{"id":29061,"date":"2025-03-18T23:02:34","date_gmt":"2025-03-18T15:02:34","guid":{"rendered":"https:\/\/www.aisharenet.com\/?p=29061"},"modified":"2025-03-18T23:02:34","modified_gmt":"2025-03-18T15:02:34","slug":"humanomni","status":"publish","type":"post","link":"https:\/\/www.kdjingpai.com\/de\/humanomni\/","title":{"rendered":"HumanOmni\uff1a\u5206\u6790\u4eba\u7c7b\u89c6\u9891\u60c5\u611f\u548c\u52a8\u4f5c\u7684\u591a\u6a21\u6001\u5927\u6a21\u578b"},"content":{"rendered":"<p>HumanOmni \u662f\u7531 HumanMLLM \u56e2\u961f\u5f00\u53d1\u7684\u4e00\u4e2a\u5f00\u6e90\u591a\u6a21\u6001\u5927\u6a21\u578b\uff0c\u6258\u7ba1\u5728 GitHub \u4e0a\u3002\u5b83\u4e13\u6ce8\u4e8e\u5206\u6790\u4eba\u7c7b\u89c6\u9891\uff0c\u80fd\u540c\u65f6\u5904\u7406\u753b\u9762\u548c\u58f0\u97f3\uff0c\u5e2e\u52a9\u7406\u89e3\u60c5\u611f\u3001\u52a8\u4f5c\u548c\u5bf9\u8bdd\u5185\u5bb9\u3002\u9879\u76ee\u7528\u4e86 240 \u4e07\u4e2a\u4ee5\u4eba\u4e3a\u4e2d\u5fc3\u7684\u89c6\u9891\u7247\u6bb5\u548c 1400 \u4e07\u6761\u6307\u4ee4\u6570\u636e\u8fdb\u884c\u9884\u8bad\u7ec3\uff0c\u8fd8\u7528 5 \u4e07\u4e2a\u624b\u5de5\u6807\u6ce8\u7684\u89c6\u9891\u7247\u6bb5\uff08\u542b 10 \u4e07\u591a\u6761\u6307\u4ee4\uff09\u8fdb\u884c\u5fae\u8c03\u3002HumanOmni \u5206\u4e09\u4e2a\u5206\u652f\u5904\u7406\u9762\u90e8\u3001\u8eab\u4f53\u548c\u4e92\u52a8\u573a\u666f\uff0c\u80fd\u6839\u636e\u8f93\u5165\u52a8\u6001\u8c03\u6574\u878d\u5408\u65b9\u5f0f\u3002\u5b83\u662f\u4e1a\u754c\u9996\u4e2a\u4ee5\u4eba\u7c7b\u4e3a\u4e2d\u5fc3\u7684\u591a\u6a21\u6001\u6a21\u578b\uff0c\u6027\u80fd\u8d85\u8fc7\u8bb8\u591a\u540c\u7c7b\u6a21\u578b\u3002\u56e2\u961f\u8fd8\u63a8\u51fa\u4e86\u57fa\u4e8e\u5b83\u7684 R1-Omni\uff0c\u9996\u6b21\u7ed3\u5408\u5f3a\u5316\u5b66\u4e60\u63d0\u5347\u63a8\u7406\u80fd\u529b\u3002\u4ee3\u7801\u548c\u90e8\u5206\u6570\u636e\u96c6\u90fd\u5f00\u653e\uff0c\u65b9\u4fbf\u7814\u7a76\u8005\u548c\u5f00\u53d1\u8005\u4f7f\u7528\u3002<\/p>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"aligncenter size-full wp-image-29063\" title=\"HumanOmni\uff1a\u5206\u6790\u4eba\u7c7b\u89c6\u9891\u60c5\u611f\u548c\u52a8\u4f5c\u7684\u591a\u6a21\u6001\u5927\u6a21\u578b-1\" src=\"https:\/\/www.kdjingpai.com\/wp-content\/uploads\/2025\/03\/2290871387ab214.png\" alt=\"HumanOmni\uff1a\u5206\u6790\u4eba\u7c7b\u89c6\u9891\u60c5\u611f\u548c\u52a8\u4f5c\u7684\u591a\u6a21\u6001\u5927\u6a21\u578b-1\" width=\"1311\" height=\"531\" srcset=\"https:\/\/www.kdjingpai.com\/wp-content\/uploads\/2025\/03\/2290871387ab214.png 