{"kind":"guides","slug":"train-tts-voice","path":"/guides/train-tts-voice","title":"用自己的 GPU 训练 TTS 声音 | SuperClonica","description":"如何用 5 到 10 秒录音在 NVIDIA GPU 上训练自己的语音合成声音：配置要求、步骤、检查点，以及用 Supertonic 3 离线使用（暂不支持中文）。","h1":"在自己的 NVIDIA GPU 上训练 TTS 声音","intro":"从零训练语音合成模型需要数小时的干净录音和强大的硬件。大多数人并不需要这样：他们需要的是一个能让好模型说话的新声音。SuperClonica 训练的正是这个：为开源多语言模型 Supertonic 3 训练一个声音，只需一段 5 到 10 秒的干净录音。请注意：Supertonic 3 暂不支持中文，可朗读英语、日语、韩语等 31 种语言。\n\n本指南介绍所需条件、如何把带 NVIDIA 显卡的电脑接入为 worker、工作室中的训练流程，以及如何使用或出售训练好的声音。用自己的硬件训练是免费的。","sections":[{"heading":"训练的是什么，为什么这么快","paragraphs":["基础模型已经会说 31 种语言。训练是为它找到一个声音：一个描述录音音色和说话方式的小 JSON 文件。所以一段短参考录音就够了，训练好的声音也能在普通处理器上离线运行。"]},{"heading":"电脑配置要求","paragraphs":["worker 客户端运行在 Windows 上；模式在 worker 卡片上选择，“自动”模式下 worker 会等待空闲内存，15 分钟后切换到“迷你”模式。"],"points":["Windows 10/11（64 位），NVIDIA RTX 20 系列或更新的显卡（GTX 16 也可以），驱动 580 及以上。","完整质量：12 GB 显存（最好 16 GB）和 32 GB 内存。","迷你模式：8 GB 显存和 16 GB 内存，声音相似度会略低。","15 GB 可用磁盘空间，并开启 Windows 页面文件。"]},{"heading":"第 1 步：创建声音和参考录音","paragraphs":["在工作室中创建声音并添加参考录音：一段 5 到 10 秒的干净录音及其逐字准确的文本。在安静的房间录制，只有一个说话人，音量稳定，用声音应有的说话方式。请使用自己的声音或已获对方同意的声音。"]},{"heading":"第 2 步：把电脑接入为 worker","paragraphs":["打开“我的 worker”，添加一个 worker 并下载 SuperClonicaWorker.zip：其中包含客户端和你的 worker.json（含平台地址和个人令牌）。解压后运行 SuperClonicaWorker.exe，首次启动会下载其余所需文件，之后自动更新。worker 只训练你的声音。"]},{"heading":"第 3 步：训练并比较检查点","paragraphs":["在参考录音上点击“训练”，选择步数并开始；页面会根据以往训练估算 500 步所需时间。训练过程中，按试听和与参考录音的相似度比较检查点，并继续训练最有希望的那个。"]},{"heading":"第 4 步：使用或出售声音","paragraphs":["在训练页面免费下载 JSON，用免费的 SuperClonica 引擎离线朗读任何文本：python -m pip install superclonica-cpu，然后 superclonica --text \"Hello!\" --voice my-voice.json --lang en --output hello.wav。如果愿意，可以配上封面、说明和价格在目录中发布，或在交易所接其他用户的声音订单。"]}],"faq":[{"question":"训练免费吗？","answer":"免费。在自己的电脑上训练免费，训练好的 JSON 也可免费下载。"},{"question":"我没有 NVIDIA 显卡怎么办？","answer":"可以在交易所下单：作者用你的录音在自己的硬件上训练，你验收结果后才付款。使用训练好的声音不需要显卡。"},{"question":"训练需要多长时间？","answer":"取决于显卡和步数。worker 页面会根据以往训练显示 500 步的估计时间。"},{"question":"worker 能在 Linux 或 macOS 上运行吗？","answer":"worker 客户端面向带 NVIDIA 显卡的 Windows 10/11。训练好的声音和 SuperClonica 引擎可在 Windows、macOS 和 Linux 上运行。"}],"cta":{"catalog":"浏览声音","install":"安装 SuperClonica 引擎"},"sources":[{"label":"Supertonic 3 模型卡片","url":"https://huggingface.co/supertone-oss-archive/supertonic-3"},{"label":"PyPI 上的 SuperClonica 引擎","url":"https://pypi.org/project/superclonica-cpu/"}],"published":"2026-10-24","updated":"2026-10-24","pages":[{"slug":"offline-text-to-speech","path":"/guides/offline-text-to-speech","h1":"如何在自己的电脑上离线把文字转成语音"},{"slug":"voice-cloning","path":"/guides/voice-cloning","h1":"声音克隆：如何得到一个像录音里那样的声音"},{"slug":"best-free-tts-models","path":"/guides/best-free-tts-models","h1":"2026 年适合本地运行的免费语音合成模型"},{"slug":"text-to-speech-for-videos","path":"/guides/text-to-speech-for-videos","h1":"为视频、有声书、游戏和应用做文字转语音配音"},{"slug":"supertonic-3-voices","path":"/guides/supertonic-3-voices","h1":"Supertonic 