{"id":6041,"date":"2025-07-03T12:27:54","date_gmt":"2025-07-03T04:27:54","guid":{"rendered":"https:\/\/nullthought.net\/?p=6041"},"modified":"2025-07-03T14:00:27","modified_gmt":"2025-07-03T06:00:27","slug":"llamafactory%ef%bc%9a%e6%94%af%e6%8c%81%e8%b6%85%e8%bf%87100%e7%a7%8d%e5%a4%a7%e6%a8%a1%e5%9e%8b%e7%9a%84llm%e9%ab%98%e6%95%88%e5%be%ae%e8%b0%83%e6%a1%86%e6%9e%b6","status":"publish","type":"post","link":"https:\/\/nullthought.net\/?p=6041","title":{"rendered":"LLaMAFactory\uff1a\u652f\u6301\u8d85\u8fc7100\u79cd\u5927\u6a21\u578b\u7684LLM\u9ad8\u6548\u5fae\u8c03\u6846\u67b6"},"content":{"rendered":"\n<p>\u8bba\u6587<strong><a href=\"https:\/\/arxiv.org\/abs\/2403.13372\" target=\"_blank\" rel=\"noreferrer noopener\">LlamaFactory: Unified Efficient Fine-Tuning of 100+ Language Models<\/a><\/strong>\u63d0\u51fa\u7684LLaMAFactory\uff0c\u662f\u5f53\u524d\u6700\u5168\u9762\u3001\u53ef\u6269\u5c55\u6027\u6781\u5f3a\u7684LLM\u9ad8\u6548<a href=\"https:\/\/nullthought.net\/?p=6045\" target=\"_blank\" rel=\"noreferrer noopener\"><strong>\u5fae\u8c03<\/strong><\/a>\u6846\u67b6\u4e4b\u4e00\uff0c\u517c\u5177\u6280\u672f\u6df1\u5ea6\u4e0e\u5b9e\u7528\u6027\u3002\u5728\u89e3\u51b3\u7edf\u4e00\u6027\u3001\u6613\u7528\u6027\u4e0e\u8bad\u7ec3\u6548\u7387\u8fd9\u4e09\u5927\u96be\u9898\u65b9\u9762\u63d0\u4f9b\u4e86\u7cfb\u7edf\u5316\u3001\u5de5\u7a0b\u5316\u7684\u89e3\u51b3\u65b9\u6848\uff0c\u5e76\u901a\u8fc7\u4e30\u5bcc\u5b9e\u9a8c\u8bc1\u660e\u5176\u6709\u6548\u6027\uff0c\u4ee3\u8868\u4e86\u5f53\u524d\u5f00\u6e90LLM\u5fae\u8c03\u5de5\u5177\u7684\u53d1\u5c55\u524d\u6cbf\u3002<\/p>\n\n\n\n<p>\u8bba\u6587\u4f5c\u8005\u4e3aYaowei Zheng\uff08\u90d1\u8000\u5a01\uff09, Richong Zhang, Junhao Zhang, Yanhan Ye, Zheyan Luo, Zhangchi Feng, Yongqiang Ma\uff0c\u4e3b\u8981\u6765\u81eaBeihang University\uff08\u5317\u4eac\u822a\u7a7a\u822a\u5929\u5927\u5b66\uff09\u3002<\/p>\n\n\n\n<figure class=\"wp-block-image aligncenter size-full is-resized\"><img loading=\"lazy\" decoding=\"async\" width=\"489\" height=\"609\" src=\"https:\/\/nullthought.net\/wp-content\/uploads\/2025\/07\/image-3.png\" alt=\"\" class=\"wp-image-6043\" style=\"width:332px;height:auto\" srcset=\"https:\/\/nullthought.net\/wp-content\/uploads\/2025\/07\/image-3.png 489w, https:\/\/nullthought.net\/wp-content\/uploads\/2025\/07\/image-3-241x300.png 