1311w, https:\/\/www.kdjingpai.com\/wp-content\/uploads\/2025\/03\/2290871387ab214-768x311.png 768w, https:\/\/www.kdjingpai.com\/wp-content\/uploads\/2025\/03\/2290871387ab214-18x7.png 18w\" sizes=\"auto, (max-width: 1311px) 100vw, 1311px\" \/><\/p>\n<p>&nbsp;<\/p>\n<h2>\u529f\u80fd\u5217\u8868<\/h2>\n<ul>\n<li><strong>\u60c5\u611f\u8bc6\u522b<\/strong>: \u5206\u6790\u89c6\u9891\u4e2d\u7684\u9762\u90e8\u8868\u60c5\u548c\u58f0\u97f3\u8bed\u8c03\uff0c\u5224\u65ad\u4eba\u7269\u60c5\u7eea\uff0c\u5982\u5f00\u5fc3\u3001\u751f\u6c14\u6216\u60b2\u4f24\u3002<\/li>\n<li><strong>\u9762\u90e8\u8868\u60c5\u63cf\u8ff0<\/strong>: \u8bc6\u522b\u5e76\u63cf\u8ff0\u4eba\u7269\u9762\u90e8\u7ec6\u8282\uff0c\u6bd4\u5982\u5fae\u7b11\u6216\u76b1\u7709\u3002<\/li>\n<li><strong>\u52a8\u4f5c\u7406\u89e3<\/strong>: \u5206\u6790\u89c6\u9891\u4e2d\u4eba\u7269\u7684\u52a8\u4f5c\uff0c\u63cf\u8ff0\u4ed6\u4eec\u5728\u505a\u4ec0\u4e48\uff0c\u6bd4\u5982\u8d70\u8def\u6216\u6325\u624b\u3002<\/li>\n<li><strong>\u8bed\u97f3\u5904\u7406<\/strong>: \u4ece\u97f3\u9891\u4e2d\u63d0\u53d6\u5185\u5bb9\uff0c\u652f\u6301\u8bed\u97f3\u8bc6\u522b\u548c\u8bed\u8c03\u5206\u6790\u3002<\/li>\n<li><strong>\u591a\u6a21\u6001\u878d\u5408<\/strong>: \u7ed3\u5408\u753b\u9762\u548c\u58f0\u97f3\uff0c\u7406\u89e3\u590d\u6742\u573a\u666f\uff0c\u63d0\u4f9b\u66f4\u51c6\u786e\u7684\u5206\u6790\u3002<\/li>\n<li><strong>\u52a8\u6001\u5206\u652f\u8c03\u6574<\/strong>: \u7528\u9762\u90e8\u3001\u8eab\u4f53\u3001\u4e92\u52a8\u4e09\u4e2a\u5206\u652f\u5904\u7406\u4e0d\u540c\u573a\u666f\uff0c\u81ea\u52a8\u8c03\u6574\u6743\u91cd\u3002<\/li>\n<li><strong>\u5f00\u6e90\u652f\u6301<\/strong>: \u63d0\u4f9b\u4ee3\u7801\u3001\u9884\u8bad\u7ec3\u6a21\u578b\u548c\u90e8\u5206\u6570\u636e\u96c6\uff0c\u652f\u6301\u4e8c\u6b21\u5f00\u53d1\u3002<\/li>\n<\/ul>\n<p>&nbsp;<\/p>\n<h2>\u4f7f\u7528\u5e2e\u52a9<\/h2>\n<p>HumanOmni \u9002\u5408\u6709\u6280\u672f\u57fa\u7840\u7684\u7528\u6237\uff0c\u6bd4\u5982\u5f00\u53d1\u8005\u6216\u7814\u7a76\u8005\u3002\u4ee5\u4e0b\u662f\u5b89\u88c5\u548c\u4f7f\u7528\u6b65\u9aa4\uff0c\u8be6\u7ec6\u5230\u53ef\u4ee5\u76f4\u63a5\u4e0a\u624b\u3002<\/p>\n<h3>\u5b89\u88c5\u6d41\u7a0b<\/h3>\n<p>\u8981\u8fd0\u884c