3 音色：自定义音色的来源与使用方法"},{"slug":"python-text-to-speech","path":"/guides/python-text-to-speech","h1":"Python 文字转语音：离线、免费、自然的音色"},{"slug":"text-to-speech-api-docker","path":"/guides/text-to-speech-api-docker","h1":"在 Docker 中自建文字转语音 API"},{"slug":"character-voice-generator","path":"/guides/character-voice-generator","h1":"用角色声音朗读文字：让你喜欢的角色说出任何台词"},{"slug":"free-unlimited-text-to-speech","path":"/guides/free-unlimited-text-to-speech","h1":"免费、无限制、无需注册的文字转语音：在自己的电脑上运行"},{"slug":"narrator-voice-generator","path":"/guides/narrator-voice-generator","h1":"旁白配音生成器：为视频、有声书和播客配旁白"},{"slug":"child-voice-text-to-speech","path":"/guides/child-voice-text-to-speech","h1":"用儿童声音朗读文字：儿童与青少年 AI 声音"},{"slug":"anime-voice-generator","path":"/guides/anime-voice-generator","h1":"动漫声音生成器：用动漫角色的声音朗读文字"},{"slug":"text-to-speech-languages","path":"/guides/text-to-speech-languages","h1":"31 种语言的文字转语音：一个声音说遍所有语言"},{"slug":"female-voice-text-to-speech","path":"/guides/female-voice-text-to-speech","h1":"女声文字转语音：自然的 AI 女声"},{"slug":"male-voice-text-to-speech","path":"/guides/male-voice-text-to-speech","h1":"男声文字转语音：低沉自然的 AI 男声"},{"slug":"realistic-text-to-speech","path":"/guides/realistic-text-to-speech","h1":"逼真的文字转语音：像真人一样说话的 AI 声音"},{"slug":"ai-voice-generator","path":"/guides/ai-voice-generator","h1":"AI 声音生成器：用现成的 AI 声音把文字变成语音"},{"slug":"text-to-speech-reader","path":"/guides/text-to-speech-reader","h1":"文字朗读软件：用耳朵听书籍、PDF 和文章"},{"slug":"tiktok-voice-generator","path":"/guides/tiktok-voice-generator","h1":"TikTok 配音生成器：TikTok、Shorts 和 Reels 的声音"},{"slug":"dota-2-voice-generator","path":"/guides/dota-2-voice-generator","h1":"DOTA2 英雄声音：让任何英雄说出你的台词"},{"slug":"cartoon-voice-generator","path":"/guides/cartoon-voice-generator","h1":"卡通声音生成器：让卡通角色说出你的文字"},{"slug":"jarvis-voice-generator","path":"/guides/jarvis-voice-generator","h1":"贾维斯声音生成器：用 AI 助手的声音朗读你的文字"},{"slug":"darth-vader-voice-generator","path":"/guides/darth-vader-voice-generator","h1":"达斯·维达声音生成器：用星球大战角色的声音朗读你的文字"},{"slug":"train-tts-voice","path":"/guides/train-tts-voice","h1":"在自己的 NVIDIA GPU 上训练 TTS 声音"},{"slug":"emotional-text-to-speech","path":"/guides/emotional-text-to-speech","h1":"有感情的文字转语音：如何让 AI 声音富有表现力"},{"slug":"hindi-text-to-speech","path":"/guides/hindi-text-to-speech","h1":"印地语文字转语音：会读印地语的 AI 声音"},{"slug":"spanish-text-to-speech","path":"/guides/spanish-text-to-speech","h1":"西班牙语文字转语音：会读西班牙语的 AI 声音"},{"slug":"japanese-text-to-speech","path":"/guides/japanese-text-to-speech","h1":"日语文字转语音：会读日语的 AI 声音"},{"slug":"russian-text-to-speech","path":"/guides/russian-text-to-speech","h1":"用俄语录制的声音进行俄语文字转语音"},{"slug":"portuguese-text-to-speech","path":"/guides/portuguese-text-to-speech","h1":"葡萄牙语文字转语音：会读葡萄牙语的 AI 声音"},{"slug":"german-text-to-speech","path":"/guides/german-text-to-speech","h1":"德语文字转语音：会读德语的 AI 