u50cf\u7279\u5f81\u3002\u89c6\u89c9\u7f16\u7801\u5668\u7684\u8f93\u51fa\u662f\u56fe\u50cf\u7684\u7279\u5f81\u5411\u91cf\uff0c\u5b83\u80fd\u591f\u8868\u5f81\u56fe\u50cf\u4e2d\u7684\u4e0d\u540c\u533a\u57df\u6216\u7269\u4f53\uff0c\u4ee5\u53ca\u5b83\u4eec\u7684\u7a7a\u95f4\u5173\u7cfb\u3002<\/p>\n\n\n\n<h5 class=\"wp-block-heading\">2.2 \u8bed\u8a00\u7279\u5f81\u63d0\u53d6<\/h5>\n\n\n\n<p>\u8bed\u8a00\u7279\u5f81\u63d0\u53d6\u7531\u8bed\u8a00\u7f16\u7801\u5668\u5b8c\u6210\uff0c\u901a\u5e38\u662f\u57fa\u4e8eTransformer\u67b6\u6784\uff08\u5982BERT\u6216GPT\uff09\u3002\u6587\u672c\u6570\u636e\u88ab\u5904\u7406\u6210token\uff0c\u7ecf\u8fc7\u7f16\u7801\u5668\u540e\u5f97\u5230\u6587\u672c\u7684\u4e0a\u4e0b\u6587\u5d4c\u5165\u3002\u8fd9\u4e9b\u5d4c\u5165\u4ee3\u8868\u4e86\u6587\u672c\u7684\u8bed\u4e49\u4fe1\u606f\u4ee5\u53ca\u8bcd\u8bed\u4e4b\u95f4\u7684\u8054\u7cfb\u3002<\/p>\n\n\n\n<h5 class=\"wp-block-heading\">2.3 \u591a\u6a21\u6001\u878d\u5408<\/h5>\n\n\n\n<p>\u591a\u6a21\u6001\u878d\u5408\u662fVLM\u7684\u6838\u5fc3\u90e8\u5206\uff0c\u5b83\u7528\u4e8e\u5c06\u89c6\u89c9\u548c\u8bed\u8a00\u7279\u5f81\u8fdb\u884c\u6574\u5408\uff0c\u4ee5\u4fbf\u6a21\u578b\u80fd\u591f\u7406\u89e3\u4e24\u8005\u4e4b\u95f4\u7684\u5173\u8054\u3002\u5178\u578b\u7684\u878d\u5408\u65b9\u6cd5\u5305\u62ec\u4ee5\u4e0b\u51e0\u79cd\uff1a<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>\u5bf9\u6bd4\u5b66\u4e60<\/strong>\uff1a\u5982CLIP\uff08Contrastive Language-Image Pre-training\uff09\uff0c\u901a\u8fc7\u5bf9\u6bd4\u5b66\u4e60\u8ba9\u6a21\u578b\u5b66\u4e60\u5c06\u63cf\u8ff0\u76f8\u4f3c\u7684\u56fe\u50cf\u548c\u6587\u672c\u6620\u5c04\u5230\u76f8\u8fd1\u7684\u7279\u5f81\u7a7a\u95f4\u4e2d\u3002<\/li>\n\n\n\n<li><strong>\u8de8\u6a21\u6001\u6ce8\u610f\u529b<\/strong>\uff1a\u5982LXMERT\u3001UNITER\u7b49\u6a21\u578b\uff0c\u901a\u8fc7\u8de8\u6a21\u6001\u6ce8\u610f\u529b\u673a\u5236\u6355\u6349\u89c6\u89c9\u548c\u8bed\u8a00\u4e4b\u95f4\u7684\u4ea4\u4e92\u5173\u7cfb\uff0c\u4ece\u800c\u8ba9\u6a21\u578b\u80fd\u591f\u5728\u89c6\u89c9\u548c\u8bed\u8a00\u4e4b\u95f4\u5efa\u7acb\u66f4\u5f3a\u7684\u5173\u8054\u3002<\/li>\n\n\n\n<li><strong>\u8054\u5408\u7f16\u7801\u5668<\/strong>\uff1a\u6709\u4e9b\u6a21\u578b\u76f4\u63a5\u5c06\u56fe\u50cf\u548c\u6587\u672c\u4f5c\u4e3a\u8f93\u5165\uff0c\u7ecf\u8fc7\u7edf\u4e00\u7684\u7f16\u7801\u5668\u8fdb\u884c\u8054\u5408\u7f16\u7801\uff08\u5982ViLT\uff09\uff0c\u8fd9\u6837\u53ef\u4ee5\u8ba9\u6a21\u578b\u5728\u7f16\u7801\u8fc7\u7a0b\u4e2d\u66f4\u597d\u5730\u5904\u7406\u591a\u6a21\u6001\u7684\u4ea4\u4e92\u4fe1\u606f\u3002<\/li>\n<\/ul>\n\n\n\n<h4 