{"id":29355,"date":"2025-03-25T16:51:49","date_gmt":"2025-03-25T08:51:49","guid":{"rendered":"https:\/\/www.aisharenet.com\/?p=29355"},"modified":"2025-03-25T16:51:49","modified_gmt":"2025-03-25T08:51:49","slug":"rf-detr","status":"publish","type":"post","link":"https:\/\/www.kdjingpai.com\/ja\/rf-detr\/","title":{"rendered":"RF-DETR\uff1a\u5b9e\u65f6\u89c6\u89c9\u5bf9\u8c61\u68c0\u6d4b\u5f00\u6e90\u6a21\u578b"},"content":{"rendered":"<p>RF-DETR \u662f Roboflow \u56e2\u961f\u5f00\u53d1\u7684\u4e00\u4e2a\u5f00\u6e90\u5bf9\u8c61\u68c0\u6d4b\u6a21\u578b\u3002\u5b83\u57fa\u4e8e <a href=\"https:\/\/www.kdjingpai.com\/ja\/transformer\/\">Transformer<\/a> \u67b6\u6784\uff0c\u6838\u5fc3\u7279\u70b9\u662f\u5b9e\u65f6\u9ad8\u6548\u3002\u6a21\u578b\u5728\u5fae\u8f6f COCO \u6570\u636e\u96c6\u4e0a\u9996\u6b21\u5b9e\u73b0\u8d85\u8fc7 60 AP \u7684\u5b9e\u65f6\u68c0\u6d4b\uff0c\u540c\u65f6\u5728 RF100-VL \u57fa\u51c6\u6d4b\u8bd5\u4e2d\u8868\u73b0\u7a81\u51fa\uff0c\u9002\u5e94\u591a\u79cd\u5b9e\u9645\u573a\u666f\u3002\u5b83\u6709\u4e24\u79cd\u7248\u672c\uff1aRF-DETR-base\uff082900 \u4e07\u53c2\u6570\uff09\u548c RF-DETR-large\uff081.28 \u4ebf\u53c2\u6570\uff09\u3002\u6a21\u578b\u4f53\u79ef\u5c0f\uff0c\u9002\u5408\u8fb9\u7f18\u8bbe\u5907\u90e8\u7f72\u3002\u4ee3\u7801\u548c\u9884\u8bad\u7ec3\u6743\u91cd\u4f7f\u7528 Apache 2.0 \u8bb8\u53ef\uff0c\u514d\u8d39\u5f00\u653e\u7ed9\u793e\u533a\u4f7f\u7528\u3002\u7528\u6237\u53ef\u4ee5\u4ece GitHub \u83b7\u53d6\u8d44\u6e90\uff0c\u8f7b\u677e\u8bad\u7ec3\u6216\u90e8\u7f72\u3002<\/p>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"aligncenter size-full wp-image-29356\" title=\"RF-DETR\uff1a\u5b9e\u65f6\u89c6\u89c9\u5bf9\u8c61\u68c0\u6d4b\u5f00\u6e90\u6a21\u578b-1\" src=\"https:\/\/www.kdjingpai.com\/wp-content\/uploads\/2025\/03\/9433146588d333c.png\" alt=\"RF-DETR\uff1a\u5b9e\u65f6\u89c6\u89c9\u5bf9\u8c61\u68c0\u6d4b\u5f00\u6e90\u6a21\u578b-1\" width=\"1031\" height=\"616\" srcset=\"https:\/\/www.kdjingpai.com\/wp-content\/uploads\/2025\/03\/9433146588d333c.png 1031w, https:\/\/www.kdjingpai.com\/wp-content\/uploads\/2025\/03\/9433146588d333c-768x459.png 768w, https:\/\/www.kdjingpai.com\/wp-content\/uploads\/2025\/03\/9433146588d333c-18x12.png 18w\" sizes=\"auto, (max-width: 1031px) 100vw, 