{"id":29410,"date":"2025-03-27T20:06:57","date_gmt":"2025-03-27T12:06:57","guid":{"rendered":"https:\/\/www.aisharenet.com\/?p=29410"},"modified":"2025-11-18T17:06:38","modified_gmt":"2025-11-18T09:06:38","slug":"qwen25-omni","status":"publish","type":"post","link":"https:\/\/www.kdjingpai.com\/en\/qwen25-omni\/","title":{"rendered":"Qwen2.5-Omni"},"content":{"rendered":"<p>Qwen2.5-Omni \u662f\u963f\u91cc\u5df4\u5df4\u4e91 Qwen \u56e2\u961f\u5f00\u53d1\u7684\u4e00\u6b3e\u5f00\u6e90\u591a\u6a21\u6001 AI \u6a21\u578b\u3002\u5b83\u80fd\u5904\u7406\u6587\u672c\u3001\u56fe\u50cf\u3001\u97f3\u9891\u548c\u89c6\u9891\u7b49\u591a\u79cd\u8f93\u5165\uff0c\u5e76\u5b9e\u65f6\u751f\u6210\u6587\u672c\u6216\u81ea\u7136\u8bed\u97f3\u54cd\u5e94\u3002\u8fd9\u6b3e\u6a21\u578b\u4e8e 2025 \u5e74 3 \u6708 26 \u65e5\u53d1\u5e03\uff0c\u4ee3\u7801\u548c\u6a21\u578b\u6587\u4ef6\u6258\u7ba1\u5728 GitHub \u4e0a\uff0c\u514d\u8d39\u4f9b\u6240\u6709\u4eba\u4e0b\u8f7d\u4f7f\u7528\u3002\u5b83\u91c7\u7528 Thinker-Talker \u67b6\u6784\u548c TMRoPE \u6280\u672f\uff0c\u786e\u4fdd\u591a\u6a21\u6001\u6570\u636e\u7684\u9ad8\u6548\u5904\u7406\u3002Qwen2.5-Omni \u5728\u8bed\u97f3\u8bc6\u522b\u3001\u56fe\u50cf\u7406\u89e3\u548c\u89c6\u9891\u5206\u6790\u7b49\u4efb\u52a1\u4e2d\u8868\u73b0\u4f18\u5f02\uff0c\u9002\u7528\u4e8e\u667a\u80fd\u52a9\u624b\u3001\u591a\u5a92\u4f53\u5904\u7406\u7b49\u573a\u666f\u3002<\/p>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"aligncenter size-full wp-image-29412\" title=\"Qwen2.5-Omni\uff1a\u591a\u6a21\u6001\u8f93\u5165\u548c\u5b9e\u65f6\u8bed\u97f3\u4ea4\u4e92\u7684\u7aef\u6d4b\u6a21\u578b-1\" src=\"https:\/\/www.kdjingpai.com\/wp-content\/uploads\/2025\/03\/d3a7ae3f2ae0056.png\" alt=\"Qwen2.5-Omni\uff1a\u591a\u6a21\u6001\u8f93\u5165\u548c\u5b9e\u65f6\u8bed\u97f3\u4ea4\u4e92\u7684\u7aef\u6d4b\u6a21\u578b-1\" width=\"1289\" height=\"692\" srcset=\"https:\/\/www.kdjingpai.com\/wp-content\/uploads\/2025\/03\/d3a7ae3f2ae0056.png 1289w, https:\/\/www.kdjingpai.com\/wp-content\/uploads\/2025\/03\/d3a7ae3f2ae0056-768x412.png 768w, https:\/\/www.kdjingpai.com\/wp-content\/uploads\/2025\/03\/d3a7ae3f2ae0056-18x10.png 18w\" sizes=\"auto, (max-width: 1289px) 100vw, 