{"id":5977,"date":"2025-06-17T09:30:43","date_gmt":"2025-06-17T01:30:43","guid":{"rendered":"https:\/\/nullthought.net\/?p=5977"},"modified":"2025-10-07T15:45:18","modified_gmt":"2025-10-07T07:45:18","slug":"textok%ef%bc%9a%e5%9f%ba%e4%ba%8e%e6%96%87%e6%9c%ac%e6%9d%a1%e4%bb%b6%e5%bc%95%e5%af%bc%e7%9a%84%e5%9b%be%e5%83%8f%e5%88%86%e8%af%8d%e6%a1%86%e6%9e%b6%e3%80%82%e5%9c%a8%e7%bc%96%e7%a0%81%e5%99%a8","status":"publish","type":"post","link":"https:\/\/nullthought.net\/?p=5977","title":{"rendered":"TexTok\uff1a\u57fa\u4e8e\u6587\u672c\u6761\u4ef6\u5f15\u5bfc\u7684\u56fe\u50cf\u5206\u8bcd\u6846\u67b6\u3002\u5728\u7f16\u7801\u5668\uff08Tokenizer\uff09\u4e0e\u89e3\u7801\u5668\uff08Detokenizer\uff09\u4e2d\u5747\u5f15\u5165\u56fe\u50cf\u5bf9\u5e94\u7684\u6587\u5b57\u63cf\u8ff0\u5d4c\u5165\uff0c\u4f5c\u4e3a\u8bed\u4e49\u6761\u4ef6\uff0c\u6307\u5bfc\u56fe\u50cf\u7684\u538b\u7f29\u4e0e\u91cd\u5efa"},"content":{"rendered":"\n<p class=\"wp-block-paragraph\">\u8bba\u6587<strong><a href=\"https:\/\/arxiv.org\/abs\/2412.05796\" target=\"_blank\" rel=\"noreferrer noopener\">Language-Guided Image Tokenization for Generation<\/a><\/strong>\u63d0\u51fa\u7684TexTok\u4e3a\u4e00\u79cd\u6709\u6548\u4e14\u9ad8\u6548\u7684\u56fe\u50cf\u5206\u8bcd\u65b0\u8303\u5f0f\uff0c\u5176\u901a\u8fc7\u6587\u672c\u8bed\u4e49\u5728\u5206\u8bcd\u9636\u6bb5\u7684\u5f15\u5165\uff0c\u663e\u8457\u63d0\u5347\u4e86\u56fe\u50cf\u91cd\u5efa\u8d28\u91cf\u548c\u751f\u6210\u6027\u80fd\uff0c\u5e76\u5728\u4e0d\u540ctoken\u538b\u7f29\u7387\u4e0b\u5747\u4fdd\u6301\u9886\u5148\u3002\u5728ImageNet\u6761\u4ef6\u751f\u6210\u4e0e\u6587\u672c\u751f\u6210\u4efb\u52a1\u4e2d\u5747\u8fbe\u6210\u4e86SOTA\u6c34\u51c6\uff0c\u5e76\u5927\u5e45\u51cf\u5c11\u63a8\u7406\u65f6\u95f4\u548c\u8ba1\u7b97\u6210\u672c\u3002TexTok\u5c55\u73b0\u51fa\u5c06\u8bed\u8a00\u4fe1\u606f\u7528\u4e8e\u89c6\u89c9\u4efb\u52a1\u65e9\u671f\u7f16\u7801\u9636\u6bb5\u7684\u6f5c\u529b\uff0c\u672a\u6765\u53ef\u62d3\u5c55\u81f3\u66f4\u591a\u591a\u6a21\u6001\u751f\u6210\u4e0e\u7406\u89e3\u4efb\u52a1\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u8bba\u6587\u4f5c\u8005\u4e3aKaiwen Zha, Lijun Yu, Alireza Fathi, David A. Ross, Cordelia Schmid, Dina Katabi, Xiuye Gu\uff0c\u6765\u81eaGoogle DeepMind\u548cMIT CSAIL\u3002<\/p>\n\n\n\n<figure class=\"wp-block-image aligncenter size-full\"><img