Power Users/Automators 折腾/讨论/分享各种开源工具/脚本/自动化工作流👥 @zhetengsha_group📌 合集 https://t.me/zhetengsha/2🎁 恰饭 https://t.me/zhetengsha/957📢 广告投放 @xream @xream_botBuy ads: https://telega.io/c/zhetengshafeedId:55438372655431680+userId:62307599601855488http://telegram.zhetengsha.eu.orghttp://telegram.zhetengsha.eu.org/posts/5175http://telegram.zhetengsha.eu.org/posts/5175Tue, 24 Mar 2026 07:00:00 GMT<u>JANG: 面向 Apple Silicon 的 MLX 混合精度量化格式</u><br /><br />• 针对 MoE 与超大模型做按张量分配比特宽度，在接近 MLX 体积下优先保留 attention 与 router 精度，低比特场景下稳定性和效果更强<br /><br />• 支持 Qwen、Nemotron、MiniMax、DeepSeek 等架构，部分模型可在 16 GB 或 64 GB Mac 上运行，甚至实现 397B 级模型在 128 GB Mac 上推理<br /><br />• 原生兼容 MLX 生态，提供推理模式、VLM 支持、bfloat16 自动检测和开发者集成方案，适合在 Apple Silicon 上部署高压缩本地模型<br /><br /><a href="https://github.com/jjang-ai/jangq" target="_blank" rel="noopener" title="https://github.com/jjang-ai/jangq">https://github.com/jjang-ai/jangq</a><br /><br /><a href="/search/result?q=%23AppleSilicon" title="#AppleSilicon">#AppleSilicon</a> <a href="/search/result?q=%23MLX" title="#MLX">#MLX</a> <a href="/search/result?q=%23%E6%A8%A1%E5%9E%8B%E9%87%8F%E5%8C%96" title="#模型量化">#模型量化</a> <a href="/search/result?q=%23%E6%B7%B7%E5%90%88%E7%B2%BE%E5%BA%A6%E9%87%8F%E5%8C%96" title="#混合精度量化">#混合精度量化</a> <a href="/search/result?q=%23%E6%9C%AC%E5%9C%B0%E5%A4%A7%E6%A8%A1%E5%9E%8B" title="#本地大模型">#本地大模型</a> <a href="/search/result?q=%23MoE" title="#MoE">#MoE</a> <a href="/search/result?q=%23%E6%8E%A8%E7%90%86%E4%BC%98%E5%8C%96" title="#推理优化">#推理优化</a> <a href="/search/result?q=%23Mac" title="#Mac">#Mac</a> <a href="/search/result?q=%23AI" title="#AI">#AI</a> <a href="/search/result?q=%23GitHub" title="#GitHub">#GitHub</a><a class="tgme_widget_message_link_preview" href="https://github.com/jjang-ai/jangq" target="_blank" rel="noopener" title="JANG — GGUF for MLX. YOU MUST USE JANG_Q RUNTIME. Adaptive Mixed-Precision Quantization + Runtime for Apple Silicon - jjang-ai/jangq"> <div class="link_preview_site_name accent_color">GitHub</div> <img class="link_preview_image" alt="GitHub - jjang-ai/jangq: JANG — GGUF for MLX. YOU MUST USE JANG_Q RUNTIME. Adaptive Mixed-Precision Quantization + Runtime for…" src="/static/https://cdn4.telesco.pe/file/a-4YuxjLJ1eo_HWZGr2PJ7XP6HDagnzWTHLT5H5tStVN1ruInK-zNbrjEAkD-SPLRo5pj3nTx4NIlmm-yvqUlVNs-BZGtB-OUjl5qrxQsy9XW-rhWgXiNlzMwLGUgIvu4mDHUgVEBzUXV_PNwTnxGJYotQ2eVLZX3CvtFsmadpoHK2_hCRQPrtmZncOwdgjCtPrxwe4zWhYR0d-Z1DEgXNBbmvGNLOzl1cvHrIHgG1POHJOH4soSn2MDAA1qKR8w7AJMPc-zlz9BbmhET9w9taHlJ-uizzGRfPywYdJTqD62ftJvTWbmJQEDHf6j4_wqzDz6ckzaM5snCjn_6vlsaw.jpg" width="1200" height="630" loading="eager" /> <div class="link_preview_title">GitHub - jjang-ai/jangq: JANG — GGUF for MLX. YOU MUST USE JANG_Q RUNTIME. Adaptive Mixed-Precision Quantization + Runtime for…</div> <div class="link_preview_description">JANG — GGUF for MLX. YOU MUST USE JANG_Q RUNTIME. Adaptive Mixed-Precision Quantization + Runtime for Apple Silicon - jjang-ai/jangq</div> </a>http://telegram.zhetengsha.eu.org/posts/5106http://telegram.zhetengsha.eu.org/posts/5106Tue, 17 Mar 2026 14:02:40 GMT<div class="image-list-container image-list-odd"> <button type="button" class="image-preview-button image-preview-wrap" popovertarget="modal-5106-0" popovertargetaction="show" aria-label="Open image preview: Metapi: 将分散的 AI 中转站聚合为统一入口的元网关上科技 自动签到• 聚合 New API、One API、OneHub、DoneHub、Veloera、AnyRouter、Sub2API 等多种上游平台，对下游提供一个 API Key 和统一的 /v1/* 入口• 自动发现模型并基于成本、余额、使用率进行智能路由，支持故障切换、失败重试与多通道分流，适合长期稳定供给• 自带可视化控制台，覆盖站点管理、余额看板、签到任务、Token 生命周期、代理日志与告警通知，部署上也支持单 Docker 轻量运行"> <img src="/static/https://cdn5.telesco.pe/file/iHz3Q7k5GwwZzKN6zog8kgJ5dERzBrnonR2JPbbnW_FGxWNtGjk7-NJ06aXNBF5bszjq1sypI2K4nV2LDd9ipIkQIvPawUFqn2BvfnY-UfpjKsfoj2jKtBPWZ4isCslwPQ97ISL-n4wBmiC8QGa1yEm2m0S0h88bmon273O3J3CIoF73CED2mn3QdM3c14_HT_vMOQtjykl_jdDVNDRRCmGHlVrO0MauBK7sOR5mJs97YcsZBlu9_v0U5V32Nt3RrSRWpF6MVNs-S6Km-Ai74mf7MUmiskbstbNYgSkeXr7Jf2WTblH39RwxIa6J01uSjmLw2vWTzhklpFDZAIzi1A.jpg" alt="Metapi: 将分散的 AI 中转站聚合为统一入口的元网关上科技 自动签到• 聚合 New API、One API、OneHub、DoneHub、Veloera、AnyRouter、Sub2API 等多种上游平台，对下游提供一个 API Key 和统一的 /v1/* 入口• 自动发现模型并基于成本、余额、使用率进行智能路由，支持故障切换、失败重试与多通道分流，适合长期稳定供给• 自带可视化控制台，覆盖站点管理、余额看板、签到任务、Token 生命周期、代理日志与告警通知，部署上也支持单 Docker 轻量运行" width="800" height="272" loading="eager" /> </button> <div class="modal" id="modal-5106-0" popover="auto" aria-label="Image preview"> <button type="button" class="modal__backdrop" popovertarget="modal-5106-0" popovertargetaction="hide" aria-label="Close image preview"></button> <button type="button" class="modal__close" popovertarget="modal-5106-0" popovertargetaction="hide" aria-label="Close image preview">×</button> <div class="modal__surface"> </div> </div> </div><u>Metapi: 将分散的 AI 中转站聚合为统一入口的元网关</u><br /><br />上科技 自动签到<br /><br />• 聚合 New API、One API、OneHub、DoneHub、Veloera、AnyRouter、Sub2API 等多种上游平台，对下游提供一个 API Key 和统一的 /v1/* 入口<br /><br />• 自动发现模型并基于成本、余额、使用率进行智能路由，支持故障切换、失败重试与多通道分流，适合长期稳定供给<br /><br />• 自带可视化控制台，覆盖站点管理、余额看板、签到任务、Token 生命周期、代理日志与告警通知，部署上也支持单 Docker 轻量运行<br /><br /><a href="https://github.com/cita-777/metapi" target="_blank" rel="noopener" title="https://github.com/cita-777/metapi">https://github.com/cita-777/metapi</a><br /><br /><a href="/search/result?q=%23AI" title="#AI">#AI</a> <a href="/search/result?q=%23%E7%BD%91%E5%85%B3" title="#网关">#网关</a> <a href="/search/result?q=%23API" title="#API">#API</a> <a href="/search/result?q=%23%E8%81%9A%E5%90%88" title="#聚合">#聚合</a> <a href="/search/result?q=%23%E6%99%BA%E8%83%BD%E8%B7%AF%E7%94%B1" title="#智能路由">#智能路由</a> <a href="/search/result?q=%23%E6%A8%A1%E5%9E%8B%E4%B8%AD%E8%BD%AC" title="#模型中转">#模型中转</a> <a href="/search/result?q=%23%E8%87%AA%E6%89%98%E7%AE%A1" title="#自托管">#自托管</a> <a href="/search/result?q=%23Docker" title="#Docker">#Docker</a> <a href="/search/result?q=%23TypeScript" title="#TypeScript">#TypeScript</a> <a