241w\" sizes=\"auto, (max-width: 489px) 100vw, 489px\" \/><figcaption class=\"wp-element-caption\"><strong><a href=\"https:\/\/arxiv.org\/abs\/2403.13372\" target=\"_blank\" rel=\"noreferrer noopener\">LlamaFactory: Unified Efficient Fine-Tuning of 100+ Language Models<\/a><\/strong><\/figcaption><\/figure>\n\n\n\n<p>\u4e00\u3001\u7814\u7a76\u80cc\u666f\u4e0e\u52a8\u673a<\/p>\n\n\n\n<p>\u5f53\u524d\u5927\u8bed\u8a00\u6a21\u578b\uff08LLM\uff09\u5728\u8bf8\u591a\u81ea\u7136\u8bed\u8a00\u5904\u7406\u4efb\u52a1\u4e2d\u5c55\u73b0\u51fa\u5353\u8d8a\u80fd\u529b\uff0c\u4f46\u5c06\u5b83\u4eec\u9ad8\u6548\u5730\u5fae\u8c03\u5230\u4e0b\u6e38\u4efb\u52a1\u4ecd\u9762\u4e34\u663e\u8457\u6311\u6218\uff0c\u5c24\u5176\u5728\u8d44\u6e90\u6709\u9650\u7684\u6761\u4ef6\u4e0b\u3002\u5c3d\u7ba1\u793e\u533a\u5df2\u7ecf\u63d0\u51fa\u591a\u79cd\u53c2\u6570\u9ad8\u6548\u5fae\u8c03\u65b9\u6cd5\uff08\u5982LoRA\u3001QLoRA\u3001DoRA\u7b49\uff09\uff0c\u4f46\u7f3a\u4e4f\u4e00\u4e2a\u7edf\u4e00\u7684\u3001\u6a21\u5757\u5316\u7684\u5e73\u53f0\u6765\u6574\u5408\u8fd9\u4e9b\u65b9\u6cd5\u5e76\u9002\u914d\u4e0d\u540c\u6a21\u578b\u7ed3\u6784\u3002\u6b64\u5916\uff0c\u5927\u591a\u6570\u73b0\u6709\u5de5\u5177\u94fe\u4ecd\u9700\u8f83\u9ad8\u7684\u4ee3\u7801\u95e8\u69db\uff0c\u8fd9\u9650\u5236\u4e86\u975e\u4e13\u4e1a\u7528\u6237\u7684\u4f7f\u7528\u3002<\/p>\n\n\n\n<p>\u4e8c\u3001LLaMAFactory\u6846\u67b6\u6982\u8ff0<\/p>\n\n\n\n<p>LLaMAFactory\u662f\u4e00\u4e2a\u7edf\u4e00\u7684\u9ad8\u6548\u5fae\u8c03\u6846\u67b6\uff0c\u652f\u6301\u8d85\u8fc7100\u79cd\u4e3b\u6d41\u5f00\u6e90LLM\u6a21\u578b\uff0c\u5e76\u6574\u5408\u4e86\u591a\u4e2a\u9ad8\u6548\u5fae\u8c03\u6280\u672f\u3002\u8be5\u6846\u67b6\u7684\u6838\u5fc3\u4f18\u52bf\u5305\u62ec\uff1a<\/p>\n\n\n\n<ol class=\"wp-block-list\">\n<li><strong>\u6a21\u5757\u5316\u8bbe\u8ba1<\/strong>\uff1a\u5206\u4e3aModel Loader\u3001Data Worker\u4e0eTrainer\u4e09\u5927\u6a21\u5757\uff0c\u4e92\u76f8\u89e3\u8026\uff0c\u4fbf\u4e8e\u6269\u5c55\u4e0e\u7ef4\u62a4\u3002<\/li>\n\n\n\n<li><strong>\u53ef\u89c6\u5316\u754c\u9762LLaMABoard<\/strong>\uff1a\u57fa\u4e8eGradio\u6784\u5efa\uff0c\u7528\u6237\u65e0\u9700\u7f16\u7a0b\u5373\u53ef\u914d\u7f6e\u548c\u76d1\u63a7\u5fae\u8c03\u4efb\u52a1\u3002<\/li>\n\n\n\n<li><strong>\u5e7f\u6cdb\u652f\u6301\u6a21\u578b\u4e0e\u65b9\u6cd5<\/strong>\uff1a\u517c\u5bb9LoRA\