loading=\"lazy\" decoding=\"async\" width=\"470\" height=\"722\" src=\"https:\/\/nullthought.net\/wp-content\/uploads\/2025\/06\/image.png\" alt=\"\" class=\"wp-image-5978\" srcset=\"https:\/\/nullthought.net\/wp-content\/uploads\/2025\/06\/image.png 470w, https:\/\/nullthought.net\/wp-content\/uploads\/2025\/06\/image-195x300.png 195w\" sizes=\"auto, (max-width: 470px) 100vw, 470px\" \/><figcaption class=\"wp-element-caption\"><strong><a href=\"https:\/\/arxiv.org\/abs\/2412.05796\" target=\"_blank\" rel=\"noreferrer noopener\">Language-Guided Image Tokenization for Generation<\/a><\/strong><\/figcaption><\/figure>\n\n\n\n<p class=\"wp-block-paragraph\">\u4e00\u3001\u7814\u7a76\u80cc\u666f\u4e0e\u52a8\u673a<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u8fd1\u5e74\u6765\u56fe\u50cf\u751f\u6210\u9886\u57df\u53d6\u5f97\u4e86\u663e\u8457\u8fdb\u5c55\uff0c\u5176\u4e2d\u4e00\u4e2a\u5173\u952e\u56e0\u7d20\u662f\u56fe\u50cf\u5206\u8bcd\uff08tokenization\uff09\u6280\u672f\u7684\u53d1\u5c55\uff0c\u5373\u5c06\u539f\u59cb\u56fe\u50cf\u538b\u7f29\u4e3a\u4f4e\u7ef4\u6f5c\u5728\u8868\u793a\u3002\u8fd9\u4e00\u8fc7\u7a0b\u901a\u5e38\u501f\u52a9\u81ea\u52a8\u7f16\u7801\u5668\uff08AutoEncoder\uff09\u5b8c\u6210\uff0c\u4f7f\u751f\u6210\u6a21\u578b\uff08\u5982\u6269\u6563\u6a21\u578b\u548c\u81ea\u56de\u5f52\u6a21\u578b\uff09\u53ef\u4ee5\u5728\u6f5c\u5728\u7a7a\u95f4\u4e2d\u751f\u6210\u56fe\u50cf\uff0c\u4ece\u800c\u5927\u5e45\u964d\u4f4e\u8ba1\u7b97\u5f00\u9500\u3002\u7136\u800c\uff0c\u73b0\u6709\u56fe\u50cf\u5206\u8bcd\u65b9\u6cd5\u5728\u9ad8\u538b\u7f29\u7387\u4e0b\u5f80\u5f80\u727a\u7272\u91cd\u5efa\u8d28\u91cf\uff0c\u5bfc\u81f4\u9ad8\u5206\u8fa8\u7387\u56fe\u50cf\u751f\u6210\u7684\u8868\u73b0\u53d7\u9650\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u672c\u6587\u63d0\u51fa\u7684\u6838\u5fc3\u89c2\u70b9\u662f\uff1a\u8bed\u8a00\u63cf\u8ff0\u5929\u7136\u5177\u5907\u56fe\u50cf\u8bed\u4e49\u8868\u8fbe\u80fd\u529b\uff0c\u82e5\u80fd\u5728\u5206\u8bcd\u9636\u6bb5\u5f15\u5165\u56fe\u50cf\u5bf9\u5e94\u7684\u6587\u5b57\u63cf\u8ff0\uff0c\u6216\u53ef\u63d0\u5347\u8bed\u4e49\u5b66\u4e60\u6548\u7387\uff0c\u4f7f\u5f97\u66f4\u591atoken\u7a7a\u95f4\u7528\u4e8e\u7cbe\u7ec6\u89c6\u89c9\u7ec6\u8282\u7684\u7f16\u7801\uff0c\u4ece\u800c\u517c\u987e\u538b\u7f29\u7387\u4e0e\u91cd\u5efa\u8d28\u91cf\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u4e8c\u3001\u65b9\u6cd5\u6982\u8ff0\uff1aTexTok \u6846\u67b6<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">TexTok\uff08Text-Conditioned