HumanOmni\uff0c\u9700\u8981\u5148\u51c6\u5907\u597d\u73af\u5883\u3002\u4ee5\u4e0b\u662f\u5177\u4f53\u6b65\u9aa4\uff1a<\/p>\n<ol>\n<li><strong>\u68c0\u67e5\u786c\u4ef6\u548c\u8f6f\u4ef6\u8981\u6c42<\/strong>\n<ul>\n<li>\u64cd\u4f5c\u7cfb\u7edf\uff1a\u652f\u6301 Linux\u3001Windows \u6216 macOS\u3002<\/li>\n<li>Python\uff1a\u9700\u8981 3.10 \u6216\u66f4\u9ad8\u7248\u672c\u3002<\/li>\n<li>CUDA\uff1a\u5efa\u8bae 12.1 \u6216\u66f4\u9ad8\uff08\u82e5\u7528 GPU\uff09\u3002<\/li>\n<li>PyTorch\uff1a\u9700\u8981 2.2 \u6216\u66f4\u9ad8\u7248\u672c\uff0c\u652f\u6301 CUDA\u3002<\/li>\n<li>\u786c\u4ef6\uff1a\u63a8\u8350 NVIDIA GPU\uff0cCPU \u4e5f\u80fd\u7528\u4f46\u901f\u5ea6\u6162\u3002<\/li>\n<\/ul>\n<\/li>\n<li><strong>\u4e0b\u8f7d\u4ee3\u7801<\/strong><br \/>\n\u6253\u5f00\u7ec8\u7aef\uff0c\u8f93\u5165\u547d\u4ee4\u4e0b\u8f7d\u9879\u76ee\uff1a<\/li>\n<\/ol>\n<pre><code>git clone https:\/\/github.com\/HumanMLLM\/HumanOmni.git\r\ncd HumanOmni\r\n<\/code><\/pre>\n<ol start=\"3\">\n<li><strong>\u521b\u5efa\u865a\u62df\u73af\u5883<\/strong><br \/>\n\u7528 Conda \u521b\u5efa\u72ec\u7acb\u73af\u5883\uff0c\u907f\u514d\u51b2\u7a81\uff1a<\/li>\n<\/ol>\n<pre><code>conda create -n humanOmni python=3.10 -y\r\nconda activate humanOmni\r\n<\/code><\/pre>\n<ol start=\"4\">\n<li><strong>\u5b89\u88c5\u4f9d\u8d56<\/strong><br \/>\n\u9879\u76ee\u6709\u4e2a\u00a0<code>requirements.txt<\/code>\u00a0\u6587\u4ef6\uff0c\u5217\u51fa\u4e86\u6240\u9700\u5e93\u3002\u8fd0\u884c\u4ee5\u4e0b\u547d\u4ee4\u5b89\u88c5\uff1a<\/li>\n<\/ol>\n<pre><code>pip install --upgrade pip\r\npip install -r requirements.txt\r\npip install flash-attn --no-build-isolation\r\n<\/code><\/pre>\n<ol start=\"5\">\n<li><strong>\u4e0b\u8f7d\u6a21\u578b\u6743\u91cd<\/strong><br \/>\nHumanOmni \u6709\u4e09\u79cd\u6a21\u578b\uff1a<\/li>\n<\/ol>\n<ul>\n<li><code>HumanOmni-Video<\/code>\uff1a\u5904\u7406\u89c6\u9891\uff0c7B \u53c2\u6570\u3002<\/li>\n<li><code>HumanOmni-Audio<\/code>\uff1a\u5904\u7406\u97f3\u9891\uff0c7B \u53c2\u6570\u3002<\/li>\n<li><code>HumanOmni-Omni<\/code>\uff1a\u878d\u5408\u89c6\u9891\u548c\u97f3\u9891\uff0c7B \u53c2\u6570\uff08\u7b80\u79f0 HumanOmni\uff09\u3002<br \/>\n\u4ece Hugging Face \u6216 ModelScope \u4e0b\u8f7d\uff0c\u6bd4\u5982\uff1a<\/li>\n<li><a href=\"https:\/\/hf.co\/StarJiaxing\/HumanOmni-7B\">HumanOmni-7B<\/a><\/li>\n<li><a href=\"https:\/\/modelscope.cn\/models\/iic\/HumanOmni-7B-Video\">HumanOmni-7B-Video<\/a><br \/>\n\u4e0b\u8f7d\u540e\u653e\u5230\u9879\u76ee\u6587\u4ef6\u5939\u3002<\/li>\n<\/ul>\n<ol start=\"6\">\n<li><strong>\u9a8c\u8bc1\u5b89\u88c5<\/strong><br \/>\n\u7528\u6d4b\u8bd5\u547d\u4ee4\u68c0\u67e5\u73af\u5883\uff1a<\/li>\n<\/ol>\n<pre><code>python inference.py --modal video --model_path .