1289px\" \/><\/p>\n<p>&nbsp;<\/p>\n<h2>\u529f\u80fd\u5217\u8868<\/h2>\n<ul>\n<li>\u652f\u6301\u591a\u6a21\u6001\u8f93\u5165\uff1a\u80fd\u540c\u65f6\u5904\u7406\u6587\u672c\u3001\u56fe\u50cf\u3001\u97f3\u9891\u548c\u89c6\u9891\u3002<\/li>\n<li>\u5b9e\u65f6\u6d41\u5f0f\u54cd\u5e94\uff1a\u8f93\u5165\u540e\u7acb\u5373\u751f\u6210\u6587\u672c\u6216\u8bed\u97f3\u53cd\u9988\u3002<\/li>\n<li>\u81ea\u7136\u8bed\u97f3\u5408\u6210\uff1a\u751f\u6210\u6e05\u6670\u81ea\u7136\u7684\u8bed\u97f3\uff0c\u652f\u6301\u591a\u79cd\u97f3\u8272\u3002<\/li>\n<li>\u56fe\u50cf\u548c\u89c6\u9891\u7406\u89e3\uff1a\u8bc6\u522b\u56fe\u50cf\u5185\u5bb9\u6216\u5206\u6790\u89c6\u9891\u7247\u6bb5\u3002<\/li>\n<li>\u7aef\u5230\u7aef\u6307\u4ee4\u8ddf\u968f\uff1a\u76f4\u63a5\u6839\u636e\u8bed\u97f3\u6216\u6587\u672c\u6307\u4ee4\u5b8c\u6210\u4efb\u52a1\u3002<\/li>\n<li>\u5f00\u6e90\u514d\u8d39\uff1a\u63d0\u4f9b\u5b8c\u6574\u4ee3\u7801\u548c\u6a21\u578b\uff0c\u652f\u6301\u7528\u6237\u81ea\u5b9a\u4e49\u3002<\/li>\n<\/ul>\n<p>&nbsp;<\/p>\n<p>\u6211\u4eec\u5bf9 Qwen2.5-Omni \u8fdb\u884c\u4e86\u5168\u9762\u8bc4\u4f30\u3002\u7ed3\u679c\u663e\u793a\uff0c\u5b83\u5728\u6240\u6709\u6a21\u6001\u4e0a\u7684\u8868\u73b0\u90fd\u5f88\u5f3a\uff0c\u4e0e\u540c\u7b49\u89c4\u6a21\u7684\u5355\u4e00\u6a21\u6001\u6a21\u578b\u4ee5\u53ca\u95ed\u6e90\u6a21\u578b\uff08\u5982 Qwen2.5-VL-7B\u3001Qwen2-Audio \u548c Gemini-1.5-Pro\uff09\u76f8\u6bd4\u5177\u6709\u4f18\u52bf\u3002\u5728\u9700\u8981\u6574\u5408\u591a\u79cd\u6a21\u6001\u7684\u4efb\u52a1\u4e2d\uff0c\u4f8b\u5982 OmniBench\uff0cQwen2.5-Omni \u8fbe\u5230\u4e86\u6700\u5148\u8fdb\u7684\u6027\u80fd\u3002\u6b64\u5916\uff0c\u5728\u5355\u4e00\u6a21\u6001\u4efb\u52a1\u4e2d\uff0c\u5b83\u5728\u8bed\u97f3\u8bc6\u522b\uff08Common Voice\uff09\u3001\u7ffb\u8bd1\uff08CoVoST2\uff09\u3001\u97f3\u9891\u7406\u89e3\uff08MMAU\uff09\u3001\u56fe\u50cf\u63a8\u7406\uff08MMMU\u3001MMStar\uff09\u3001\u89c6\u9891\u7406\u89e3\uff08MVBench\uff09\u548c\u8bed\u97f3\u751f\u6210\uff08Seed-tts-eval \u548c\u4e3b\u89c2\u81ea\u7136\u5ea6\uff09\u7b49\u9886\u57df\u8868\u73b0\u51fa\u8272\u3002<\/p>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"aligncenter wp-image-29413\" title=\"Qwen2.5-Omni\uff1a\u591a\u6a21\u6001\u8f93\u5165\u548c\u5b9e\u65f6\u8bed\u97f3\u4ea4\u4e92\u7684\u7aef\u6d4b\u6a21\u578b-1\" src=\"https:\/\/www.kdjingpai.com\/wp-content\/uploads\/2025\/03\/122736e6df1a367.png\" alt=\"Qwen2.5-Omni\uff1a\u591a\u6a21\u6001\u8f93\u5165\u548c\u5b9e\u65f6\u8bed\u97f3\u4ea4\u4e92\u7684\u7aef\u6d4b\u6a21\u578b-1\" width=\"748\" height=\"536\" srcset=\"https:\/\/www.kdjingpai.com\/wp-content\/uploads\/2025\/03\/122736e6df1a367.png 945w, https:\/\/www.kdjingpai.com\/wp-content\/uploads\/2025\/03\/122736e6df1a367-768x550.png 