u3001QLoRA\u3001DoRA\u3001PiSSA\u3001GaLore\u3001BAdam\u7b49\u5fae\u8c03\u65b9\u6cd5\uff0c\u4ee5\u53caSFT\u3001RLHF\u3001DPO\u3001KTO\u3001ORPO\u7b49\u8bad\u7ec3\u7b56\u7565\u3002<\/li>\n<\/ol>\n\n\n\n<p>\u4e09\u3001\u9ad8\u6548\u5fae\u8c03\u65b9\u6cd5\u8be6\u89e3<\/p>\n\n\n\n<p>\u8bba\u6587\u5c06\u5fae\u8c03\u65b9\u6cd5\u5206\u4e3a\u4e24\u7c7b\uff1a<\/p>\n\n\n\n<ol class=\"wp-block-list\">\n<li><strong>\u4f18\u5316\u6548\u7387\u5bfc\u5411\uff08Efficient Optimization\uff09<\/strong>\uff1a\n<ul class=\"wp-block-list\">\n<li><em>Freeze-tuning<\/em>\uff1a\u51bb\u7ed3\u5927\u90e8\u5206\u53c2\u6570\uff0c\u4ec5\u8bad\u7ec3\u5c11\u91cf\u89e3\u7801\u5668\u5c42\u3002<\/li>\n\n\n\n<li><em>LoRA \/ QLoRA<\/em>\uff1a\u5f15\u5165\u53ef\u8bad\u7ec3\u7684\u4f4e\u79e9\u77e9\u9635\uff0c\u540c\u65f6QLoRA\u7ed3\u5408\u6743\u91cd\u91cf\u5316\u8fdb\u4e00\u6b65\u964d\u4f4e\u5185\u5b58\u3002<\/li>\n\n\n\n<li><em>DoRA<\/em>\uff1a\u5bf9\u9884\u8bad\u7ec3\u6743\u91cd\u5206\u89e3\u4e3a\u65b9\u5411\u4e0e\u6a21\u957f\uff0c\u4ec5\u66f4\u65b0\u65b9\u5411\u90e8\u5206\u3002<\/li>\n\n\n\n<li><em>LoRA+<\/em> \u4e0e <em>PiSSA<\/em>\uff1a\u5206\u522b\u901a\u8fc7\u6539\u8fdb\u4f4e\u79e9\u521d\u59cb\u5316\u4e0e\u4e3b\u6210\u5206\u521d\u59cb\u5316\u52a0\u5feb\u6536\u655b\u3002<\/li>\n\n\n\n<li><em>GaLore<\/em>\uff1a\u5c06\u68af\u5ea6\u6295\u5f71\u81f3\u4f4e\u7ef4\u5b50\u7a7a\u95f4\u4ee5\u51cf\u5c11\u5185\u5b58\u3002<\/li>\n\n\n\n<li><em>BAdam<\/em>\uff1a\u57fa\u4e8e\u5757\u5750\u6807\u4e0b\u964d\u7684\u4f18\u5316\u5668\uff0c\u9002\u5408\u5168\u53c2\u6570\u5fae\u8c03\u3002<\/li>\n<\/ul>\n<\/li>\n\n\n\n<li><strong>\u8ba1\u7b97\u6548\u7387\u5bfc\u5411\uff08Efficient Computation\uff09<\/strong>\uff1a\n<ul class=\"wp-block-list\">\n<li>\u6df7\u5408\u7cbe\u5ea6\u8bad\u7ec3\uff08bfloat16\/float16\uff09\u3001\u6fc0\u6d3b\u68c0\u67e5\u70b9\u3001FlashAttention\u3001S2 Attention\u7b49\u3002<\/li>\n\n\n\n<li><em>Unsloth<\/em>\uff1a\u7528\u4e8eLoRA\u53cd\u5411\u4f20\u64ad\u7684\u4f18\u5316\u6280\u672f\uff0c\u5927\u5e45\u52a0\u901f\u8bad\u7ec3\u3002<\/li>\n\n\n\n<li>\u91cf\u5316\u652f\u6301\uff1a4\/8-bit \u6743\u91cd\u91cf\u5316\u652f\u6301\u4e0eLoRA\u517c\u5bb9\u3002<\/li>\n<\/ul>\n<\/li>\n<\/ol>\n\n\n\n<p>\u56db\u3001\u4e09\u5927\u6838\u5fc3\u6a21\u5757\u8bbe\u8ba1<\/p>\n\n\n\n<ol class=\"wp-block-list\">\n<li><strong>Model