class=\"wp-block-heading\">3. \u5178\u578b\u7684\u89c6\u89c9\u8bed\u8a00\u6a21\u578b<\/h4>\n\n\n\n<p>\u4ee5\u4e0b\u662f\u4e00\u4e9b\u5178\u578b\u7684VLM\uff0c\u5b83\u4eec\u4ee3\u8868\u4e86\u89c6\u89c9\u8bed\u8a00\u6a21\u578b\u7684\u53d1\u5c55\u5386\u7a0b\u53ca\u5176\u5728\u591a\u6a21\u6001\u7406\u89e3\u4e2d\u7684\u521b\u65b0\uff1a<\/p>\n\n\n\n<h5 class=\"wp-block-heading\">3.1 CLIP<\/h5>\n\n\n\n<p><strong>CLIP<\/strong>\uff08Contrastive Language-Image Pre-training\uff09\u7531OpenAI\u63d0\u51fa\uff0c\u662f\u4e00\u79cd\u901a\u8fc7\u5bf9\u6bd4\u5b66\u4e60\u6765\u9884\u8bad\u7ec3\u89c6\u89c9\u548c\u6587\u672c\u6a21\u578b\u7684\u65b9\u6cd5\u3002CLIP\u7684\u76ee\u6807\u662f\u5c06\u6587\u672c\u548c\u56fe\u50cf\u6620\u5c04\u5230\u540c\u4e00\u4e2a\u7279\u5f81\u7a7a\u95f4\u4e2d\uff0c\u4f7f\u5f97\u76f8\u5e94\u7684\u6587\u672c\u548c\u56fe\u50cf\u5177\u6709\u76f8\u4f3c\u7684\u7279\u5f81\u5411\u91cf\uff0c\u4ece\u800c\u5b9e\u73b0\u89c6\u89c9-\u8bed\u8a00\u68c0\u7d22\u7684\u80fd\u529b\u3002CLIP\u901a\u8fc7\u5bf9\u6210\u5bf9\u7684\u56fe\u50cf\u548c\u6587\u672c\u8fdb\u884c\u5927\u89c4\u6a21\u9884\u8bad\u7ec3\uff0c\u5b66\u4e60\u5230\u4e86\u5982\u4f55\u5728\u540c\u4e00\u4e2a\u7279\u5f81\u7a7a\u95f4\u4e2d\u5bf9\u9f50\u89c6\u89c9\u548c\u8bed\u8a00\u7279\u5f81\u3002<\/p>\n\n\n\n<h5 class=\"wp-block-heading\">3.2 Flamingo<\/h5>\n\n\n\n<p><strong>Flamingo<\/strong>\u662f\u7531DeepMind\u63d0\u51fa\u7684\u4e00\u79cd\u89c6\u89c9\u8bed\u8a00\u6a21\u578b\uff0c\u4e3b\u8981\u7528\u4e8e\u89e3\u51b3\u5c11\u6837\u672c\u5b66\u4e60\u4e2d\u7684\u89c6\u89c9\u8bed\u8a00\u7406\u89e3\u95ee\u9898\u3002Flamingo\u5c06\u8bed\u8a00\u6a21\u578b\uff08\u5982Transformer\uff09\u4e0e\u89c6\u89c9\u7f16\u7801\u5668\u76f8\u7ed3\u5408\uff0c\u53ef\u4ee5\u5728\u6709\u9650\u7684\u6570\u636e\u4e0b\u8fdb\u884c\u591a\u6a21\u6001\u4efb\u52a1\u7684\u9ad8\u6548\u5b66\u4e60\u3002Flamingo\u91c7\u7528\u4e86\u4e00\u79cd\u57fa\u4e8e\u6ce8\u610f\u529b\u7684\u878d\u5408\u65b9\u6cd5\u6765\u7ed3\u5408\u89c6\u89c9\u548c\u6587\u672c\u7279\u5f81\uff0c\u4ee5\u5b9e\u73b0\u5bf9\u590d\u6742\u591a\u6a21\u6001\u8f93\u5165\u7684\u7406\u89e3\u3002<\/p>\n\n\n\n<p>\u533b\u7597\u65b9\u9762\u5e94\u7528\uff1a<a href=\"https:\/\/nullthought.net\/?p=5181\" target=\"_blank\" rel=\"noreferrer noopener\">Flamingo-CXR\uff0c\u7ed3\u5408\u89c6\u89c9\u548c\u8bed\u8a00\u8868\u793a\uff0c\u53ef\u81ea\u52a8\u5316\u751f\u6210\u653e\u5c04\u62a5\u544a\u7684\u5148\u8fdbAI\u6a21\u578b<\/a>\u3002<\/p>\n\n\n\n<h5 class=\"wp-block-heading\">3.3 LLaVA<\/h5>\n\n\n\n<p><strong>LLaVA<\/strong>\uff08Large Language and