{"id":13679,"date":"2025-03-19T11:59:24","date_gmt":"2025-03-19T18:59:24","guid":{"rendered":"https:\/\/dev.phison.com\/?p=13679"},"modified":"2026-06-08T15:06:02","modified_gmt":"2026-06-08T22:06:02","slug":"phison-expands-aidaptiv-gpu-memory-extension-capabilities-for-additional-platforms-to-enable-llm-training-and-improve-inferencing-on-premises","status":"publish","type":"post","link":"https:\/\/www.phison.com\/zh\/phison-expands-aidaptiv-gpu-memory-extension-capabilities-for-additional-platforms-to-enable-llm-training-and-improve-inferencing-on-premises\/","title":{"rendered":"\u7fa4\u8054\u6269\u5c55 aiDAPTIV+ GPU \u5185\u5b58\u6269\u5145\u80fd\u529b \u652f\u6301\u66f4\u591a\u5e73\u53f0 \u63d0\u5347\u8fb9\u7f18 LLM \u8bad\u7ec3\u4e0e\u63a8\u8bba\u6548\u80fd"},"content":{"rendered":"<p style=\"text-align: center;\">\u5168\u7403 NAND \u63a7\u5236\u82af\u7247\u66a8\u50a8\u5b58\u65b9\u6848\u6280\u672f\u9886\u5bfc\u5382\u5546 \u7fa4\u8054 (Phison, 8299TT) \u4eca\u65e5\u5ba3\u5e03\uff0c\u5176\u8fb9\u7f18 AI \u8bad\u7ec3\u4e0e\u63a8\u8bba\u89e3\u51b3\u65b9\u6848 aiDAPTIV+ \u518d\u5ea6\u5347\u7ea7\uff0c\u63d0\u4f9b\u66f4\u591a\u6269\u5145\u80fd\u529b\u3002<a href=\"https:\/\/www.phison.com\/zh-tw\/aidaptiv-plus-ai-data-storage-solution\">aiDAPTIV+<\/a>\u00a0\u5c06\u6574\u5408\u81f3 Maingear \u7684ML \u7cfb\u5217\u7b14\u7535\uff0c\u8fd9\u6b3e AI \u7b14\u7535\u5c06\u6210\u4e3a\u5168\u7403\u9996\u6b3e\u652f\u6301 LLMOps (Large Language Model Operations; \u5927\u578b\u8bed\u8a00\u6a21\u578b\u4f5c\u4e1a) \u7684\u6d88\u8d39\u7ea7\u7b14\u7535\uff0c\u7528\u6237\u7528\u6237\u5c06\u53ef\u4f7f\u7528\u81ea\u8eab\u4f01\u4e1a\u6216\u4e2a\u4eba\u8d44\u6599\u6765\u5fae\u8c03 (Post-training\/Fine-tuning) \u6700\u591a 80 \u4ebf (8B) \u53c2\u6570\u7684\u5927\u578b\u8bed\u8a00\u6a21\u578b (LLM)\uff0c\u8ba9\u9884\u8bad\u7ec3 (Pre-trained) \u7684AI\u6a21\u578b\u80fd\u4f9d\u636e\u8fb9\u7f18\u5e94\u7528\u573a\u666f\u4e0e\u4f7f\u7528\u8005\u9700\u6c42\u8fdb\u884c\u5ba2\u5236\u5316 (Post-training)\uff0c\u4ee5\u8fbe\u5230\u771f\u6b63\u7684AI\u843d\u5730\u666e\u53ca\u5316\u3002\u6b64\u6b3e\u642d\u8f7d NVIDIA GPU\u7684AI\u7b14\u7535\u5c06\u4e8e\u672c\u5468 (3\/17-21, San Jose, USA) \u4e8eNVIDIA GTC 2025 \u5c55\u793a\u6982\u5ff5\u673a\u5e76\u5f00\u653e\u767b\u8bb0\u8d2d\u4e70\u3002\u6b64\u5916\uff0c\u7fa4\u8054\u4e5f\u5c06 aiDAPTIV+ \u6269\u5c55\u81f3\u652f\u6301\u00a0<a href=\"https:\/\/www.nvidia.com\/en-us\/autonomous-machines\/embedded-systems\/\" target=\"_blank\" rel=\"noopener\">NVIDIA