1031px\" \/><\/p>\n<p>&nbsp;<\/p>\n<h2>\u529f\u80fd\u5217\u8868<\/h2>\n<ul>\n<li>\u5b9e\u65f6\u5bf9\u8c61\u68c0\u6d4b\uff1a\u5728\u56fe\u50cf\u6216\u89c6\u9891\u4e2d\u5feb\u901f\u8bc6\u522b\u7269\u4f53\uff0c\u5ef6\u8fdf\u4f4e\u3002<\/li>\n<li>\u81ea\u5b9a\u4e49\u6570\u636e\u96c6\u8bad\u7ec3\uff1a\u652f\u6301\u7528\u81ea\u5df1\u7684\u6570\u636e\u8c03\u6574\u6a21\u578b\u3002<\/li>\n<li>\u8fb9\u7f18\u8bbe\u5907\u8fd0\u884c\uff1a\u6a21\u578b\u8f7b\u91cf\uff0c\u9002\u5408\u8d44\u6e90\u6709\u9650\u7684\u8bbe\u5907\u3002<\/li>\n<li>\u53ef\u8c03\u5206\u8fa8\u7387\uff1a\u7528\u6237\u53ef\u4ee5\u5e73\u8861\u68c0\u6d4b\u901f\u5ea6\u548c\u7cbe\u5ea6\u3002<\/li>\n<li>\u9884\u8bad\u7ec3\u6a21\u578b\u652f\u6301\uff1a\u63d0\u4f9b\u57fa\u4e8e COCO \u6570\u636e\u96c6\u7684\u9884\u8bad\u7ec3\u6743\u91cd\u3002<\/li>\n<li>\u89c6\u9891\u6d41\u5904\u7406\uff1a\u80fd\u5b9e\u65f6\u5206\u6790\u89c6\u9891\u5e76\u8f93\u51fa\u7ed3\u679c\u3002<\/li>\n<li>ONNX \u5bfc\u51fa\uff1a\u652f\u6301\u8f6c\u4e3a ONNX \u683c\u5f0f\uff0c\u65b9\u4fbf\u8de8\u5e73\u53f0\u90e8\u7f72\u3002<\/li>\n<li>\u591a GPU \u8bad\u7ec3\uff1a\u53ef\u4ee5\u7528\u591a\u5f20\u663e\u5361\u52a0\u901f\u8bad\u7ec3\u8fc7\u7a0b\u3002<\/li>\n<\/ul>\n<p>&nbsp;<\/p>\n<h2>\u4f7f\u7528\u5e2e\u52a9<\/h2>\n<p>RF-DETR \u7684\u4f7f\u7528\u5206\u4e3a\u5b89\u88c5\u3001\u63a8\u7406\u548c\u8bad\u7ec3\u4e09\u90e8\u5206\u3002\u4ee5\u4e0b\u662f\u8be6\u7ec6\u6b65\u9aa4\uff0c\u5e2e\u52a9\u4f60\u5feb\u901f\u4e0a\u624b\u3002<\/p>\n<h3>\u5b89\u88c5\u6d41\u7a0b<\/h3>\n<ol>\n<li><strong>\u73af\u5883\u51c6\u5907<\/strong><br \/>\n\u9700\u8981 Python 3.9 \u6216\u4ee5\u4e0a\u7248\u672c\uff0c\u4ee5\u53ca PyTorch 1.13.0 \u6216\u66f4\u9ad8\u3002\u5982\u679c\u7528 GPU\uff0c\u8fd0\u884c\u00a0<code>nvidia-smi<\/code>\u00a0\u68c0\u67e5\u9a71\u52a8\u3002<\/p>\n<ul>\n<li>\u5b89\u88c5 PyTorch\uff1a\n<pre><code>pip install torch&gt;=1.13.0 torchvision&gt;=0.14.0\r\n<\/code><\/pre>\n<\/li>\n<li>\u4e0b\u8f7d\u4ee3\u7801\uff1a\n<pre><code>git clone https:\/\/github.com\/roboflow\/rf-detr.git\r\ncd rf-detr\r\n<\/code><\/pre>\n<\/li>\n<li>\u5b89\u88c5\u4f9d\u8d56\uff1a\n<pre><code>pip install rfdetr\r\n<\/code><\/pre>\n<p>\u8fd9\u4f1a\u81ea\u52a8\u5b89\u88c5\u00a0<code>numpy<\/code>\u3001<code>supervision<\/code>\u00a0\u7b49\u5fc5\u8981\u5e93\u3002<\/li>\n<\/ul>\n<\/li>\n<li><strong>\u9a8c\u8bc1\u5b89\u88c5<\/strong><br \/>\n\u8fd0\u884c\u4ee5\u4e0b\u4ee3\u7801\uff1a<\/p>\n<pre><code>from rfdetr import RFDETRBase\r\nprint(\"\u5b89\u88c5\u6210\u529f\")<\/code><\/pre>\n<\/li>\n<\/ol>\n<p>\u5982\u679c\u6ca1\u6709\u62a5\u9519\uff0c\u5b89\u88c5\u5c31\u5b8c\u6210\u4e86\u3002<\/p>\n<h3>\u63a8\u7406\u64cd\u4f5c<\/h3>\n<p>RF-DETR \u81ea\u5e26 COCO \u6570\u636e\u96c6\u7684\u9884\u8bad\u7ec3\u6a21\u578b\uff0c\u53ef\u4ee5\u76f4\u63a5\u68c0\u6d4b\u56fe\u50cf\u6216\u89c6\u9891\u3002<\/p>\n<ol>\n<li><strong>\u56fe\u50cf\u68c0\u6d4b<\/strong>\n<ul>\n<li>\u793a\u4f8b\u4ee3\u7801\uff1a\n<pre><code>import io\r\nimport requests\r\nfrom PIL import Image\r\nfrom rfdetr import RFDETRBase\r\nimport supervision as sv\r\nmodel = RFDETRBase()\r\nurl = \"https:\/\/media.roboflow.com\/notebooks\/examples\/dog-2.jpeg\"\r\nimage = Image.open(io.BytesIO(requests.get(url).content))\r\ndetections = model.predict(image, threshold=0.5)\r\nlabels = [f\"{class_id} {confidence:.2f}\" for class_id, confidence in zip(detections.class_id, detections.confidence)]\r\nannotated_image = image.copy()\r\nannotated_image = sv.BoxAnnotator().annotate(annotated_image, detections)\r\nannotated_image = sv.LabelAnnotator().annotate(annotated_image, detections, labels)\r\nsv.plot_image(annotated_image)\r\n<\/code><\/pre>\n<\/li>\n<li>\u8fd9\u6bb5\u4ee3\u7801\u4f1a\u68c0\u6d4b\u56fe\u7247\u4e2d\u7684\u7269\u4f53\uff0c\u6807\u6ce8\u8fb9\u754c\u6846\u548c\u7f6e\u4fe1\u5ea6\uff0c\u7136\u540e\u663e\u793a\u7ed3\u679c\u3002<\/li>\n<\/ul>\n<\/li>\n<li><strong>\u89c6\u9891\u68c0\u6d4b<\/strong>\n<ul>\n<li>\u5148\u5b89\u88c5\u00a0<code>opencv-python<\/code>\uff1a\n<pre><code>pip install opencv-python\r\n<\/code><\/pre>\n<\/li>\n<li>\u793a\u4f8b\u4ee3\u7801\uff1a\n<pre><code>import cv2\r\nfrom rfdetr import RFDETRBase\r\nimport supervision as sv\r\nmodel = RFDETRBase()\r\ncap = cv2.VideoCapture(\"video.mp4\")  # \u66ff\u6362\u4e3a\u4f60\u7684\u89c6\u9891\u8def\u5f84\r\nwhile cap.isOpened():\r\nret, frame = cap.read()\r\nif not ret:\r\nbreak\r\nimage = Image.fromarray(cv2.cvtColor(frame, cv2.COLOR_BGR2RGB))\r\ndetections = model.predict(image, threshold=0.5)\r\nannotated_frame = sv.BoxAnnotator().annotate(frame, detections)\r\ncv2.imshow(\"RF-DETR Detection\", annotated_frame)\r\nif cv2.waitKey(1) &amp; 0xFF == ord('q'):\r\nbreak\r\ncap.release()\r\ncv2.destroyAllWindows()\r\n<\/code><\/pre>\n<\/li>\n<li>\u8fd9\u4f1a\u9010\u5e27\u68c0\u6d4b\u89c6\u9891\u4e2d\u7684\u7269\u4f53\u5e76\u5b9e\u65f6\u663e\u793a\u3002<\/li>\n<\/ul>\n<\/li>\n<li><strong>\u8c03\u6574\u5206\u8fa8\u7387<\/strong>\n<ul>\n<li>\u521d\u59cb\u5316\u65f6\u53ef\u8bbe\u7f6e\u5206\u8fa8\u7387\uff08\u5fc5\u987b\u662f 56 \u7684\u500d\u6570\uff09\uff1a\n<pre><code>model = RFDETRBase(resolution=560)\r\n<\/code><\/pre>\n<\/li>\n<li>\u5206\u8fa8\u7387\u8d8a\u9ad8\uff0c\u7cbe\u5ea6\u8d8a\u597d\uff0c\u4f46\u901f\u5ea6\u4f1a\u53d8\u6162\u3002<\/li>\n<\/ul>\n<\/li>\n<\/ol>\n<h3>\u8bad\u7ec3\u81ea\u5b9a\u4e49\u6a21\u578b<\/h3>\n<p>RF-DETR \u652f\u6301\u7528\u81ea\u5df1\u7684\u6570\u636e\u96c6\u5fae\u8c03\uff0c\u4f46\u6570\u636e\u96c6\u9700\u662f COCO \u683c\u5f0f\uff0c\u5305\u542b\u00a0<code>train<\/code>\u3001<code>valid<\/code>\u00a0\u548c\u00a0<code>test<\/code>\u00a0\u4e09\u4e2a\u5b50\u76ee\u5f55\u3002<\/p>\n<ol>\n<li><strong>\u51c6\u5907\u6570\u636e\u96c6<\/strong>\n<ul>\n<li>\u793a\u4f8b\u76ee\u5f55\u7ed3\u6784\uff1a\n<pre><code>dataset\/\r\n\u251c\u2500\u2500 train\/\r\n\u2502   \u251c\u2500\u2500 _annotations.coco.json\r\n\u2502   \u251c\u2500\u2500 image1.jpg\r\n\u2502   \u2514\u2500\u2500 image2.jpg\r\n\u251c\u2500\u2500 valid\/\r\n\u2502   \u251c\u2500\u2500 _annotations.coco.json\r\n\u2502   \u251c\u2500\u2500 image1.jpg\r\n\u2502   \u2514\u2500\u2500 image2.jpg\r\n\u2514\u2500\u2500 test\/\r\n\u251c\u2500\u2500 _annotations.coco.json\r\n\u251c\u2500\u2500 image1.jpg\r\n\u2514\u2500\u2500 image2.jpg\r\n<\/code><\/pre>\n<\/li>\n<li>\u53ef\u4ee5\u7528 Roboflow \u5e73\u53f0\u751f\u6210 COCO \u683c\u5f0f\u6570\u636e\u96c6\uff1a\n<pre><code>from roboflow import Roboflow\r\nrf = Roboflow(api_key=\"\u4f60\u7684API\u5bc6\u94a5\")\r\nproject = rf.workspace(\"rf-100-vl\").project(\"mahjong-vtacs-mexax-m4vyu-sjtd\")\r\ndataset = project.version(2).download(\"coco\")\r\n<\/code><\/pre>\n<\/li>\n<\/ul>\n<\/li>\n<li><strong>\u5f00\u59cb\u8bad\u7ec3<\/strong>\n<ul>\n<li>\u793a\u4f8b\u4ee3\u7801\uff1a\n<pre><code>from rfdetr import RFDETRBase\r\nmodel = RFDETRBase()\r\nmodel.train(dataset_dir=\".\/mahjong-vtacs-mexax-m4vyu-sjtd-2\", epochs=10, batch_size=4, grad_accum_steps=4, lr=1e-4)\r\n<\/code><\/pre>\n<\/li>\n<li>\u8bad\u7ec3\u65f6\uff0c\u63a8\u8350\u603b\u6279\u6b21\u5927\u5c0f\uff08<code>batch_size * grad_accum_steps<\/code>\uff09\u4e3a 16\u3002\u4f8b\u5982\uff0cA100 GPU \u7528\u00a0<code>batch_size=16, grad_accum_steps=1<\/code>\uff1bT4 GPU \u7528\u00a0<code>batch_size=4, grad_accum_steps=4<\/code>\u3002<\/li>\n<\/ul>\n<\/li>\n<li><strong>\u591a GPU \u8bad\u7ec3<\/strong>\n<ul>\n<li>\u521b\u5efa\u00a0<code>main.py<\/code>\u00a0\u6587\u4ef6\uff1a\n<pre><code>from rfdetr import RFDETRBase\r\nmodel = RFDETRBase()\r\nmodel.train(dataset_dir=\".