Vision Assistant\uff09\u662f\u4e00\u79cd\u7ed3\u5408\u89c6\u89c9\u548c\u8bed\u8a00\u7684\u5927\u578b\u6a21\u578b\uff0c\u80fd\u591f\u6267\u884c\u89c6\u89c9\u95ee\u7b54\u3001\u56fe\u50cf\u63cf\u8ff0\u7b49\u4efb\u52a1\u3002LLaVA\u901a\u8fc7\u4e0e\u89c6\u89c9\u7f16\u7801\u5668\uff08\u5982CLIP\uff09\u7684\u96c6\u6210\uff0c\u5c06\u56fe\u50cf\u7684\u7279\u5f81\u4fe1\u606f\u4e0e\u8bed\u8a00\u6a21\u578b\u8fdb\u884c\u6709\u6548\u878d\u5408\uff0c\u5b9e\u73b0\u4e86\u8f83\u5f3a\u7684\u591a\u6a21\u6001\u63a8\u7406\u80fd\u529b\u3002<\/p>\n\n\n\n<h5 class=\"wp-block-heading\">3.4 <a href=\"https:\/\/nullthought.net\/?p=4084\" target=\"_blank\" rel=\"noreferrer noopener\">PaliGemma<\/a><\/h5>\n\n\n\n<p>PaliGemma\u662f\u6765\u81eaGoogle Deepmind\u56e2\u961f\u7684\u7814\u7a76\u6210\u679c\u3002<strong>PaliGemma<\/strong>\u662f\u4e00\u4e2a\u5f00\u653e\u7684\u89c6\u89c9\u8bed\u8a00\u6a21\u578b\uff08VLM\uff09\uff0c\u5b83\u7ed3\u5408\u4e86 SigLIP-So400m \u89c6\u89c9\u7f16\u7801\u5668\u548c Gemma-2B \u8bed\u8a00\u6a21\u578b\u3002\u5176\u76ee\u6807\u662f\u63d0\u4f9b\u4e00\u4e2a\u591a\u529f\u80fd\u7684\u57fa\u7840\u6a21\u578b\uff0c\u80fd\u591f\u6709\u6548\u8fc1\u79fb\u5230\u5404\u79cd\u4efb\u52a1\uff0c\u5728\u8fd140\u4e2a\u4e0d\u540c\u7684\u4efb\u52a1\u4e2d\u8868\u73b0\u51fa\u8272\uff0c\u5305\u62ec\u6807\u51c6\u7684 VLM \u57fa\u51c6\u6d4b\u8bd5\u4ee5\u53ca\u66f4\u4e13\u4e1a\u7684\u4efb\u52a1\u5982\u9065\u611f\u548c\u5206\u5272\u3002<\/p>\n\n\n\n<h4 class=\"wp-block-heading\">4. \u89c6\u89c9\u8bed\u8a00\u6a21\u578b\u7684\u5e94\u7528\u573a\u666f<\/h4>\n\n\n\n<p>\u89c6\u89c9\u8bed\u8a00\u6a21\u578b\u5728\u591a\u6a21\u6001\u4efb\u52a1\u4e2d\u8868\u73b0\u51fa\u8272\uff0c\u4ee5\u4e0b\u662f\u5176\u4e00\u4e9b\u5178\u578b\u7684\u5e94\u7528\u573a\u666f\uff1a<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>\u56fe\u50cf\u63cf\u8ff0<\/strong>\uff1aVLM\u80fd\u591f\u6839\u636e\u56fe\u50cf\u751f\u6210\u81ea\u7136\u8bed\u8a00\u63cf\u8ff0\uff0c\u4f8b\u5982\u201c\u56fe\u7247\u4e2d\u6709\u4e00\u53ea\u5728\u8dd1\u6b65\u7684\u72d7\u201d\u3002<\/li>\n\n\n\n<li><strong>\u89c6\u89c9\u95ee\u7b54\uff08VQA\uff09<\/strong>\uff1aVLM\u53ef\u4ee5\u56de\u7b54\u5173\u4e8e\u56fe\u7247\u5185\u5bb9\u7684\u95ee\u9898\uff0c\u4f8b\u5982\u201c\u56fe\u7247\u4e2d\u6709\u591a\u5c11\u4eba\uff1f\u201d\u3002<\/li>\n\n\n\n<li><strong>\u8de8\u6a21\u6001\u68c0\u7d22<\/strong>\uff1aVLM\u53ef\u4ee5\u5b9e\u73b0\u8de8\u6a21\u6001\u68c0\u7d22\uff0c\u4f8b\u5982\u901a\u8fc7\u8f93\u5165\u6587\u672c\u6765\u68c0\u7d22\u4e0e\u4e4b\u5339\u914d\u7684\u56fe\u50cf\uff0c\u6216\u901a\u8fc7\u8f93\u5165\u56fe\u50cf\u6765\u68c0\u7d22\u4e0e\u4e4b\u63cf\u8ff0\u76f8\u7b26\u7684\u6587\u672c\u3002<\/li>\n\n\n\n<li><strong>\u591a\u6a21\u6001\u63a8\u7406<\/strong>\uff1a\u4f8b\u5982\u5728\u533b\u5b66\u5f71\u50cf\u4e2d\uff0cVLM\u53ef\u4ee5\u7ed3\u5408\u5f71\u50cf\u548c\u533b\u751f\u7684\u6ce8\u91ca\u8fdb\u884c\u8bca\u65ad\u5206\u6790\u3002<\/li>\n<\/ul>\n\n\n\n<h4 