Image Tokenization\uff09\u662f\u672c\u6587\u63d0\u51fa\u7684\u57fa\u4e8e\u6587\u672c\u6761\u4ef6\u5f15\u5bfc\u7684\u56fe\u50cf\u5206\u8bcd\u6846\u67b6\uff0c\u5176\u6838\u5fc3\u521b\u65b0\u5728\u4e8e\u5728\u7f16\u7801\u5668\uff08Tokenizer\uff09\u4e0e\u89e3\u7801\u5668\uff08Detokenizer\uff09\u4e2d\u5747\u5f15\u5165\u56fe\u50cf\u5bf9\u5e94\u7684\u6587\u5b57\u63cf\u8ff0\u5d4c\u5165\uff0c\u4f5c\u4e3a\u8bed\u4e49\u6761\u4ef6\uff0c\u6307\u5bfc\u56fe\u50cf\u7684\u538b\u7f29\u4e0e\u91cd\u5efa\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u5176\u7ed3\u6784\u5305\u62ec\uff1a<\/p>\n\n\n\n<ol class=\"wp-block-list\">\n<li>\u6587\u672c\u7f16\u7801\u5668\u4f7f\u7528\u51bb\u7ed3\u7684 T5 \u6a21\u578b\uff0c\u5c06\u56fe\u50cf\u63cf\u8ff0\u8f6c\u5316\u4e3a\u5d4c\u5165\u5411\u91cf\u3002<\/li>\n\n\n\n<li>\u5206\u8bcd\u5668\uff08Tokenizer\uff09\u8f93\u5165\u7531\u4e09\u4e2a\u90e8\u5206\u7ec4\u6210\uff1a\u56fe\u50cf patch token\u3001\u53ef\u5b66\u4e60\u7684\u56fe\u50cf token \u4ee5\u53ca\u6587\u672c token\u3002\u8f93\u51fa\u4ec5\u4fdd\u7559\u56fe\u50cf token\u3002<\/li>\n\n\n\n<li>\u89e3\u7801\u5668\u8f93\u5165\u540c\u6837\u5305\u62ec\u56fe\u50cf token\u3001\u6587\u672c token \u4ee5\u53ca\u53ef\u5b66\u4e60\u7684 patch token\uff0c\u6700\u7ec8\u6062\u590d\u51fa\u539f\u56fe\u3002<\/li>\n\n\n\n<li>\u6a21\u578b\u8bad\u7ec3\u91c7\u7528 L2 \u91cd\u5efa\u635f\u5931\u3001\u611f\u77e5\u635f\u5931\uff08LPIPS\uff09\u3001GAN \u635f\u5931\u53ca LeCAM \u6b63\u5219\u9879\u3002<\/li>\n<\/ol>\n\n\n\n<p class=\"wp-block-paragraph\">TexTok \u4e0e\u4f20\u7edf\u65b9\u6cd5\u7684\u4e0d\u540c\u5728\u4e8e\uff0c\u5b83\u5c06\u56fe\u50cf\u7684\u9ad8\u5c42\u8bed\u4e49\u4ea4\u7531\u6587\u672c token \u8868\u8fbe\uff0c\u4f7f\u6a21\u578b\u66f4\u591a\u5173\u6ce8\u7ec6\u8282\u7eb9\u7406\uff0c\u5c24\u5176\u5728 token \u6570\u91cf\u53d7\u9650\u7684\u573a\u666f\u4e0b\u4f18\u52bf\u66f4\u660e\u663e\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u4e09\u3001\u56fe\u50cf\u751f\u6210\u6d41\u7a0b<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u5728\u56fe\u50cf\u751f\u6210\u9636\u6bb5\uff0cTexTok \u4ec5\u9700\u751f\u6210\u56fe\u50cf token\uff0c\u800c\u6587\u672c token \u53ef\u7531\u9884\u8bbecaption\u76f4\u63a5\u63d0\u4f9b\u3002\u751f\u6210\u6a21\u578b\u9009\u7528 <a href=\"https:\/\/nullthought.net\/?p=6239\" target=\"_blank\" rel=\"noreferrer noopener\">Diffusion