声音"},{"slug":"french-text-to-speech","path":"/guides/french-text-to-speech","h1":"法语文字转语音：会读法语的 AI 声音"},{"slug":"korean-text-to-speech","path":"/guides/korean-text-to-speech","h1":"韩语文字转语音：会读韩语的 AI 声音"},{"slug":"english-text-to-speech","path":"/guides/english-text-to-speech","h1":"英语文字转语音：会读英语的 AI 声音"},{"slug":"arabic-text-to-speech","path":"/guides/arabic-text-to-speech","h1":"阿拉伯语文字转语音：会读阿拉伯语的 AI 声音"},{"slug":"witcher-voice-generator","path":"/guides/witcher-voice-generator","h1":"杰洛特声音生成器：用巫师系列的声音朗读你的文字"},{"slug":"gta-5-voice-generator","path":"/guides/gta-5-voice-generator","h1":"GTA5 声音生成器：让崔佛、富兰克林和莱斯特说出你的文字"},{"slug":"audiobook-ai-voice","path":"/guides/audiobook-ai-voice","h1":"如何用 AI 声音制作有声书"},{"slug":"text-to-speech-for-games","path":"/guides/text-to-speech-for-games","h1":"游戏开发用文字转语音：无需录音棚的 NPC 声音"},{"slug":"italian-text-to-speech","path":"/guides/italian-text-to-speech","h1":"意大利语文字转语音：会读意大利语的 AI 声音"},{"slug":"polish-text-to-speech","path":"/guides/polish-text-to-speech","h1":"波兰语文字转语音：会读波兰语的 AI 声音"},{"slug":"turkish-text-to-speech","path":"/guides/turkish-text-to-speech","h1":"土耳其语文字转语音：会读土耳其语的 AI 声音"},{"slug":"ukrainian-text-to-speech","path":"/guides/ukrainian-text-to-speech","h1":"乌克兰语文字转语音：会读乌克兰语的 AI 声音"},{"slug":"vietnamese-text-to-speech","path":"/guides/vietnamese-text-to-speech","h1":"越南语文字转语音：会读越南语的 AI 声音"},{"slug":"indonesian-text-to-speech","path":"/guides/indonesian-text-to-speech","h1":"印尼语文字转语音：会读印尼语的 AI 声音"},{"slug":"meme-voice-generator","path":"/guides/meme-voice-generator","h1":"表情包配音生成器：为你的梗配上搞笑声音"},{"slug":"vhs-voice-over-dub","path":"/guides/vhs-voice-over-dub","h1":"VHS 配音：90 年代单人翻译配音风格"},{"slug":"stream-alert-voices","path":"/guides/stream-alert-voices","h1":"直播提醒配音：用角色声音播报关注、订阅和打赏"}],"related":[{"slug":"text-to-speech","path":"/text-to-speech","h1":"在自己电脑上运行的语音合成"},{"slug":"supertonic","path":"/compare/supertonic","h1":"Supertonic 与 SuperClonica：同一模型，更快更干净"},{"slug":"fish-audio","path":"/compare/fish-audio","h1":"Fish Audio、Fish Speech 与 SuperClonica 对比"},{"slug":"omnivoice","path":"/compare/omnivoice","h1":"OmniVoice 与 SuperClonica：语言更多，还是速度更快"},{"slug":"qwen3-tts","path":"/compare/qwen3-tts","h1":"本地语音合成：Qwen3-TTS 还是 SuperClonica"},{"slug":"voxcpm","path":"/compare/voxcpm","h1":"VoxCPM 2 与 SuperClonica：录音棚音质还是 CPU 速度"},{"slug":"f5-tts","path":"/compare/f5-tts","h1":"F5-TTS 与 SuperClonica：零样本克隆还是 CPU 上的快速合成"},{"slug":"elevenlabs","path":"/compare/elevenlabs","h1":"可离线使用的 ElevenLabs 替代方案"},{"slug":"kokoro","path":"/compare/kokoro","h1":"Kokoro TTS 与 SuperClonica：两个小模型的本地语音合成"},{"slug":"piper","path":"/compare/piper","h1":"Piper TTS 与 SuperClonica：小设备还是定制音色"},{"slug":"xtts","path":"/compare/xtts","h1":"Coqui XTTS-v2 与 SuperClonica：Coqui TTS 的替代方案"},{"slug":"chatterbox","path":"/compare/chatterbox","h1":"Chatterbox TTS 与 SuperClonica：即时克隆还是训练好的音色文件"}]}