class=\"wp-block-heading\">5. \u89c6\u89c9\u8bed\u8a00\u6a21\u578b\u7684\u6311\u6218\u548c\u672a\u6765\u53d1\u5c55<\/h4>\n\n\n\n<p>\u5c3d\u7ba1\u89c6\u89c9\u8bed\u8a00\u6a21\u578b\u5728\u591a\u6a21\u6001\u9886\u57df\u53d6\u5f97\u4e86\u663e\u8457\u7684\u8fdb\u5c55\uff0c\u4f46\u5b83\u4eec\u4ecd\u9762\u4e34\u4e00\u4e9b\u6311\u6218\uff1a<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>\u591a\u6a21\u6001\u6570\u636e\u7684\u6807\u6ce8<\/strong>\uff1a\u8bad\u7ec3VLM\u9700\u8981\u5927\u91cf\u7684\u56fe\u50cf-\u6587\u672c\u914d\u5bf9\u6570\u636e\uff0c\u8fd9\u79cd\u6570\u636e\u7684\u83b7\u53d6\u548c\u6807\u6ce8\u6210\u672c\u8f83\u9ad8\u3002<\/li>\n\n\n\n<li><strong>\u8de8\u6a21\u6001\u63a8\u7406\u7684\u590d\u6742\u6027<\/strong>\uff1a\u5728\u9700\u8981\u591a\u4e2a\u4fe1\u606f\u6765\u6e90\u8fdb\u884c\u63a8\u7406\u7684\u590d\u6742\u573a\u666f\u4e2d\uff0c\u5982\u4f55\u6709\u6548\u878d\u5408\u591a\u5e45\u56fe\u50cf\u548c\u6587\u672c\u4ecd\u7136\u662f\u4e00\u4e2a\u96be\u70b9\u3002<\/li>\n\n\n\n<li><strong>\u9ad8\u6548\u7684\u6a21\u578b\u7ed3\u6784<\/strong>\uff1aVLM\u901a\u5e38\u9700\u8981\u540c\u65f6\u5904\u7406\u56fe\u50cf\u548c\u6587\u672c\uff0c\u5176\u8ba1\u7b97\u91cf\u548c\u5185\u5b58\u9700\u6c42\u5f88\u9ad8\uff0c\u56e0\u6b64\u5982\u4f55\u8bbe\u8ba1\u66f4\u52a0\u9ad8\u6548\u7684\u7ed3\u6784\u4ee5\u5e94\u5bf9\u5927\u89c4\u6a21\u63a8\u7406\u9700\u6c42\u662f\u4e00\u4e2a\u91cd\u8981\u7814\u7a76\u65b9\u5411\u3002<\/li>\n<\/ul>\n","protected":false},"excerpt":{"rendered":"<p>\u89c6\u89c9\u8bed\u8a00\u6a21\u578b\uff08Vision-Language Models, VLMs\uff09\u662f\u540c\u65f6\u5904\u7406\u89c6\u89c9\u4fe1\u606f\u548c\u6587\u672c\u4fe1\u606f\u7684\u6df1\u5ea6\u5b66\u4e60 [&hellip;]<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"site-sidebar-layout":"default","site-content-layout":"","ast-site-content-layout":"default","site-content-style":"default","site-sidebar-style":"default","ast-global-header-display":"","ast-banner-title-visibility":"","ast-main-header-display":"","ast-hfb-above-header-display":"","ast-hfb-below-header-display":"","ast-hfb-mobile-header-display":"","site-post-title":"","ast-breadcrumbs-content":"","ast-featured-img":"","footer-sml-layout":"","ast-disable-related