Transformer\uff08DiT\uff09<\/a>\uff0c\u5728 class-conditional \u4efb\u52a1\u4e2d\u4f7f\u7528\u7c7b\u522b\u6807\u7b7e\u4f5c\u4e3a\u6761\u4ef6\uff0c\u5728 text-to-image \u4efb\u52a1\u4e2d\u4f7f\u7528\u6587\u672c\u63cf\u8ff0\u4f5c\u4e3a\u6761\u4ef6\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u5728\u63a8\u7406\u9636\u6bb5\uff0cTexTok \u5177\u6709\u4ee5\u4e0b\u7279\u70b9\uff1a<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>\u7c7b\u522b\u751f\u6210\u4efb\u52a1\u4e2d\uff0c\u7528\u79bb\u7ebf\u751f\u6210\u7684\u7c7b\u63cf\u8ff0caption\u4e0e\u751f\u6210\u7684\u56fe\u50cftoken\u5171\u540c\u9001\u5165\u89e3\u7801\u5668\u3002<\/li>\n\n\n\n<li>\u6587\u672c\u751f\u6210\u4efb\u52a1\u4e2d\uff0ccaption\u540c\u6837\u9001\u5165\u751f\u6210\u5668\u4e0e\u89e3\u7801\u5668\uff0c\u751f\u6210\u8fc7\u7a0b\u65e0\u989d\u5916\u6ce8\u91ca\u4ee3\u4ef7\u3002<\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\">\u56db\u3001\u5b9e\u9a8c\u8bbe\u7f6e\u4e0e\u5b9e\u73b0\u7ec6\u8282<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">TexTok \u7684\u8bad\u7ec3\u4e3b\u8981\u57fa\u4e8e ImageNet-256 \u4e0e ImageNet-512 \u6570\u636e\u96c6\uff0c\u5e76\u4f7f\u7528 Gemini VLM \u81ea\u52a8\u4e3a\u56fe\u50cf\u751f\u6210\u6587\u5b57\u63cf\u8ff0\uff08caption\uff09\u3002\u6587\u672c\u957f\u5ea6\u8bbe\u7f6e\u4e3a\u6700\u591a75\u8bcd\uff0c\u901a\u8fc7T5-XL\u6216T5-XXL\u7f16\u7801\u4e3a128\u7ef4token\u5e8f\u5217\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u7f51\u7edc\u7ed3\u6784\u91c7\u7528ViT-Base\uff0c\u5177\u670912\u5c42Transformer\u7ed3\u6784\u3001768\u7ef4\u9690\u85cf\u5c42\u300112\u5934\u6ce8\u610f\u529b\uff0ctoken\u7ef4\u5ea6\u8bbe\u4e3a8\u3002\u56fe\u50cftoken\u6570\u8bbe\u7f6e\u4e3a32\/64\/128\/256\uff0c\u5206\u522b\u7528\u4e8e\u8bc4\u4f30\u538b\u7f29\u7387\u4e0e\u751f\u6210\u8d28\u91cf\u7684\u6743\u8861\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u751f\u6210\u5668\u4e3a\u539f\u59cbDiT\u7ed3\u6784\uff0c\u6587\u672c\u751f\u6210\u4efb\u52a1\u4e2d\u6dfb\u52a0\u591a\u5934cross-attention\u7528\u4e8e\u63a5\u6536\u6587\u672c\u5d4c\u5165\uff0c\u79f0\u4e3aDiT-T2I\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u4e94\u3001\u5b9e\u9a8c\u7ed3\u679c\u4e0e\u5206\u6790<\/p>\n\n\n\n<ol class=\"wp-block-list\">\n<li>\u56fe\u50cf\u91cd\u5efa\u4e0e\u751f\u6210\u8d28\u91cf\uff1aTexTok \u5728\u5404\u4e2atoken\u6570\u91cf\u8bbe\u5b9a\u4e0b\u5747\u4f18\u4e8e\u672a\u4f7f\u7528\u6587\u672c\u6761\u4ef6\u7684Baseline\uff08w\/o