\/HumanOmni_7B --video_path test.mp4 --instruct \"Describe this video.\"\r\n<\/code><\/pre>\n<p>\u5982\u679c\u8f93\u51fa\u89c6\u9891\u63cf\u8ff0\uff0c\u5b89\u88c5\u5c31\u6210\u529f\u4e86\u3002<\/p>\n<h3>\u529f\u80fd\u64cd\u4f5c\u6d41\u7a0b<\/h3>\n<p>HumanOmni \u7684\u6838\u5fc3\u662f\u5206\u6790\u89c6\u9891\u548c\u97f3\u9891\u3002\u4ee5\u4e0b\u662f\u4e3b\u8981\u529f\u80fd\u7684\u8be6\u7ec6\u64cd\u4f5c\u3002<\/p>\n<h4>1. \u60c5\u611f\u8bc6\u522b<\/h4>\n<ul>\n<li><strong>\u6b65\u9aa4<\/strong><\/li>\n<li>\u51c6\u5907\u4e00\u4e2a\u542b\u4eba\u7269\u7684\u89c6\u9891\uff08\u6bd4\u5982\u00a0<code>sample.mp4<\/code>\uff09\u3002<\/li>\n<li>\u8fd0\u884c\u547d\u4ee4\uff1a<\/li>\n<\/ul>\n<pre><code>python inference.py --modal video_audio --model_path .\/HumanOmni_7B --video_path sample.mp4 --instruct \"Which emotion is most obvious?\"\r\n<\/code><\/pre>\n<ul>\n<li>\u6a21\u578b\u4f1a\u8f93\u51fa\u60c5\u7eea\uff0c\u6bd4\u5982 \u201cangry\u201d \u6216 \u201chappy\u201d\u3002<\/li>\n<li><strong>\u6ce8\u610f<\/strong><\/li>\n<li>\u89c6\u9891\u8981\u6e05\u6670\uff0c\u4eba\u7269\u8868\u60c5\u548c\u58f0\u97f3\u9700\u53ef\u8fa8\u8bc6\u3002<\/li>\n<li>\u957f\u89c6\u9891\u53ef\u80fd\u9700\u8981\u66f4\u591a\u8ba1\u7b97\u65f6\u95f4\u3002<\/li>\n<\/ul>\n<h4>2. \u9762\u90e8\u8868\u60c5\u63cf\u8ff0<\/h4>\n<ul>\n<li><strong>\u6b65\u9aa4<\/strong><\/li>\n<li>\u8f93\u5165\u89c6\u9891\uff0c\u8fd0\u884c\uff1a<\/li>\n<\/ul>\n<pre><code>python inference.py --modal video --model_path .\/HumanOmni_7B --video_path sample.mp4 --instruct \"What\u2019s the major facial expression?\"\r\n<\/code><\/pre>\n<ul>\n<li>\u8f93\u51fa\u53ef\u80fd\u662f \u201csmile\u201d \u6216 \u201cfrown\u201d\uff0c\u5e26\u7b80\u5355\u63cf\u8ff0\u3002<\/li>\n<li><strong>\u5efa\u8bae<\/strong><\/li>\n<li>\u7528 10-30 \u79d2\u7684\u77ed\u89c6\u9891\u6d4b\u8bd5\u6548\u679c\u66f4\u597d\u3002<\/li>\n<\/ul>\n<h4>3. \u52a8\u4f5c\u7406\u89e3<\/h4>\n<ul>\n<li><strong>\u6b65\u9aa4<\/strong><\/li>\n<li>\u8f93\u5165\u89c6\u9891\uff0c\u8fd0\u884c\uff1a<\/li>\n<\/ul>\n<pre><code>python inference.py --modal video --model_path .