-posts":"","theme-transparent-header-meta":"","adv-header-id-meta":"","stick-header-meta":"","header-above-stick-meta":"","header-main-stick-meta":"","header-below-stick-meta":"","astra-migrate-meta-layouts":"set","ast-page-background-enabled":"default","ast-page-background-meta":{"desktop":{"background-color":"","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"tablet":{"background-color":"","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"mobile":{"background-color":"","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""}},"ast-content-background-meta":{"desktop":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"tablet":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"mobile":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""}},"footnotes":""},"categories":[8],"tags":[39,95,90,80],"class_list":["post-4920","post","type-post","status-publish","format-standard","hentry","category-tech","tag-ai","tag-transformer","tag-multimodal","tag-cv"],"rttpg_featured_image_url":null,"rttpg_author":{"display_name":"NullThought","author_link":"https:\/\/nullthought.net\/?author=1"},"rttpg_comment":0,"rttpg_category":"<a href=\"https:\/\/nullthought.net\/?cat=8\" rel=\"category\">Tech<\/a>","rttpg_excerpt":"\u89c6\u89c9\u8bed\u8a00\u6a21\u578b\uff08Vision-Language Models, VLMs\uff09\u662f\u540c\u65f6\u5904\u7406\u89c6\u89c9\u4fe1\u606f\u548c\u6587\u672c\u4fe1\u606f\u7684\u6df1\u5ea6\u5b66\u4e60&hellip;","_links":{"self":[{"href":"https:\/\/nullthought.net\/index.php?rest_route=\/wp\/v2\/posts\/4920","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/nullthought.net\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/nullthought.net\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/nullthought.net\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/nullthought.net\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=4920"}],"version-history":[{"count":2,"href":"https:\/\/nullthought.net\/index.php?rest_route=\/wp\/v2\/posts\/4920\/revisions"}],"predecessor-version":[{"id":5185,"href":"https:\/\/nullthought.net\/index.php?rest_route=\/wp\/v2\/posts\/4920\/revisions\/5185"}],"wp:attachment":[{"href":"https:\/\/nullthought.net\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=4920"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/nullthought.net\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=4920"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/nullthought.net\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=4920"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}