\/dataset\", epochs=10, batch_size=4, grad_accum_steps=4, lr=1e-4)\r\n<\/code><\/pre>\n<\/li>\n<li>\u5728\u7ec8\u7aef\u8fd0\u884c\uff1a\n<pre><code>python -m torch.distributed.launch --nproc_per_node=8 --use_env main.py\r\n<\/code><\/pre>\n<\/li>\n<li>\u5c06\u00a0<code>8<\/code>\u00a0\u66ff\u6362\u4e3a\u4f60\u4f7f\u7528\u7684 GPU \u6570\u91cf\u3002\u6ce8\u610f\u8c03\u6574\u00a0<code>batch_size<\/code>\u00a0\u4ee5\u4fdd\u6301\u603b\u6279\u6b21\u5927\u5c0f\u7a33\u5b9a\u3002<\/li>\n<\/ul>\n<\/li>\n<li><strong>\u52a0\u8f7d\u8bad\u7ec3\u7ed3\u679c<\/strong>\n<ul>\n<li>\u8bad\u7ec3\u540e\u4f1a\u751f\u6210\u4e24\u4e2a\u6743\u91cd\u6587\u4ef6\uff1a\u5e38\u89c4\u6743\u91cd\u548c EMA \u6743\u91cd\uff08\u66f4\u7a33\u5b9a\uff09\u3002\u52a0\u8f7d\u65b9\u5f0f\uff1a\n<pre><code>model = RFDETRBase(pretrain_weights=\".\/output\/model_ema.pt\")\r\ndetections = model.predict(\"image.jpg\")\r\n<\/code><\/pre>\n<\/li>\n<\/ul>\n<\/li>\n<\/ol>\n<h3>ONNX \u5bfc\u51fa<\/h3>\n<ul>\n<li>\u5bfc\u51fa\u4e3a ONNX \u683c\u5f0f\uff0c\u65b9\u4fbf\u5728\u5176\u4ed6\u5e73\u53f0\u90e8\u7f72\uff1a\n<pre><code>from rfdetr import RFDETRBase\r\nmodel = RFDETRBase()\r\nmodel.export()\r\n<\/code><\/pre>\n<\/li>\n<li>\u5bfc\u51fa\u7684\u6587\u4ef6\u4f1a\u4fdd\u5b58\u5728\u00a0<code>output<\/code>\u00a0\u76ee\u5f55\uff0c\u9002\u5408\u8fb9\u7f18\u8bbe\u5907\u4f18\u5316\u63a8\u7406\u3002<\/li>\n<\/ul>\n<p>&nbsp;<\/p>\n<h2>\u5e94\u7528\u573a\u666f<\/h2>\n<ol>\n<li><strong>\u81ea\u52a8\u9a7e\u9a76<\/strong><br \/>\nRF-DETR \u80fd\u5b9e\u65f6\u68c0\u6d4b\u9053\u8def\u4e0a\u7684\u8f66\u8f86\u548c\u884c\u4eba\u3002\u5b83\u7684\u4f4e\u5ef6\u8fdf\u548c\u9ad8\u7cbe\u5ea6\u9002\u5408\u5d4c\u5165\u5f0f\u7cfb\u7edf\u4f7f\u7528\u3002<\/li>\n<li><strong>\u5de5\u4e1a\u8d28\u68c0<\/strong><br \/>\n\u5728\u5de5\u5382\u6d41\u6c34\u7ebf\u4e0a\uff0cRF-DETR \u53ef\u4ee5\u5feb\u901f\u8bc6\u522b\u96f6\u4ef6\u7f3a\u9677\u3002\u6a21\u578b\u8f7b\u91cf\uff0c\u80fd\u76f4\u63a5\u5728\u8bbe\u5907\u4e0a\u8fd0\u884c\u3002<\/li>\n<li><strong>\u89c6\u9891\u76d1\u63a7<\/strong><br \/>\nRF-DETR \u5904\u7406\u76d1\u63a7\u89c6\u9891\u65f6\uff0c\u80fd\u5b9e\u65f6\u68c0\u6d4b\u5f02\u5e38\u7269\u4f53\u6216\u884c\u4e3a\u3002\u5b83\u652f\u6301\u89c6\u9891\u6d41\uff0c\u9002\u5408\u5168\u5929\u5019\u5b89\u9632\u3002<\/li>\n<\/ol>\n<p>&nbsp;<\/p>\n<h2>QA<\/h2>\n<ol>\n<li><strong>\u652f\u6301\u54ea\u4e9b\u6570\u636e\u96c6\u683c\u5f0f\uff1f<\/strong><br \/>\n\u53ea\u652f\u6301 COCO \u683c\u5f0f\u3002\u6570\u636e\u96c6\u9700\u5305\u542b\u00a0<code>train<\/code>\u3001<code>valid<\/code>\u00a0\u548c\u00a0<code>test<\/code>\u00a0\u5b50\u76ee\u5f55\uff0c\u6bcf\u4e2a\u76ee\u5f55\u6709\u5bf9\u5e94\u7684\u00a0<code>_annotations.coco.json<\/code>\u00a0\u6587\u4ef6\u3002<\/li>\n<li><strong>\u5982\u4f55\u83b7\u53d6 Roboflow API \u5bc6\u94a5\uff1f<\/strong><br \/>\n\u767b\u5f55\u00a0https:\/\/app.roboflow.com\uff0c\u5728\u8d26\u6237\u8bbe\u7f6e\u4e2d\u627e\u5230 API \u5bc6\u94a5\uff0c\u590d\u5236\u540e\u8bbe\u7f6e\u5230\u73af\u5883\u53d8\u91cf\u00a0<code>ROBOFLOW_API_KEY<\/code>\u3002<\/li>\n<li><strong>\u8bad\u7ec3\u9700\u8981\u591a\u957f\u65f6\u95f4\uff1f<\/strong><br \/>\n\u53d6\u51b3\u4e8e\u786c\u4ef6\u548c\u6570\u636e\u96c6\u5927\u5c0f\u3002\u5728 T4 GPU \u4e0a\uff0c10 \u4e2a epoch \u53ef\u80fd\u9700\u8981\u51e0\u5c0f\u65f6\u3002\u5c0f\u6570\u636e\u96c6\u7528 CPU \u4e5f\u80fd\u8dd1\uff0c\u4f46\u901f\u5ea6\u6162\u3002<\/li>\n<\/ol>\n","protected":false},"excerpt":{"rendered":"<p>RF-DETR \u662f Roboflow \u56e2\u961f\u5f00\u53d1\u7684\u4e00\u4e2a\u5f00\u6e90\u5bf9\u8c61\u68c0\u6d4b\u6a21\u578b\u3002\u5b83\u57fa\u4e8e Transformer \u67b6\u6784\uff0c\u6838\u5fc3\u7279\u70b9\u662f\u5b9e\u65f6\u9ad8\u6548\u3002\u6a21\u578b\u5728\u5fae\u8f6f COCO \u6570\u636e\u96c6\u4e0a\u9996\u6b21\u5b9e\u73b0\u8d85\u8fc7 60 AP \u7684\u5b9e\u65f6\u68c0\u6d4b\uff0c\u540c\u65f6\u5728 RF100-VL \u57fa\u51c6\u6d4b\u8bd5\u4e2d\u8868\u73b0\u7a81\u51fa&#8230;<\/p>\n","protected":false},"author":1,"featured_media":62106,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[20],"tags":[230,378],"class_list":["post-29355","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-tool","tag-aikaiyuanxiangmu","tag-shijuemubiaojiance"],"_links":{"self":[{"href":"https:\/\/www.kdjingpai.com\/ja\/wp-json\/wp\/v2\/posts\/29355","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=29355"}],"version-history":[{"count":0,"href":"https:\/\/www.kdjingpai.com\/ja\/wp-json\/wp\/v2\/posts\/29355\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.kdjingpai.com\/ja\/wp-json\/wp\/v2\/media\/62106"}],"wp:attachment":[{"href":"https:\/\/www.kdjingpai.com\/ja\/wp-json\/wp\/v2\/media?parent=29355"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.kdjingpai.com\/ja\/wp-json\/wp\/v2\/categories?post=29355"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.kdjingpai.com\/ja\/wp-json\/wp\/v2\/tags?post=29355"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}