\/HumanOmni_7B --video_path sample.mp4 --instruct \"Describe the major action in detail.\"\r\n<\/code><\/pre>\n<ul>\n<li>\u8f93\u51fa\u52a8\u4f5c\u63cf\u8ff0\uff0c\u6bd4\u5982 \u201ca person is walking\u201d\u3002<\/li>\n<li><strong>\u6280\u5de7<\/strong><\/li>\n<li>\u786e\u4fdd\u52a8\u4f5c\u660e\u663e\uff0c\u907f\u514d\u80cc\u666f\u6742\u4e71\u3002<\/li>\n<\/ul>\n<h4>4. \u8bed\u97f3\u5904\u7406<\/h4>\n<ul>\n<li><strong>\u6b65\u9aa4<\/strong><\/li>\n<li>\u8f93\u5165\u542b\u97f3\u9891\u7684\u89c6\u9891\uff0c\u8fd0\u884c\uff1a<\/li>\n<\/ul>\n<pre><code>python inference.py --modal audio --model_path .\/HumanOmni_7B --video_path sample.mp4 --instruct \"What did the person say?\"\r\n<\/code><\/pre>\n<ul>\n<li>\u8f93\u51fa\u8bed\u97f3\u5185\u5bb9\uff0c\u6bd4\u5982 \u201cDogs are sitting by the door\u201d\u3002<\/li>\n<li><strong>\u6ce8\u610f<\/strong><\/li>\n<li>\u97f3\u9891\u8981\u6e05\u695a\uff0c\u65e0\u6742\u97f3\u6548\u679c\u6700\u4f73\u3002<\/li>\n<\/ul>\n<h4>5. \u591a\u6a21\u6001\u878d\u5408<\/h4>\n<ul>\n<li><strong>\u6b65\u9aa4<\/strong><\/li>\n<li>\u8f93\u5165\u89c6\u9891\u548c\u97f3\u9891\uff0c\u8fd0\u884c\uff1a<\/li>\n<\/ul>\n<pre><code>python inference.py --modal video_audio --model_path .\/HumanOmni_7B --video_path sample.mp4 --instruct \"Describe this video.\"\r\n<\/code><\/pre>\n<ul>\n<li>\u6a21\u578b\u4f1a\u7ed3\u5408\u753b\u9762\u548c\u58f0\u97f3\uff0c\u7ed9\u51fa\u5b8c\u6574\u63cf\u8ff0\u3002<\/li>\n<li><strong>\u4f18\u52bf<\/strong><\/li>\n<li>\u80fd\u6355\u6349\u60c5\u7eea\u548c\u52a8\u4f5c\u7684\u5173\u8054\uff0c\u5206\u6790\u66f4\u5168\u9762\u3002<\/li>\n<\/ul>\n<h4>6. \u81ea\u5b9a\u4e49\u6570\u636e\u96c6\u8bad\u7ec3<\/h4>\n<ul>\n<li><strong>\u6b65\u9aa4<\/strong><\/li>\n<li>\u51c6\u5907 JSON \u683c\u5f0f\u7684\u6570\u636e\u6587\u4ef6\uff0c\u5305\u542b\u89c6\u9891\u8def\u5f84\u548c\u6307\u4ee4\u5bf9\u8bdd\u3002\u6bd4\u5982\uff1a<\/li>\n<\/ul>\n<pre><code>[\r\n{\r\n\"video\": \"path\/to\/video.mp4\",\r\n\"conversations\": [\r\n{\"from\": \"human\", \"value\": \"What\u2019s the emotion?