768w, https:\/\/www.kdjingpai.com\/wp-content\/uploads\/2025\/03\/122736e6df1a367-18x12.png 18w\" sizes=\"auto, (max-width: 748px) 100vw, 748px\" \/><\/p>\n<p>&nbsp;<\/p>\n<h2>\u4f7f\u7528\u5e2e\u52a9<\/h2>\n<p>Qwen2.5-Omni \u7684\u4f7f\u7528\u9700\u8981\u4e00\u5b9a\u6280\u672f\u57fa\u7840\u3002\u4ee5\u4e0b\u662f\u8be6\u7ec6\u7684\u5b89\u88c5\u548c\u64cd\u4f5c\u6307\u5357\uff0c\u5e2e\u52a9\u7528\u6237\u5feb\u901f\u4e0a\u624b\u3002<\/p>\n<h3>\u5b89\u88c5\u6d41\u7a0b<\/h3>\n<ol>\n<li><strong>\u51c6\u5907\u73af\u5883<\/strong>\n<ul>\n<li>\u786e\u4fdd\u5b89\u88c5 Python 3.10 \u6216\u66f4\u9ad8\u7248\u672c\u3002<\/li>\n<li>\u9700\u8981 Git \u6765\u4e0b\u8f7d\u4ee3\u7801\u3002<\/li>\n<li>\u63a8\u8350\u4f7f\u7528 Linux \u7cfb\u7edf\uff0c\u975e Linux \u7cfb\u7edf\u53ef\u80fd\u9700\u8981\u989d\u5916\u914d\u7f6e\u3002<\/li>\n<\/ul>\n<\/li>\n<li><strong>\u4e0b\u8f7d\u4ee3\u7801<\/strong>\n<ul>\n<li>\u5728\u7ec8\u7aef\u8f93\u5165\u547d\u4ee4\u514b\u9686 GitHub \u4ed3\u5e93\uff1a\n<pre><code>git clone https:\/\/github.com\/QwenLM\/Qwen2.5-Omni.git\r\n<\/code><\/pre>\n<\/li>\n<li>\u8fdb\u5165\u9879\u76ee\u76ee\u5f55\uff1a\n<pre><code>cd Qwen2.5-Omni\r\n<\/code><\/pre>\n<\/li>\n<\/ul>\n<\/li>\n<li><strong>\u5b89\u88c5\u4f9d\u8d56<\/strong>\n<ul>\n<li>\u56e0\u4e3a\u4ee3\u7801\u5c1a\u672a\u5b8c\u5168\u5408\u5e76\u5230 Hugging Face \u4e3b\u5206\u652f\uff0c\u9700\u5b89\u88c5\u7279\u5b9a\u7248\u672c\u7684 Transformers\uff1a\n<pre><code>pip uninstall transformers\r\npip install git+https:\/\/github.com\/huggingface\/transformers@3a1ead0aabed473eafe527915eea8c197d424356\r\npip install accelerate\r\n<\/code><\/pre>\n<\/li>\n<li>\u5b89\u88c5\u591a\u6a21\u6001\u5904\u7406\u5de5\u5177\uff1a\n<pre><code>pip install qwen-omni-utils[decord]\r\n<\/code><\/pre>\n<ul>\n<li>\u6ce8\u610f\uff1a\u9700\u9884\u88c5\u00a0<code>ffmpeg<\/code>\uff0cLinux \u7528\u6237\u53ef\u8fd0\u884c\u00a0<code>sudo apt install ffmpeg<\/code>\u3002<\/li>\n<li>\u975e Linux \u7528\u6237\u82e5\u65e0\u6cd5\u5b89\u88c5\u00a0<code>decord<\/code>\uff0c\u53ef\u6539\u4e3a\uff1a\n<pre><code>pip install qwen-omni-utils\r\n<\/code><\/pre>\n<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n<\/li>\n<li><strong>\u4e0b\u8f7d\u6a21\u578b<\/strong>\n<ul>\n<li>\u4ece Hugging Face \u4e0b\u8f7d Qwen2.5-Omni-7B \u6a21\u578b\uff08<a