Jetson \u5e73\u53f0<\/a>\u7684\u8fb9\u7f18\u8fd0\u7b97\u8bbe\u5907\uff0c\u5f3a\u5316\u5728\u8fb9\u7f18\u7aef\u7684\u751f\u6210\u5f0f AI \u63a8\u8bba\u4e0e\u673a\u5668\u4eba\u5e94\u7528\u3002<\/p>\n<p><em><img fetchpriority=\"high\" decoding=\"async\" class=\"aligncenter\" src=\"\/wp-content\/uploads\/2025\/03\/GTC-PR-Image.jpg\" width=\"786\" height=\"438\" data-path=\"local-images:\/image\/Press-Releases\/GTC PR Image.jpg\" \/><\/em><\/p>\n<p>\u968f\u7740\u6b64\u6b21\u5347\u7ea7\uff0caiDAPTIV+ \u7528\u6237\u5c06\u8fce\u6765\u5168\u65b0\u7684 aiDAPTIVLink 3.0 \u4e2d\u95f4\u4ef6 (Middleware Software)\uff0c\u63d0\u4f9b\u66f4\u5feb\u7684\u9996\u6b21\u54cd\u5e94\u65f6\u95f4 (Time to First Token, TTFT)\uff0c\u5e76\u5ef6\u957f Token \u957f\u5ea6\uff0c\u4ee5\u63d0\u5347\u63a8\u8bba\u6548\u80fd\u4e0e\u51c6\u786e\u6027\u3002\u6b64\u6269\u5c55\u5c06\u8fdb\u4e00\u6b65\u5f00\u542f\u66f4\u591a\u5e94\u7528\u673a\u4f1a\uff0c\u65e0\u8bba\u662f\u5927\u5b66\u751f\u6216AI \u4ea7\u4e1a\u4e13\u4e1a\u4eba\u58eb\u5b66\u4e60 LLM \u8bad\u7ec3\uff0c\u6216\u662f\u7814\u7a76\u4eba\u5458\u900f\u8fc7 PC \u6df1\u5165\u89e3\u6790\u81ea\u8eab\u6570\u636e\uff0c\u751a\u81f3\u662f\u5236\u9020\u5de5\u7a0b\u5e08\u900f\u8fc7\u8fb9\u7f18\u8fd0\u7b97\u8bbe\u5907\u81ea\u52a8\u5316\u5de5\u5382\u751f\u4ea7\u6d41\u7a0b\uff0c\u7686\u80fd\u53d7\u76ca\u4e8e aiDAPTIV+\u6b64\u6b21\u7684\u529f\u80fd\u5347\u7ea7\u3002<\/p>\n<p>AI \u4e0e\u8fb9\u7f18\u8fd0\u7b97\u5e94\u7528\u573a\u666f\u65e5\u76ca\u589e\u591a\uff0c\u5e02\u573a\u5bf9 AI \u5f00\u53d1\u4eba\u624d\u7684\u9700\u6c42\u6301\u7eed\u6500\u5347\u3002\u5f00\u53d1\u8005\u9700\u8981\u4eb2\u8eab\u64cd\u4f5c LLM \u8bad\u7ec3\u89e3\u51b3\u65b9\u6848\uff0c\u4ee5\u5efa\u7acb\u672a\u6765AI\u5e94\u7528\u6240\u9700\u8981\u7684AI\u8f6f\u4ef6\u6216\u662f\u4e13\u5c5e\u8fb9\u7f18AI\u6a21\u578b\uff0c\u800c\u653f\u5e9c\u673a\u6784\u3001\u7814\u7a76\u673a\u6784\u3001\u533b\u7597\u4e0e\u5de5\u4e1a\u7ec4\u7ec7\u7b49\u9ad8\u5ea6\u53d7\u76d1\u7ba1\u4e4b\u4ea7\u4e1a\uff0c\u5219\u9700\u8981\u8fb9\u7f18\u3001\u5b89\u5168\u7684\u5730\u7aefAI\u8bbe\u5907\u6765\u8fdb\u884c\u79c1\u6709\u6570\u636e\u7684 AI\u6a21\u578b\u5fae\u8c03\u8bad\u7ec3 (Post-training\/Fine-tuning)\u3002\u6b64\u5916\uff0c\u5e02\u573a\u5bf9 LLM \u63a8\u8bba\u901f\u5ea6\u4e0e\u51c6\u786e\u5ea6\u7684\u8981\u6c42\u4e0d\u65ad\u63d0\u5347\uff0c\u4f01\u4e1a\u66f4\u5e0c\u671b\u5728\u53ef\u9884\u6d4b\u7684\u6210\u672c\u4e0b\uff0c\u83b7\u5f97\u6700\u4f73\u89e3\u51b3\u65b9\u6848\u3002aiDAPTIV+ \u4ee5\u7ecf\u6d4e\u5b9e\u60e0\u7684 GPU \u5185\u5b58\u6269\u5145\u80fd\u529b\uff0c\u8ba9\u7528\u6237\u53ef\u5728\u8fb9\u7f18\u5c01\u95ed\u5f0f\u5b89\u5168\u7f51\u7edc\u5185\u5fae\u8c03\u8bad\u7ec3\u5927\u8bed\u8a00\u6a21\u578b (LLM)\uff0c\u5e76\u900f\u8fc7\u76f4\u89c9\u5316\u8f6f\u4ef6\u63a5\u53e3\u4e0eAI\u6a21\u578b\u4e92\u52a8\uff0c\u83b7\u53d6\u6240\u9700\u7684AI\u751f\u6210\u5173\u952e\u4fe1\u606f\u3002<\/p>\n<p>\u300c\u5ef6\u7eed\u53bb\u5e74 GTC \u5c55\u51fa\u7684 aiDAPTIV+ AI \u8bad\u7ec3\u5de5\u4f5c\u7ad9\u70ed\u6f6e\uff0c\u6211\u4eec\u4e0e\u7fa4\u8054\u6df1\u5316\u5408\u4f5c\uff0c\u63a8\u51fa\u6982\u5ff5\u6027 AI \u7b14\u7535\uff0c\u8ba9\u5f3a\u5927\u7684\u8fb9\u7f18 AI \u63a8\u8bba\u4e0e\u88c5\u7f6e\u7aef\u6a21\u578b\u8bad\u7ec3\u65e0\u5904\u4e0d\u5728\u3002\u300dMaingear \u6267\u884c\u957f Wallace Santos \u8868\u793a\uff0c\u300c\u8fd9\u9879\u6280\u672f\u7a81\u7834\u5c06\u8fb9\u7f18 AI \u80fd\u529b\u5e26\u5165\u8f7b\u4fbf\u7684\u884c\u52a8\u5e73\u53f0\uff0c\u4e3a\u4f01\u4e1a\u3001\u6559\u80b2\u673a\u6784\u4e0e\u5b66\u751f\u63d0\u4f9b\u7ecf\u6d4e\u9ad8\u6548\u7684 AI \u5f00\u53d1\u5de5\u5177\uff0c\u65e0\u9700\u501a\u8d56\u4f20\u7edf\u6570\u636e\u4e2d\u5fc3\u57fa\u7840\u8bbe\u65bd\u3002\u300d<\/p>\n<p>(\u82f1\u6587\u539f\u6587: \u201cBuilding on the momentum of last year\u2019s AI training workstation featuring aiDAPTIV+ at GTC, we\u2019ve expanded our collaboration with Phison to take AI performance even further by introducing a proof-of-concept laptop that enables both powerful edge AI inference and on-device model training from anywhere,\u201d said Wallace Santos, CEO of Maingear.\u00a0 \u201cThis brings the capabilities of on-prem AI to a compact, mobile platform, giving businesses, educators, and students a cost-effective tool to develop AI models without the complexity of traditional data center infrastructure.\u201d)<\/p>\n<p>\u7fa4\u8054\u7535\u5b50\u6267\u884c\u957f\u6f58\u5065\u6210\u8868\u793a\uff0cAI\u7684\u53d1\u5c55\u5df2\u7ecf\u6765\u5230\u4e00\u4e2a\u5173\u952e\u65f6\u523b\uff0c\u4f01\u4e1a\u3001\u5b66\u672f\u754c\u4e0e\u5f00\u53d1\u8005\u90fd\u9700\u8981\u66f4\u7075\u6d3b\u3001\u9ad8\u6548\u7684\u8fb9\u7f18AI\u89e3\u51b3\u65b9\u6848\u6765\u6269\u5927AI\u7684\u843d\u5730\u5e94\u7528\u3002\u7fa4\u8054\u7684aiDAPTIV+\u4e0d\u4ec5\u8ba9\u7528\u6237\u80fd\u5728\u4e2a\u4eba\u7b14\u7535\u6216\u8fb9\u7f18\u8bbe\u5907\u4e0a\u8bad\u7ec3\u4e0e\u63a8\u7406AI\u6a21\u578b\uff0c\u8fd8\u786e\u4fdd\u6570\u636e\u5b89\u5168\u3001\u964d\u4f4e\u6210\u672c\uff0c\u771f\u6b63\u5b9e\u73b0AI\u8fb9\u7f18\u8fd0\u7b97\u7684\u4fbf\u5229\u6027\u3002\u4eca\u5e74\uff0c\u6211\u4eec\u4e0eMaingear\u643a\u624b\u63a8\u51fa\u5168\u7403\u9996\u6b3e\u652f\u6301LLMOps\u7684AI\u7b14\u7535\uff0c\u5e76\u5f3a\u5316aiDAPTIVLink 