\"},\r\n{\"from\": \"gpt\", \"value\": \"sad\"}\r\n]\r\n}\r\n]\r\n<\/code><\/pre>\n<ul>\n<li>\u4e0b\u8f7d\u00a0<code>HumanOmni-7B-Video<\/code>\u00a0\u548c\u00a0<code>HumanOmni-7B-Audio<\/code>\u00a0\u6743\u91cd\u3002<\/li>\n<li>\u8fd0\u884c\u8bad\u7ec3\u811a\u672c\uff1a<\/li>\n<\/ul>\n<pre><code>bash scripts\/train\/finetune_humanomni.sh\r\n<\/code><\/pre>\n<ul>\n<li><strong>\u7528\u9014<\/strong><\/li>\n<li>\u53ef\u4ee5\u7528\u81ea\u5df1\u7684\u89c6\u9891\u6570\u636e\u4f18\u5316\u6a21\u578b\u3002<\/li>\n<\/ul>\n<h3>\u5e38\u89c1\u95ee\u9898\u89e3\u51b3<\/h3>\n<ul>\n<li><strong>\u8fd0\u884c\u62a5\u9519<\/strong>\uff1a\u68c0\u67e5 Python \u548c PyTorch \u7248\u672c\u662f\u5426\u5339\u914d\u3002<\/li>\n<li><strong>\u6a21\u578b\u52a0\u8f7d\u5931\u8d25<\/strong>\uff1a\u786e\u8ba4\u8def\u5f84\u6b63\u786e\uff0c\u78c1\u76d8\u7a7a\u95f4\u591f\u7528\uff08\u6a21\u578b\u7ea6 10GB\uff09\u3002<\/li>\n<li><strong>\u7ed3\u679c\u4e0d\u51c6<\/strong>\uff1a\u6362\u6e05\u6670\u7684\u89c6\u9891\u6216\u8c03\u6574\u6307\u4ee4\u8868\u8ff0\u3002<\/li>\n<\/ul>\n<p>\u901a\u8fc7\u8fd9\u4e9b\u6b65\u9aa4\uff0c\u7528\u6237\u53ef\u4ee5\u8f7b\u677e\u5b89\u88c5\u548c\u4f7f\u7528 HumanOmni\uff0c\u4f53\u9a8c\u5b83\u7684\u5f3a\u5927\u529f\u80fd\u3002<\/p>\n<p>&nbsp;<\/p>\n<h2>\u5e94\u7528\u573a\u666f<\/h2>\n<ol>\n<li><strong>\u6559\u80b2\u7814\u7a76<\/strong><br \/>\n\u5206\u6790\u8bfe\u5802\u89c6\u9891\uff0c\u8bc6\u522b\u5b66\u751f\u7684\u60c5\u7eea\u548c\u53c2\u4e0e\u5ea6\uff0c\u5e2e\u52a9\u8001\u5e08\u8c03\u6574\u6559\u5b66\u65b9\u5f0f\u3002<\/li>\n<li><strong>\u533b\u7597\u8f85\u52a9<\/strong><br \/>\n\u901a\u8fc7\u60a3\u8005\u8868\u60c5\u548c\u8bed\u8c03\uff0c\u8f85\u52a9\u533b\u751f\u5224\u65ad\u5fc3\u7406\u72b6\u6001\uff0c\u6bd4\u5982\u7126\u8651\u6216\u6291\u90c1\u3002<\/li>\n<li><strong>\u5f71\u89c6\u5236\u4f5c<\/strong><br \/>\n\u5206\u6790\u89d2\u8272\u60c5\u611f\u548c\u52a8\u4f5c\uff0c\u751f\u6210\u5b57\u5e55\u6216\u5267\u60c5\u63cf\u8ff0\uff0c\u63d0\u5347\u521b\u4f5c\u6548\u7387\u3002<\/li>\n<li><strong>\u793e\u4ea4\u5206\u6790<\/strong><br \/>\n\u7528\u4e8e\u4f1a\u8bae\u89c6\u9891\uff0c\u7406\u89e3\u53c2\u4e0e\u8005\u7684\u60c5\u7eea\u548c\u884c\u4e3a\uff0c\u4f18\u5316\u6c9f\u901a\u6548\u679c\u3002<\/li>\n<\/ol>\n<p>&nbsp;<\/p>\n<h2>QA<\/h2>\n<ol>\n<li><strong>\u652f\u6301\u54ea\u4e9b\u6587\u4ef6\u683c\u5f0f\uff1f<\/strong><br \/>\n\u652f\u6301 MP4 \u683c\u5f0f\uff0c\u97f3\u9891\u9700\u5d4c\u5165\u89c6\u9891\u4e2d\u3002<\/li>\n<li><strong>\u9700\u8981\u8054\u7f51\u5417\uff1f<\/strong><br \/>\n\u4e0d\u9700\u8981\u3002\u4e0b\u8f7d\u4ee3\u7801\u548c\u6a21\u578b\u540e\u53ef\u79bb\u7ebf\u4f7f\u7528\u3002<\/li>\n<li><strong>\u6a21\u578b\u6027\u80fd\u5982\u4f55\uff1f<\/strong><br \/>\n\u5728\u60c5\u611f\u7406\u89e3\u4e0a\uff0cHumanOmni \u7684 DFEW \u6570\u636e UAR \u8fbe 74.86%\uff0c\u8fdc\u8d85 GPT4-O \u7684 50.57%\u3002\u52a8\u4f5c\u7406\u89e3\u5e73\u5747\u5f97\u5206 72.6\uff0c\u9ad8\u4e8e Qwen2-VL-7B \u7684 67.7\u3002<\/li>\n<li><strong>\u666e\u901a\u4eba\u80fd\u7528\u5417\uff1f<\/strong><br \/>\n\u9700\u8981\u57fa\u7840\u7f16\u7a0b\u80fd\u529b\u3002\u5982\u679c\u4e0d\u61c2\u4ee3\u7801\uff0c\u5efa\u8bae\u8bf7\u6280\u672f\u4eba\u5458\u5e2e\u5fd9\u3002<\/li>\n<\/ol>\n","protected":false},"excerpt":{"rendered":"<p>HumanOmni \u662f\u7531 HumanMLLM \u56e2\u961f\u5f00\u53d1\u7684\u4e00\u4e2a\u5f00\u6e90\u591a\u6a21\u6001\u5927\u6a21\u578b\uff0c\u6258\u7ba1\u5728 GitHub \u4e0a\u3002\u5b83\u4e13\u6ce8\u4e8e\u5206\u6790\u4eba\u7c7b\u89c6\u9891\uff0c\u80fd\u540c\u65f6\u5904\u7406\u753b\u9762\u548c\u58f0\u97f3\uff0c\u5e2e\u52a9\u7406\u89e3\u60c5\u611f\u3001\u52a8\u4f5c\u548c\u5bf9\u8bdd\u5185\u5bb9\u3002\u9879\u76ee\u7528\u4e86 240 \u4e07\u4e2a\u4ee5\u4eba\u4e3a\u4e2d\u5fc3\u7684\u89c6\u9891\u7247\u6bb5\u548c 1400 \u4e07\u6761\u6307&#8230;<\/p>\n","protected":false},"author":1,"featured_media":62072,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[20],"tags":[230,378],"class_list":["post-29061","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-tool","tag-aikaiyuanxiangmu","tag-shijuemubiaojiance"],"_links":{"self":[{"href":"https:\/\/www.kdjingpai.com\/de\/wp-json\/wp\/v2\/posts\/29061","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.kdjingpai.com\/de\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.kdjingpai.com\/de\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.kdjingpai.com\/de\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.kdjingpai.com\/de\/wp-json\/wp\/v2\/comments?post=29061"}],"version-history":[{"count":0,"href":"https:\/\/www.kdjingpai.com\/de\/wp-json\/wp\/v2\/posts\/29061\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.kdjingpai.com\/de\/wp-json\/wp\/v2\/media\/62072"}],"wp:attachment":[{"href":"https:\/\/www.kdjingpai.com\/de\/wp-json\/wp\/v2\/media?parent=29061"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.kdjingpai.com\/de\/wp-json\/wp\/v2\/categories?post=29061"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.kdjingpai.com\/de\/wp-json\/wp\/v2\/tags?post=29061"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}