href=\"https:\/\/huggingface.co\/Qwen\/Qwen2.5-Omni-7B\">https:\/\/huggingface.co\/Qwen\/Qwen2.5-Omni-7B<\/a>\uff09\uff0c\u4fdd\u5b58\u5230\u672c\u5730\u3002<\/li>\n<\/ul>\n<\/li>\n<li><strong>\u9a8c\u8bc1\u5b89\u88c5<\/strong>\n<ul>\n<li>\u8fd0\u884c\u4ee5\u4e0b\u547d\u4ee4\u68c0\u67e5\u73af\u5883\uff1a\n<pre><code>python -c \"from transformers import Qwen2_5OmniModel; print('\u5b89\u88c5\u6210\u529f')\"\r\n<\/code><\/pre>\n<\/li>\n<\/ul>\n<\/li>\n<\/ol>\n<h3>\u529f\u80fd\u64cd\u4f5c\u6d41\u7a0b<\/h3>\n<h4>1. \u5904\u7406\u6587\u672c\u8f93\u5165<\/h4>\n<ul>\n<li><strong>\u64cd\u4f5c\u6b65\u9aa4<\/strong>\uff1a\n<ul>\n<li>\u52a0\u8f7d\u6a21\u578b\u548c\u5904\u7406\u5668\uff1a\n<pre><code>from transformers import Qwen2_5OmniModel, Qwen2_5OmniProcessor\r\nmodel = Qwen2_5OmniModel.from_pretrained(\"Qwen\/Qwen2.5-Omni-7B\", device_map=\"auto\")\r\nprocessor = Qwen2_5OmniProcessor.from_pretrained(\"Qwen\/Qwen2.5-Omni-7B\")\r\n<\/code><\/pre>\n<\/li>\n<li>\u8f93\u5165\u6587\u672c\u5e76\u751f\u6210\u54cd\u5e94\uff1a\n<pre><code>text = \"\u4eca\u5929\u5929\u6c14\u600e\u4e48\u6837\uff1f\"\r\ninputs = processor(text=text, return_tensors=\"pt\")\r\noutputs = model.generate(**inputs)\r\nprint(processor.batch_decode(outputs, skip_special_tokens=True))\r\n<\/code><\/pre>\n<\/li>\n<\/ul>\n<\/li>\n<li><strong>\u7ed3\u679c<\/strong>\uff1a\u8fd4\u56de\u6587\u672c\u56de\u7b54\uff0c\u5982\u201c\u4eca\u5929\u5929\u6c14\u6674\u6717\uff0c\u6e29\u5ea6\u9002\u5b9c\u3002\u201d<\/li>\n<\/ul>\n<h4>2. \u5904\u7406\u56fe\u50cf\u8f93\u5165<\/h4>\n<ul>\n<li><strong>\u64cd\u4f5c\u6b65\u9aa4<\/strong>\uff1a\n<ul>\n<li>\u51c6\u5907\u56fe\u7247\u6587\u4ef6\uff08\u5982\u00a0<code>image.jpg<\/code>\uff09\u3002<\/li>\n<li>\u4fee\u6539\u811a\u672c\u52a0\u5165\u56fe\u50cf\uff1a\n<pre><code>images = [\"image.jpg\"]\r\ntext = \"\u56fe\u7247\u91cc\u6709\u4ec0\u4e48\uff1f\"\r\ninputs = processor(text=text, images=images, return_tensors=\"pt\")\r\noutputs = model.generate(**inputs)\r\nprint(processor.batch_decode(outputs, skip_special_tokens=True))\r\n<\/code><\/pre>\n<\/li>\n<\/ul>\n<\/li>\n<li><strong>\u7ed3\u679c<\/strong>\uff1a\u63cf\u8ff0\u56fe\u7247\u5185\u5bb9\uff0c\u5982\u201c\u56fe\u7247\u91cc\u6709\u4e00\u53ea\u72d7\u5728\u8349\u5730\u4e0a\u8dd1\u3002\u201d<\/li>\n<\/ul>\n<h4>3. \u5904\u7406\u97f3\u9891\u8f93\u5165<\/h4>\n<ul>\n<li><strong>\u64cd\u4f5c\u6b65\u9aa4<\/strong>\uff1a\n<ul>\n<li>\u51c6\u5907\u97f3\u9891\u6587\u4ef6\uff08\u5982\u00a0<code>audio.wav<\/code>\uff09\u3002<\/li>\n<li>\u4fee\u6539\u811a\u672c\u52a0\u5165\u97f3\u9891\uff1a\n<pre><code>audios = [\"audio.wav\"]\r\ntext = \"\u97f3\u9891\u91cc\u8bf4\u4e86\u4ec0\u4e48\uff1f\"\r\ninputs = processor(text=text, audios=audios, return_tensors=\"pt\")\r\noutputs = model.generate(**inputs)\r\nprint(processor.batch_decode(outputs, skip_special_tokens=True))\r\n<\/code><\/pre>\n<\/li>\n<\/ul>\n<\/li>\n<li><strong>\u7ed3\u679c<\/strong>\uff1a\u8f6c\u5f55\u97f3\u9891\u5185\u5bb9\uff0c\u5982\u201c\u97f3\u9891\u91cc\u8bf4\u2018\u660e\u5929\u6709\u96e8\u2019\u3002\u201d<\/li>\n<\/ul>\n<h4>4. \u5904\u7406\u89c6\u9891\u8f93\u5165<\/h4>\n<ul>\n<li><strong>\u64cd\u4f5c\u6b65\u9aa4<\/strong>\uff1a\n<ul>\n<li>\u51c6\u5907\u89c6\u9891\u6587\u4ef6\uff08\u5982\u00a0<code>video.mp4<\/code>\uff09\u3002<\/li>\n<li>\u4fee\u6539\u811a\u672c\u52a0\u5165\u89c6\u9891\uff1a\n<pre><code>videos = [\"video.mp4\"]\r\ntext = \"\u89c6\u9891\u91cc\u53d1\u751f\u4e86\u4ec0\u4e48\uff1f\"\r\ninputs = processor(text=text, videos=videos, return_tensors=\"pt\")\r\noutputs = model.generate(**inputs)\r\nprint(processor.batch_decode(outputs, skip_special_tokens=True))\r\n<\/code><\/pre>\n<\/li>\n<\/ul>\n<\/li>\n<li><strong>\u7ed3\u679c<\/strong>\uff1a\u63cf\u8ff0\u89c6\u9891\u5185\u5bb9\uff0c\u5982\u201c\u89c6\u9891\u91cc\u6709\u4eba\u5728\u753b\u753b\u3002\u201d<\/li>\n<\/ul>\n<h4>5. \u751f\u6210\u8bed\u97f3\u8f93\u51fa<\/h4>\n<ul>\n<li><strong>\u64cd\u4f5c\u6b65\u9aa4<\/strong>\uff1a\n<ul>\n<li>\u8bbe\u7f6e\u7cfb\u7edf\u63d0\u793a\u5e76\u542f\u7528\u8bed\u97f3\uff1a\n<pre><code>conversation = [\r\n{\"role\": \"system\", \"content\": \"You are Qwen, a virtual human developed by the Qwen Team, Alibaba Group, capable of perceiving auditory and visual inputs, as well as generating text and speech.\"},\r\n{\"role\": \"user\", \"content\": \"\u8bf7\u7528\u8bed\u97f3\u56de\u7b54\uff1a\u4eca\u5929\u5929\u6c14\u5982\u4f55\uff1f\"}\r\n]\r\ntext = processor.apply_chat_template(conversation, tokenize=False, add_generation_prompt=True)\r\ninputs = processor(text=text, return_tensors=\"pt\")\r\ntext_ids, audio = model.generate(**inputs, use_audio_in_video=True)\r\nimport soundfile as sf\r\nsf.write(\"output.wav\", audio.reshape(-1).detach().cpu().numpy(), samplerate=24000)\r\n<\/code><\/pre>\n<\/li>\n<\/ul>\n<\/li>\n<li><strong>\u7ed3\u679c<\/strong>\uff1a\u751f\u6210\u00a0<code>output.wav<\/code>\u00a0\u6587\u4ef6\uff0c\u5185\u5bb9\u4e3a\u8bed\u97f3\u56de\u7b54\u3002<\/li>\n<\/ul>\n<h4>6. \u8c03\u6574\u8bed\u97f3\u97f3\u8272<\/h4>\n<ul>\n<li><strong>\u64cd\u4f5c\u6b65\u9aa4<\/strong>\uff1a\n<ul>\n<li>\u5728\u751f\u6210\u65f6\u6307\u5b9a\u97f3\u8272\uff08\u5982 