3.0\u7684\u6548\u80fd\uff0c\u8ba9LLM\u7684\u54cd\u5e94\u901f\u5ea6\u66f4\u5feb\u3001\u7406\u89e3\u529b\u66f4\u5f3a\u3002\u6b64\u5916\uff0c\u7fa4\u8054\u8fdb\u4e00\u6b65\u6269\u5c55\u4e0eNVIDIA Jetson\u5e73\u53f0\u7684\u5408\u4f5c\uff0c\u5c06aiDAPTIV+\u5bfc\u5165\u8fb9\u7f18\u8fd0\u7b97\u4e0e\u673a\u5668\u4eba\u5e94\u7528\uff0c\u8ba9\u66f4\u591a\u4ea7\u4e1a\u5982\u5236\u9020\u3001\u533b\u7597\u3001\u96f6\u552e\u3001\u667a\u6167\u76d1\u63a7\u7b49\u9886\u57df\uff0c\u90fd\u80fd\u900f\u8fc7AI\u63d0\u5347\u8425\u8fd0\u6548\u7387\u4e0e\u51b3\u7b56\u7cbe\u51c6\u5ea6\u3002\u900f\u8fc7\u8fd9\u4e9b\u6280\u672f\u7a81\u7834\uff0c\u6211\u4eec\u5e0c\u671b\u5e2e\u52a9\u66f4\u591a\u4f01\u4e1a\u4e0e\u5f00\u53d1\u8005\u6253\u7834AI\u5bfc\u5165\u6280\u672f\u95e8\u574e\uff0c\u52a0\u901fAI\u843d\u5730\u5e94\u7528\uff0c\u8ba9\u667a\u6167\u8fd0\u7b97\u771f\u6b63\u878d\u5165\u65e5\u5e38\u5de5\u4f5c\u4e0e\u751f\u6d3b\u3002<\/p>\n<p><strong>aiDAPTIV+\u00a0<\/strong><strong>\u6700\u65b0\u5347\u7ea7\u529f\u80fd\uff1a<\/strong><\/p>\n<p>\u25b6\u00a0<a href=\"https:\/\/phisonblog.com\/maingear-pc-with-phisons-aidaptiv-technology-named-one-of-the-10-cool-ai-pcs-you-can-get\/\" target=\"_blank\" rel=\"noopener\"><strong>Maingear ML\u00a0<\/strong><strong>\u7cfb\u5217<\/strong><strong>\u00a0AI\u00a0<\/strong><strong>\u7b14\u7535<\/strong><\/a><strong>\uff0c\u5185\u5efa<\/strong><strong>\u00a0aiDAPTIV+\u00a0<\/strong><strong>\u652f\u63f4<\/strong><strong>\u00a0LLMOps<\/strong><\/p>\n<p>\u8fd9\u6b3e AI \u7b14\u7535\u4e3a\u5168\u7403\u9996\u6b3e\u652f\u6301 LLMOps \u5fae\u8c03\u8bad\u7ec3\u4e0e\u63a8\u8bba\u7684\u6d88\u8d39\u7ea7\u7b14\u7535\uff0c\u53ef\u5904\u7406\u6700\u591a 80 \u4ebf\u53c2\u6570\u7684\u6a21\u578b\u3002Maingear AI \u7b14\u7535\u642d\u8f7d aiDAPTIVCache SSD\u3001aiDAPTIVLink \u4e2d\u95f4\u4ef6\uff0c\u4ee5\u53ca Pro Suite \u56fe\u5f62\u5316 LLMOps \u5168\u65b9\u4f4d\u529f\u80fd\uff0c\u6db5\u76d6\u6570\u636e\u5bfc\u5165\u3001\u5fae\u8c03\u3001\u68c0\u7d22\u589e\u5f3a\u751f\u6210 (RAG)\u3001\u76d1\u63a7\u3001\u9a8c\u8bc1\u4e0e\u63a8\u8bba\u3002\u7528\u6237\u53ef\u900f\u8fc7 Maingear \u767b\u8bb0\uff0c\u83b7\u53d6\u8ba2\u8d2d\u4e0e\u51fa\u8d27\u901a\u77e5\u3002<\/p>\n<p>\u25b6\u00a0<strong>aiDAPTIVLink 