text\uff09\uff0c\u5728ImageNet-256\u4e0a\u91cd\u5efaFID\u5e73\u5747\u63d0\u534729.2%\uff0c\u751f\u6210FID\u63d0\u534716.3%\uff1b\u5728ImageNet-512\u4e0a\u5206\u522b\u63d0\u534748.1%\u4e0e34.3%\u3002<\/li>\n\n\n\n<li>\u538b\u7f29\u7387\uff1aTexTok \u5728\u4ec5\u4f7f\u7528Baseline\u4e00\u534a\u4e43\u81f3\u56db\u5206\u4e4b\u4e00token\u6570\u7684\u60c5\u51b5\u4e0b\uff0c\u4ecd\u80fd\u8fbe\u5230\u76f8\u540c\u751a\u81f3\u66f4\u4f73\u7684\u91cd\u5efa\u8d28\u91cf\uff08\u5982\u56fe3\u6240\u793a\uff09\uff0c\u9a8c\u8bc1\u4e86\u5176\u5728\u538b\u7f29\u6548\u7387\u4e0a\u7684\u663e\u8457\u4f18\u52bf\u3002<\/li>\n\n\n\n<li>\u7cfb\u7edf\u6027\u80fd\u5bf9\u6bd4\uff1aTexTok + DiT \u5728ImageNet-256\u4e0a\u8fbe\u62101.46\u7684FID\uff0c\u5728ImageNet-512\u4e0a\u8fbe\u62101.62\u7684FID\uff0c\u5747\u4e3a\u5f53\u524d\u6700\u4f73\u6c34\u5e73\u3002\u540c\u65f6\uff0c\u5728\u4f7f\u752832\u4e2atoken\u65f6\uff0cTexTok\u5b9e\u73b0\u9ad8\u8fbe93.5\u00d7\u7684\u751f\u6210\u901f\u5ea6\u63d0\u5347\u3002<\/li>\n\n\n\n<li>\u6587\u672c\u5230\u56fe\u50cf\u751f\u6210\uff08Text-to-Image\uff09\uff1aTexTok \u572832\/64\/128-token\u6761\u4ef6\u4e0b\u5206\u522b\u63d0\u5347FID\u4e0eCLIP\u5206\u6570\uff0c\u4e14\u56fe\u50cf\u4e0e\u63cf\u8ff0\u5339\u914d\u5ea6\u663e\u8457\u4f18\u4e8eBaseline\uff0c\u5c24\u5176\u5728\u590d\u6742\u89c6\u89c9\u7ec6\u8282\uff08\u5982\u9e1f\u5599\u3001\u6c7d\u8f66\u8f6e\u80ce\uff09\u8fd8\u539f\u4e0a\u6548\u679c\u7a81\u51fa\u3002<\/li>\n<\/ol>\n\n\n\n<p class=\"wp-block-paragraph\">\u516d\u3001\u6d88\u878d\u5b9e\u9a8c\u4e0e\u5173\u952e\u56e0\u7d20\u5206\u6790<\/p>\n\n\n\n<ol class=\"wp-block-list\">\n<li>\u6587\u672c\u4fe1\u606f\u4e30\u5bcc\u5ea6\uff1a\u8d8a\u8be6\u7ec6\u7684caption\uff0825\u8bcd vs 75\u8bcd\uff09\u5e26\u6765\u66f4\u4f18\u7684\u91cd\u5efa\u8d28\u91cf\u3002<\/li>\n\n\n\n<li>\u6587\u672c\u7f16\u7801\u5668\u89c4\u6a21\uff1aT5\u4eceSmall\u5230XXL\u9010\u7ea7\u63d0\u5347\uff0c\u6a21\u578b\u8868\u73b0\u4e5f\u968f\u4e4b\u589e\u5f3a\u3002<\/li>\n\n\n\n<li>\u6761\u4ef6\u6ce8\u5165\u65b9\u5f0f\uff1a\u5c06\u6587\u672ctoken\u76f4\u63a5\u62fc\u63a5\u8f93\u5165ViT\uff08in-context conditioning\uff09\u4f18\u4e8ecross-attention\u65b9\u6848\u3002<\/li>\n\n\n\n<li>\u6ce8\u5165\u4f4d\u7f6e\uff1a\u5728Tokenizer\u4e0eDetokenizer\u5747\u6ce8\u5165\u6587\u672ctoken\u6548\u679c\u6700\u4f73\u3002<\/li>\n\n\n\n<li>\u6a21\u578b\u5927\u5c0f\uff1aTexTok-Base\u5728\u6027\u80fd\u4e0e\u6548\u7387\u95f4\u53d6\u5f97\u6700\u4f73\u5e73\u8861\u3002<\/li>\n<\/ol>\n\n\n\n<p