Chelsie \u6216 Ethan\uff09\uff1a\n<pre><code>text_ids, audio = model.generate(**inputs, spk=\"Ethan\")\r\nsf.write(\"output.wav\", audio.reshape(-1).detach().cpu().numpy(), samplerate=24000)\r\n<\/code><\/pre>\n<\/li>\n<\/ul>\n<\/li>\n<li><strong>\u7ed3\u679c<\/strong>\uff1a\u751f\u6210\u6307\u5b9a\u97f3\u8272\u7684\u8bed\u97f3\u6587\u4ef6\u3002<\/li>\n<\/ul>\n<h4>7. \u4f7f\u7528 FlashAttention-2 \u52a0\u901f<\/h4>\n<ul>\n<li><strong>\u64cd\u4f5c\u6b65\u9aa4<\/strong>\uff1a\n<ul>\n<li>\u5b89\u88c5 FlashAttention-2\uff1a\n<pre><code>pip install -U flash-attn --no-build-isolation\r\n<\/code><\/pre>\n<\/li>\n<li>\u52a0\u8f7d\u6a21\u578b\u65f6\u542f\u7528\uff1a\n<pre><code>model = Qwen2_5OmniModel.from_pretrained(\"Qwen\/Qwen2.5-Omni-7B\", device_map=\"auto\", attn_implementation=\"flash_attention_2\")\r\n<\/code><\/pre>\n<\/li>\n<\/ul>\n<\/li>\n<li><strong>\u7ed3\u679c<\/strong>\uff1a\u751f\u6210\u901f\u5ea6\u66f4\u5feb\uff0c\u663e\u5b58\u5360\u7528\u66f4\u4f4e\u3002<\/li>\n<\/ul>\n<h3>\u6ce8\u610f\u4e8b\u9879<\/h3>\n<ul>\n<li><strong>\u786c\u4ef6\u8981\u6c42<\/strong>\uff1a\u63a8\u8350\u4f7f\u7528 GPU\uff0c\u81f3\u5c11 16GB \u663e\u5b58\u3002<\/li>\n<li><strong>\u7f51\u7edc\u9700\u6c42<\/strong>\uff1a\u4e0b\u8f7d\u6a21\u578b\u548c\u4f9d\u8d56\u65f6\u9700\u7a33\u5b9a\u7f51\u7edc\u3002<\/li>\n<li><strong>\u8c03\u8bd5\u652f\u6301<\/strong>\uff1a\u5982\u9047\u95ee\u9898\uff0c\u53ef\u53c2\u8003 GitHub \u7684\u00a0<code>README.md<\/code>\u00a0\u6216\u793e\u533a\u8ba8\u8bba\u3002<\/li>\n<\/ul>\n<p>&nbsp;<\/p>\n<h2>\u5e94\u7528\u573a\u666f<\/h2>\n<ol>\n<li><strong>\u5b9e\u65f6\u8bed\u97f3\u52a9\u624b<\/strong><br \/>\n\u7528\u6237\u901a\u8fc7\u8bed\u97f3\u63d0\u95ee\uff0c\u6a21\u578b\u5b9e\u65f6\u751f\u6210\u8bed\u97f3\u56de\u7b54\uff0c\u9002\u5408\u5ba2\u670d\u6216\u4e2a\u4eba\u52a9\u7406\u3002<\/li>\n<li><strong>\u89c6\u9891\u5185\u5bb9\u5206\u6790<\/strong><br \/>\n\u8f93\u5165\u89c6\u9891\uff0c\u6a21\u578b\u63d0\u53d6\u5173\u952e\u4fe1\u606f\uff0c\u5e2e\u52a9\u7528\u6237\u6574\u7406\u7d20\u6750\u6216\u751f\u6210\u62a5\u544a\u3002<\/li>\n<li><strong>\u6559\u80b2\u652f\u6301<\/strong><br \/>\n\u5b66\u751f\u4e0a\u4f20\u8bfe\u7a0b\u97f3\u9891\u6216\u89c6\u9891\uff0c\u6a21\u578b\u89e3\u7b54\u95ee\u9898\u6216\u63d0\u53d6\u91cd\u70b9\uff0c\u63d0\u5347\u5b66\u4e60\u6548\u7387\u3002<\/li>\n<\/ol>\n<p>&nbsp;<\/p>\n<h2>QA<\/h2>\n<ol>\n<li><strong>\u652f\u6301\u54ea\u4e9b\u8bed\u8a00\uff1f<\/strong><br \/>\n\u4e3b\u8981\u652f\u6301\u4e2d\u6587\u548c\u82f1\u6587\uff0c\u8bed\u97f3\u5408\u6210\u63d0\u4f9b Chelsie\u3001Ethan \u7b49\u591a\u79cd\u97f3\u8272\u3002<\/li>\n<li><strong>\u9700\u8981\u591a\u5c11\u5b58\u50a8\u7a7a\u95f4\uff1f<\/strong><br \/>\nQwen2.5-Omni-7B \u6a21\u578b\u7ea6 14GB\uff0c\u5efa\u8bae\u9884\u7559 20GB \u4ee5\u4e0a\u3002<\/li>\n<li><strong>\u53ef\u4ee5\u5546\u7528\u5417\uff1f<\/strong><br \/>\n\u53ef\u4ee5\uff0c\u57fa\u4e8e Apache 2.0 \u8bb8\u53ef\uff0c\u514d\u8d39\u5546\u7528\u9700\u9075\u5b88\u6761\u6b3e\u3002<\/li>\n<\/ol>\n","protected":false},"excerpt":{"rendered":"<p>Qwen2.5-Omni \u662f\u963f\u91cc\u5df4\u5df4\u4e91 Qwen \u56e2\u961f\u5f00\u53d1\u7684\u4e00\u6b3e\u5f00\u6e90\u591a\u6a21\u6001 AI \u6a21\u578b\u3002\u5b83\u80fd\u5904\u7406\u6587\u672c\u3001\u56fe\u50cf\u3001\u97f3\u9891\u548c\u89c6\u9891\u7b49\u591a\u79cd\u8f93\u5165\uff0c\u5e76\u5b9e\u65f6\u751f\u6210\u6587\u672c\u6216\u81ea\u7136\u8bed\u97f3\u54cd\u5e94\u3002\u8fd9\u6b3e\u6a21\u578b\u4e8e 2025 \u5e74 3 \u6708 26 \u65e5\u53d1\u5e03\uff0c\u4ee3\u7801\u548c\u6a21\u578b\u6587\u4ef6\u6258\u7ba1\u5728 GitHu&#8230;<\/p>\n","protected":false},"author":1,"featured_media":32782,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[392,402],"tags":[230,374],"class_list":["post-29410","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-models","category-multimodal-models","tag-aikaiyuanxiangmu","tag-duandaoduanyuyinzhubei"],"_links":{"self":[{"href":"https:\/\/www.kdjingpai.com\/en\/wp-json\/wp\/v2\/posts\/29410","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.kdjingpai.com\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.kdjingpai.com\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.kdjingpai.com\/en\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.kdjingpai.com\/en\/wp-json\/wp\/v2\/comments?post=29410"}],"version-history":[{"count":0,"href":"https:\/\/www.kdjingpai.com\/en\/wp-json\/wp\/v2\/posts\/29410\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.kdjingpai.com\/en\/wp-json\/wp\/v2\/media\/32782"}],"wp:attachment":[{"href":"https:\/\/www.kdjingpai.com\/en\/wp-json\/wp\/v2\/media?parent=29410"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.kdjingpai.com\/en\/wp-json\/wp\/v2\/categories?post=29410"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.kdjingpai.com\/en\/wp-json\/wp\/v2\/tags?post=29410"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}