3.0\u00a0<\/strong><strong>\u4e2d\u95f4\u4ef6<\/strong><\/p>\n<p>\u65b0\u7248 aiDAPTIVLink 3.0 \u53ef\u5f3a\u5316\u63a8\u8bba\u80fd\u529b\uff0c\u63d0\u4f9b\u66f4\u5feb\u7684 TTFT \u54cd\u5e94\u901f\u5ea6\uff0c\u5e76\u652f\u6301\u66f4\u957f\u7684 Token \u957f\u5ea6\uff0c\u63d0\u5347\u8bed\u610f\u7406\u89e3\u4e0e\u56de\u7b54\u51c6\u786e\u5ea6\u3002\u6b64\u529f\u80fd\u5c06\u4e8e 2025 \u5e74 4 \u6708\u63a8\u51fa\u3002<\/p>\n<p>\u25b6\u00a0<strong>aiDAPTIV+ AI\u00a0<\/strong><strong>\u63a8\u8bba\u652f\u6301<\/strong><strong>\u00a0IoT\u00a0<\/strong><strong>\u8fb9\u7f18\u8fd0\u7b97<\/strong><\/p>\n<p>aiDAPTIV+ \u73b0\u5df2\u652f\u6301\u8fb9\u7f18\u8fd0\u7b97\u4e0e\u673a\u5668\u4eba\u5e94\u7528\uff0c\u5e76\u901a\u8fc7\u7fa4\u8054\u9a8c\u8bc1\uff0c\u53ef\u5728 NVIDIA Jetson \u5e73\u53f0\u8bbe\u5907\u4e0a\u8fd0\u884c\u3002aiDAPTIV+ \u900f\u8fc7 aiDAPTIVCache SSD \u5f3a\u5316\u63a8\u8bba\u80fd\u529b\uff0c\u5e76\u652f\u6301 LoRA \u67b6\u6784 LLM \u8bad\u7ec3\uff0c\u5c06\u4e8e 2025 \u5e74 4 \u6708\u4e0a\u5e02\u3002\u6b64\u5347\u7ea7\u53ef\u5e94\u7528\u4e8e\u81ea\u52a8\u9a7e\u9a76\u3001\u533b\u7597\u8bca\u65ad\u3001\u5de5\u4e1a\u81ea\u52a8\u5316\u3001\u96f6\u552e\u5206\u6790\u3001\u73af\u5883\u76d1\u6d4b\u3001\u7535\u4fe1\u3001\u667a\u6167\u76d1\u63a7\u53ca\u667a\u6167\u519c\u4e1a\u7b49\u9886\u57df\u3002<\/p>\n<p>\u6b32\u4e86\u89e3\u66f4\u591a\u4fe1\u606f\uff0c\u6b22\u8fce\u8385\u4e34 NVIDIA GTC \u5c55\u89c8\u644a\u4f4d (2224)\uff0c\u73b0\u573a\u5c06\u5c55\u793a Maingear AI \u7b14\u7535\u4e0e aiDAPTIV+ \u5728\u8fb9\u7f18\u8fd0\u7b97\u4e0a\u7684 AI \u63a8\u8bba\u5e94\u7528\uff0c\u6216\u524d\u5f80\u5b98\u65b9\u7f51\u7ad9\u67e5\u8be2\u6700\u65b0\u6d88\u606f\u3002<\/p>\n","protected":false},"excerpt":{"rendered":"<p>\u5168\u7403 NAND \u63a7\u5236\u82af\u7247\u66a8\u50a8\u5b58\u65b9\u6848\u6280\u672f\u9886\u5bfc\u5382\u5546 \u7fa4\u8054 (Phison, 8299TT) \u4eca\u65e5\u5ba3\u5e03\uff0c\u5176\u8fb9\u7f18 AI \u8bad\u7ec3\u4e0e\u63a8\u8bba\u89e3\u51b3\u65b9\u6848 aiDAPTIV+ \u518d\u5ea6\u5347\u7ea7\uff0c\u63d0\u4f9b\u66f4\u591a\u6269\u5145\u80fd\u529b\u3002aiDAPTIV+\u00a0\u5c06\u6574\u5408\u81f3 Maingear \u7684ML \u7cfb\u5217\u7b14\u7535\uff0c\u8fd9\u6b3e AI \u7b14\u7535\u5c06\u6210\u4e3a\u5168\u7403\u9996\u6b3e\u652f\u6301 LLMOps (Large Language Model Operations; \u5927\u578b\u8bed\u8a00\u6a21\u578b\u4f5c\u4e1a) \u7684\u6d88\u8d39\u7ea7\u7b14\u7535\uff0c\u7528\u6237\u7528\u6237\u5c06\u53ef\u4f7f\u7528\u81ea\u8eab\u4f01\u4e1a\u6216\u4e2a\u4eba\u8d44\u6599\u6765\u5fae\u8c03 (Post-training\/Fine-tuning) \u6700\u591a 