href="/search/result?q=%23GitHub" title="#GitHub">#GitHub</a>http://telegram.zhetengsha.eu.org/posts/5093http://telegram.zhetengsha.eu.org/posts/5093Mon, 16 Mar 2026 13:21:45 GMT<u>Qwen3.5-9B-Claude-4.6-Opus-Uncensored-Distilled-GGUF: 面向本地部署的轻量级创意与推理模型</u><br /><br /><i class="emoji"><b>🔞</b></i>可用于本地涩涩等场景<br /><br />• 基于 Qwen 3.5 9B，并融入 Claude Opus 4.6 蒸馏思路，主打更强的创意表达、对话表现与角色扮演场景<br /><br />• 提供 GGUF 与低显存友好的 Q4_K_M 量化版本，作者反馈在 RTX 3060 12 GB 上可达约 38 tok/s，适合本地聊天、游戏 NPC 与 Home Lab 部署<br /><br />• 模型采用 Apache 2.0 许可证，便于社区测试与二次集成<br /><br /><a href="https://www.reddit.com/r/LocalLLaMA/comments/1runlpf/qwen359bclaude46opusuncensoreddistilledgguf" target="_blank" rel="noopener" title="https://www.reddit.com/r/LocalLLaMA/comments/1runlpf/qwen359bclaude46opusuncensoreddistilledgguf">https://www.reddit.com/r/LocalLLaMA/comments/1runlpf/qwen359bclaude46opusuncensoreddistilledgguf</a><br /><br /><a href="/search/result?q=%23AI" title="#AI">#AI</a> <a href="/search/result?q=%23Uncensored" title="#Uncensored">#Uncensored</a> <a href="/search/result?q=%23%E6%9C%AC%E5%9C%B0%E5%A4%A7%E6%A8%A1%E5%9E%8B" title="#本地大模型">#本地大模型</a> <a href="/search/result?q=%23%E6%A8%A1%E5%9E%8B%E8%92%B8%E9%A6%8F" title="#模型蒸馏">#模型蒸馏</a> <a href="/search/result?q=%23GGUF" title="#GGUF">#GGUF</a> <a href="/search/result?q=%23Qwen" title="#Qwen">#Qwen</a> <a href="/search/result?q=%23Claude" title="#Claude">#Claude</a> <a href="/search/result?q=%23LMStudio" title="#LMStudio">#LMStudio</a> <a href="/search/result?q=%23%E9%87%8F%E5%8C%96%E6%A8%A1%E5%9E%8B" title="#量化模型">#量化模型</a> <a href="/search/result?q=%23%E4%BD%8E%E6%98%BE%E5%AD%98%E9%83%A8%E7%BD%B2" title="#低显存部署">#低显存部署</a> <a href="/search/result?q=%23%E8%A7%92%E8%89%B2%E6%89%AE%E6%BC%94" title="#角色扮演">#角色扮演</a> <a href="/search/result?q=%23LocalLLaMA" title="#LocalLLaMA">#LocalLLaMA</a><a class="tgme_widget_message_link_preview" href="https://www.reddit.com/r/LocalLLaMA/comments/1runlpf/qwen359bclaude46opusuncensoreddistilledgguf/" target="_blank" rel="noopener" title="Posted by EvilEnginer - 1,409 votes and 210 comments"> <div class="link_preview_site_name accent_color">Reddit</div> <img class="link_preview_image" alt="[Mature Content] From the LocalLLaMA community on Reddit: Qwen3.5-9B-Claude-4.6-Opus-Uncensored-Distilled-GGUF" src="/static/https://cdn4.telesco.pe/file/WMPwcy796bDBbgu8DKV-RoG1-Tewa3uJpHDcb3uqbjBkD4JLD2h3nCJBxzK_6lzquVBwuSeDnN4azCm3eIPL0PUDlgh5Ve7yH3oPeBpSk-LCTPb50B_r9u5VVpD_7edJdx5HkKhHovkKKNgjRMMsl6m1HxMQZ0EKPuFyFbh5kWcESDNJXksLoficEGqzMHNkRVIEo7IRtRGSuzc3qQenErSAvTzr8IJgI5vU3MjepS1uyuUrpdlhIgw38H1t6JWHDrYrtYjJNVybq33423SodicYXE0-nvRzoe8QspZ5A01RTlKWbHVrsse4uNWrPNM4MGTLBJT6-4AOt7nZQUIq5Q.jpg" width="1200" height="630" loading="eager" /> <div class="link_preview_title">[Mature Content] From the LocalLLaMA community on Reddit: Qwen3.5-9B-Claude-4.6-Opus-Uncensored-Distilled-GGUF</div> <div class="link_preview_description">Posted by EvilEnginer - 1,409 votes and 210 comments</div> </a>