class=\"wp-block-paragraph\">\u4e03\u3001\u4e0e\u76f8\u5173\u5de5\u4f5c\u7684\u6bd4\u8f83<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u4e0e\u5176\u4ed6\u65b9\u6cd5\u5982LQAE\u3001SPAE\u7b49\u76f4\u63a5\u5c06\u56fe\u50cf\u6620\u5c04\u81f3\u8bed\u8a00\u6a21\u578btoken\u6216\u5728\u5bf9\u9f50image-text\u8868\u793a\u4e0a\u5f3a\u5236\u7ea6\u675f\u4e0d\u540c\uff0cTexTok\u5e76\u4e0d\u8bd5\u56fe\u5c06\u56fe\u50cftoken\u4e0e\u8bed\u8a00\u7a7a\u95f4\u5f3a\u884c\u5bf9\u9f50\uff0c\u800c\u662f\u91c7\u7528\u8f6f\u6761\u4ef6\u65b9\u5f0f\u5f15\u5bfc\u89c6\u89c9\u538b\u7f29\u8fc7\u7a0b\uff0c\u56e0\u800c\u5728\u4fdd\u6301\u751f\u6210\u8d28\u91cf\u7684\u540c\u65f6\u4fdd\u7559\u4e86\u89c6\u89c9\u4fe1\u606f\u7ed3\u6784\uff0c\u907f\u514d\u8bed\u4e49\u635f\u5931\u3002<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<p class=\"wp-block-paragraph\">TexTok: <a href=\"https:\/\/kaiwenzha.github.io\/textok\/\" target=\"_blank\" rel=\"noreferrer noopener\">https:\/\/kaiwenzha.github.io\/textok\/<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"<p>\u8bba\u6587Language-Guided Image Tokenization for Generation\u63d0\u51fa\u7684T [&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":[39,80,66],"class_list":["post-5977","post","type-post","status-publish","format-standard","hentry","category-tech","tag-ai","tag-cv","tag-google"],"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":"\u8bba\u6587Language-Guided Image Tokenization for Generation\u63d0\u51fa\u7684T&hellip;","_links":{"self":[{"href":"https:\/\/nullthought.net\/index.php?rest_route=\/wp\/v2\/posts\/5977","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=5977"}],"version-history":[{"count":2,"href":"https:\/\/nullthought.net\/index.php?rest_route=\/wp\/v2\/posts\/5977\/revisions"}],"predecessor-version":[{"id":6241,"href":"https:\/\/nullthought.net\/index.php?rest_route=\/wp\/v2\/posts\/5977\/revisions\/6241"}],"wp:attachment":[{"href":"https:\/\/nullthought.net\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=5977"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/nullthought.net\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=5977"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/nullthought.net\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=5977"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}