80 \u4ebf (8B) \u53c2\u6570\u7684\u5927\u578b\u8bed\u8a00\u6a21\u578b (LLM)\uff0c\u8ba9\u9884\u8bad\u7ec3 (Pre-trained) \u7684AI\u6a21\u578b\u80fd\u4f9d\u636e\u8fb9\u7f18\u5e94\u7528\u573a\u666f\u4e0e\u4f7f\u7528\u8005\u9700\u6c42\u8fdb\u884c\u5ba2\u5236\u5316 (Post-training)\uff0c\u4ee5\u8fbe\u5230\u771f\u6b63\u7684AI\u843d\u5730\u666e\u53ca\u5316\u3002\u6b64\u6b3e\u642d\u8f7d NVIDIA GPU\u7684AI\u7b14\u7535\u5c06\u4e8e\u672c\u5468 (3\/17-21, San Jose, USA) \u4e8eNVIDIA GTC 2025 \u5c55\u793a\u6982\u5ff5\u673a\u5e76\u5f00\u653e\u767b\u8bb0\u8d2d\u4e70\u3002\u6b64\u5916\uff0c\u7fa4\u8054\u4e5f\u5c06 aiDAPTIV+ \u6269\u5c55\u81f3\u652f\u6301\u00a0NVIDIA Jetson \u5e73\u53f0\u7684\u8fb9\u7f18\u8fd0\u7b97\u8bbe\u5907\uff0c\u5f3a\u5316\u5728\u8fb9\u7f18\u7aef\u7684\u751f\u6210\u5f0f AI \u63a8\u8bba\u4e0e\u673a\u5668\u4eba\u5e94\u7528\u3002 \u968f\u7740\u6b64\u6b21\u5347\u7ea7\uff0caiDAPTIV+ \u7528\u6237\u5c06\u8fce\u6765\u5168\u65b0\u7684 aiDAPTIVLink 3.0 \u4e2d\u95f4\u4ef6 (Middleware [&hellip;]<\/p>\n","protected":false},"author":9,"featured_media":15659,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[394],"tags":[],"years":[28],"class_list":["post-13679","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-press-releases","years-28"],"_links":{"self":[{"href":"https:\/\/www.phison.com\/zh\/wp-json\/wp\/v2\/posts\/13679","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.phison.com\/zh\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.phison.com\/zh\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.phison.com\/zh\/wp-json\/wp\/v2\/users\/9"}],"replies":[{"embeddable":true,"href":"https:\/\/www.phison.com\/zh\/wp-json\/wp\/v2\/comments?post=13679"}],"version-history":[{"count":2,"href":"https:\/\/www.phison.com\/zh\/wp-json\/wp\/v2\/posts\/13679\/revisions"}],"predecessor-version":[{"id":23696,"href":"https:\/\/www.phison.com\/zh\/wp-json\/wp\/v2\/posts\/13679\/revisions\/23696"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.phison.com\/zh\/wp-json\/wp\/v2\/media\/15659"}],"wp:attachment":[{"href":"https:\/\/www.phison.com\/zh\/wp-json\/wp\/v2\/media?parent=13679"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.phison.com\/zh\/wp-json\/wp\/v2\/categories?post=13679"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.phison.com\/zh\/wp-json\/wp\/v2\/tags?post=13679"},{"taxonomy":"years","embeddable":true,"href":"https:\/\/www.phison.com\/zh\/wp-json\/wp\/v2\/years?post=13679"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}