Loader<\/strong>\uff1a\n<ul class=\"wp-block-list\">\n<li>\u652f\u6301\u81ea\u52a8\u521d\u59cb\u5316\u3001\u6a21\u578b\u6253\u8865\u4e01\uff08\u5982\u542f\u7528Flash\/S2 Attention\uff09\u3001\u6743\u91cd\u91cf\u5316\uff08GPTQ\u3001AWQ\u3001AQLM\uff09\u53ca\u9002\u914d\u5668\u6302\u8f7d\uff08LoRA\u7b49\uff09\u3002<\/li>\n\n\n\n<li>\u652f\u6301\u4e0d\u540c\u8bbe\u5907\u7684\u7cbe\u5ea6\u9002\u914d\u7b56\u7565\uff08CUDA\u3001AMD\u3001Ascend\u7b49\uff09\u3002<\/li>\n<\/ul>\n<\/li>\n\n\n\n<li><strong>Data Worker<\/strong>\uff1a\n<ul class=\"wp-block-list\">\n<li>\u6570\u636e\u52a0\u8f7d\u652f\u6301\u672c\u5730\u4e0e\u8fdc\u7a0b\uff0c\u517c\u5bb9HF Hub\u3002<\/li>\n\n\n\n<li>\u652f\u6301\u591a\u79cd\u6570\u636e\u7ed3\u6784\uff08Plain\u3001Alpaca\u3001ShareGPT\u3001Preference\u3001\u6807\u51c6\u683c\u5f0f\uff09\uff0c\u901a\u8fc7\u5bf9\u9f50\u89c4\u8303\u7edf\u4e00\u7ed3\u6784\u3002<\/li>\n\n\n\n<li>\u652f\u6301\u591a\u6570\u636e\u96c6\u5408\u5e76\u4e0e\u6a21\u677f\u9884\u5904\u7406\uff0c\u81ea\u52a8\u9009\u62e9Chat\u6a21\u677f\u4e0eTokenizer\u8fdb\u884c\u7f16\u7801\u3002<\/li>\n<\/ul>\n<\/li>\n\n\n\n<li><strong>Trainer<\/strong>\uff1a\n<ul class=\"wp-block-list\">\n<li>\u96c6\u6210Transformers\u4e0eTRL\u5e93\u7684\u8bad\u7ec3\u5668\uff0c\u5b9e\u73b0\u9884\u8bad\u7ec3\u3001SFT\u3001RLHF\u4e0eDPO\u7b49\u3002<\/li>\n\n\n\n<li>\u63d0\u51fa\u201c\u6a21\u578b\u5171\u4eab\u578bRLHF\u201d\uff0c\u9996\u6b21\u652f\u6301\u5728\u5355\u6d88\u8d39\u7ea7\u8bbe\u5907\u4e0a\u5b8c\u6210RLHF\u5168\u6d41\u7a0b\uff08\u4e00\u4e2a\u6a21\u578b\u8f6e\u6362\u62c5\u4efbPolicy\u3001Reward\u7b49\u591a\u89d2\u8272\uff09\u3002<\/li>\n\n\n\n<li>\u652f\u6301DeepSpeed\u5206\u5e03\u5f0f\u8bad\u7ec3\uff08ZeRO\u4f18\u5316\u5668\uff09\uff0c\u6781\u5927\u964d\u4f4e\u5185\u5b58\u3002<\/li>\n<\/ul>\n<\/li>\n<\/ol>\n\n\n\n<p>\u4e94\u3001LLaMABoard\u754c\u9762\u7cfb\u7edf<\/p>\n\n\n\n<ol class=\"wp-block-list\">\n<li><strong>\u6613\u914d\u7f6e<\/strong>\uff1a\u901a\u8fc7Web UI\u914d\u7f6e\u6240\u6709\u53c2\u6570\uff0c\u5185\u7f6e\u63a8\u8350\u503c\u4e0e\u6570\u636e\u9884\u89c8\u529f\u80fd\u3002<\/li>\n\n\n\n<li><strong>\u8bad\u7ec3\u76d1\u63a7<\/strong>\uff1a\u5b9e\u65f6\u53ef\u89c6\u5316\u8bad\u7ec3\u65e5\u5fd7\u4e0eloss\u66f2\u7ebf\u3002<\/li>\n\n\n\n<li><strong>\u8bc4\u4f30\u529f\u80fd<\/strong>\uff1a\u652f\u6301BLEU\u3001ROUGE\u3001MMLU\u3001CMMLU\u3001C-Eval\u7b49\u81ea\u52a8\u8bc4\u4f30\uff0c\u6216\u901a\u8fc7\u5bf9\u8bdd\u8fdb\u884c\u4eba\u5de5\u8bc4\u4f30\u3002<\/li>\n\n\n\n<li><strong>\u591a\u8bed\u8a00\u652f\u6301<\/strong>\uff1a\u76ee\u524d\u652f\u6301\u4e2d\u6587\u3001\u82f1\u6587\u4e0e\u4fc4\u8bed\u3002<\/li>\n<\/ol>\n\n\n\n<p>\u516d\u3001\u5b9e\u9a8c\u9a8c\u8bc1<\/p>\n\n\n\n<ol class=\"wp-block-list\">\n<li><strong>\u8bad\u7ec3\u6548\u7387\u8bc4\u4f30<\/strong>\uff1a\n<ul class=\"wp-block-list\">\n<li>\u4f7f\u7528PubMed\u6570\u636e\u96c6\uff0c\u5206\u522b\u5bf9Gemma-2B\u3001Llama2-7B\u4e0e13B\u6a21\u578b\u8fdb\u884c\u5fae\u8c03\u3002<\/li>\n\n\n\n<li>\u6bd4\u8f83\u4e86Full-tuning\u3001Freeze\u3001GaLore\u3001LoRA\u3001QLoRA\u7b49\u65b9\u6cd5\u5728\u5185\u5b58\u3001\u541e\u5410\u4e0e\u56f0\u60d1\u5ea6\uff08PPL\uff09\u4e0a\u7684\u5dee\u5f02\u3002<\/li>\n\n\n\n<li>QLoRA\u5185\u5b58\u5360\u7528\u6700\u4f4e\uff0cLoRA\u5728\u5c0f\u6a21\u578b\u4e0a\u6548\u679c\u66f4\u4f18\uff0cGaLore\u5728\u5927\u6a21\u578b\u8868\u73b0\u66f4\u4f73\u3002<\/li>\n<\/ul>\n<\/li>\n\n\n\n<li><strong>\u4e0b\u6e38\u4efb\u52a1\u6548\u679c\u8bc4\u4f30<\/strong>\uff1a\n<ul class=\"wp-block-list\">\n<li>\u5728CNN\/DM\u3001XSum\u3001AdGen\u4e09\u4e2a\u6458\u8981\u4e0e\u751f\u6210\u4efb\u52a1\u4e0a\u8fdb\u884c\u5b9e\u9a8c\uff0c\u4f7f\u7528ROUGE\u4f5c\u4e3a\u8bc4\u4f30\u6307\u6807\u3002<\/li>\n\n\n\n<li>\u5728\u5927\u591a\u6570\u4efb\u52a1\u4e2d\uff0cLoRA\u4e0eQLoRA\u80fd\u53d6\u5f97\u6700\u4f18\u6216\u6b21\u4f18\u7ed3\u679c\uff0c\u5c55\u793a\u51fa\u6781\u5f3a\u7684\u9002\u5e94\u6027\u3002<\/li>\n\n\n\n<li>Llama3-8B\u6574\u4f53\u6027\u80fd\u6700\u4f73\uff0cYi-6B\u4e0eMistral-7B\u5728\u540c\u89c4\u6a21\u6a21\u578b\u4e2d\u8868\u73b0\u51fa\u8272\u3002<\/li>\n<\/ul>\n<\/li>\n<\/ol>\n\n\n\n<p>\u4e03\u3001\u672a\u6765\u5de5\u4f5c\u5c55\u671b<\/p>\n\n\n\n<p>\u4f5c\u8005\u63d0\u51fa\u4e09\u70b9\u672a\u6765\u6269\u5c55\u65b9\u5411\uff1a<\/p>\n\n\n\n<ol class=\"wp-block-list\">\n<li>\u6269\u5c55\u591a\u6a21\u6001\u6a21\u578b\u652f\u6301\uff08\u5982\u8bed\u97f3\u3001\u89c6\u9891\u6a21\u578b\uff09\u3002<\/li>\n\n\n\n<li>\u5f15\u5165\u66f4\u591a\u5e76\u884c\u8bad\u7ec3\u7b56\u7565\uff08\u5982\u5e8f\u5217\u5e76\u884c\u4e0e\u5f20\u91cf\u5e76\u884c\uff09\u3002<\/li>\n\n\n\n<li>\u52a0\u5f3a\u5bf9\u8bdd\u6a21\u578b\u7684\u5fae\u8c03\u80fd\u529b\uff0c\u5982\u5f15\u5165\u81ea\u535a\u5f08\uff08self-play\uff09\u673a\u5236\u3002<\/li>\n<\/ol>\n\n\n\n<p>\u516b\u3001\u793e\u4f1a\u5f71\u54cd\u4e0e\u5f00\u6e90\u751f\u6001<\/p>\n\n\n\n<p>LLaMAFactory\u5728GitHub\u5df2\u83b7\u5f97\u8d852.5\u4e07Star\u4e0e3\u5343Fork\uff08\u622a\u81f32025\u5e747\u6708\uff0c\u5df2\u83b7\u5f97\u8d855.34\u4e07Star\u4e0e6.5\u5343Fork\uff09\uff0c\u5e76\u88ab\u6536\u5f55\u81f3HuggingFace\u7684Awesome Transformers\u5217\u8868\u3002\u5176\u53ef\u89c6\u5316\u3001\u6a21\u5757\u5316\u3001\u652f\u6301\u5e7f\u6cdb\u6a21\u578b\u4e0e\u65b9\u6cd5\u7684\u7279\u70b9\uff0c\u6781\u5927\u4fc3\u8fdb\u4e86\u666e\u901a\u7528\u6237\u4e0e\u7814\u7a76\u4eba\u5458\u5bf9LLM\u7684\u63a2\u7d22\u4e0e\u521b\u65b0\u3002\u540c\u65f6\u4f5c\u8005\u5f3a\u8c03\u4e25\u683c\u9075\u5b88\u5f00\u6e90\u534f\u8bae\uff0c\u907f\u514d\u6ee5\u7528\u3002<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<p>LLaMAFactory on GitHub: <a href=\"https:\/\/github.com\/hiyouga\/LLaMA-Factory\" target=\"_blank\" rel=\"noreferrer noopener\">https:\/\/github.com\/hiyouga\/LLaMA-Factory<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"<p>\u8bba\u6587LlamaFactory: Unified Efficient Fine-Tuning of 100+ L [&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":[39,86],"class_list":["post-6041","post","type-post","status-publish","format-standard","hentry","category-tech","tag-ai","tag-llm"],"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\u6587LlamaFactory: Unified Efficient Fine-Tuning of 100+ L&hellip;","_links":{"self":[{"href":"https:\/\/nullthought.net\/index.php?rest_route=\/wp\/v2\/posts\/6041","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=6041"}],"version-history":[{"count":3,"href":"https:\/\/nullthought.net\/index.php?rest_route=\/wp\/v2\/posts\/6041\/revisions"}],"predecessor-version":[{"id":6047,"href":"https:\/\/nullthought.net\/index.php?rest_route=\/wp\/v2\/posts\/6041\/revisions\/6047"}],"wp:attachment":[{"href":"https:\/\/nullthought.net\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=6041"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/nullthought.net\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=6041"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/nullthought.net\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=6041"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}