GodsDevProject/GLM-4.7-Flash-Deployer
0
1<!DOCTYPE html>2<html lang="en">3<head>4 <meta charset="UTF-8">5 <meta name="viewport" content="width=device-width, initial-scale=1.0">6 <title>GLM-4.7 Flash Deployment Hub</title>7 <meta name="description" content="Deployment interface for GLM-4.7-Flash-Uncensored model">8 9 <!-- Tailwind CSS for Modern Styling -->10 <script src="https://cdn.tailwindcss.com"></script>11 12 <!-- FontAwesome for Icons -->13 <link rel="stylesheet" href="https://cdnjs.cloudflare.com/ajax/libs/font-awesome/6.4.0/css/all.min.css">14 15 <!-- Google Fonts -->16 <link href="https://fonts.googleapis.com/css2?family=Inter:wght@300;400;600;700&family=Fira+Code:wght@400;600&display=swap" rel="stylesheet">17 18 <style>19 :root {20 --primary-color: #6366f1;21 --accent-color: #8b5cf6;22 --dark-bg: #0f172a;23 --panel-bg: #1e293b;24 --text-color: #e2e8f0;25 --success-color: #10b981;26 --terminal-font: 'Fira Code', monospace;27 }28 29 body {30 background-color: var(--dark-bg);31 color: var(--text-color);32 font-family: 'Inter', sans-serif;33 line-height: 1.6;34 }35 36 .glass-panel {37 background: rgba(30, 41, 59, 0.7);38 backdrop-filter: blur(10px);39 border: 1px solid rgba(255, 255, 255, 0.1);40 box-shadow: 0 4px 6px -1px rgba(0, 0, 0, 0.5);41 }42 43 .text-gradient {44 background: linear-gradient(to right, #818cf8, #c084fc);45 -webkit-background-clip: text;46 -webkit-text-fill-color: transparent;47 }48 49 .code-block {50 font-family: var(--terminal-font);51 background-color: #000;52 border-radius: 8px;53 padding: 1rem;54 position: relative;55 overflow-x: auto;56 }57 58 .code-header {59 background-color: #2d3748;60 padding: 0.5rem 1rem;61 border-radius: 8px 8px 0 0;62 font-size: 0.85rem;63 color: #a0aec0;64 display: flex;65 justify-content: space-between;66 align-items: center;67 }68 69 .terminal-window {70 background-color: #0d1117;71 border: 1px solid #30363d;72 border-radius: 12px;73 overflow: hidden;74 box-shadow: 0 10px 15px -3px rgba(0, 0, 0, 0.7);75 }76 77 .terminal-header {78 background-color: #161b22;79 padding: 0.75rem;80 border-bottom: 1px solid #30363d;81 display: flex;82 gap: 8px;83 }84 85 .dot { width: 12px; height: 12px; border-radius: 50%; }86 .red { background-color: #ff5f56; }87 .yellow { background-color: #ffbd2e; }88 .green { background-color: #27c93f; }89 90 .terminal-body {91 padding: 1rem;92 font-family: var(--terminal-font);93 font-size: 0.9rem;94 color: #c9d1d9;95 min-height: 200px;96 }97 98 .cmd-text { color: #7ee787; }99 .path-text { color: #58a6ff; }100 .output-text { color: #d2a8ff; }101 .error-text { color: #ff7b72; }102 .success-text { color: #3fb950; }103 104 .loader {105 border: 4px solid #f3f3f3;106 border-top: 4px solid var(--primary-color);107 border-radius: 50%;108 width: 24px;109 height: 24px;110 animation: spin 1s linear infinite;111 }112 113 @keyframes spin {114 0% { transform: rotate(0deg); }115 100% { transform: rotate(360deg); }116 }117 118 .feature-card:hover {119 transform: translateY(-5px);120 transition: all 0.3s ease;121 }122 </style>123</head>124<body class="min-h-screen flex flex-col">125 126 <!-- Header / Navbar -->127 <nav class="glass-panel sticky top-0 z-50 border-b border-indigo-500/30">128 <div class="max-w-7xl mx-auto px-4 sm:px-6 lg:px-8">129 <div class="flex items-center justify-between h-16">130 <div class="flex items-center gap-3">131 <i class="fa-solid fa-robot text-2xl text-indigo-400"></i>132 <div>133 <h1 class="text-xl font-bold tracking-tight">134 GLM-4.7 <span class="text-indigo-400">Flash</span> Deployer135 </h1>136 <p class="text-xs text-gray-400">Uncensored Code & Chat Interface</p>137 </div>138 </div>139 140 <div class="hidden md:flex items-center gap-6">141 <a href="#dashboard" class="text-sm font-medium text-gray-300 hover:text-white transition">Dashboard</a>142 <a href="#install" class="text-sm font-medium text-gray-300 hover:text-white transition">Installation</a>143 <a href="#cloud" class="text-sm font-medium text-gray-300 hover:text-white transition">Cloud Options</a>144 <a href="https://huggingface.co/DavidAU/GLM-4.7-Flash-Uncensored-Heretic-NEO-CODE-Imatrix-MAX-GGUF" target="_blank" class="text-sm font-medium text-indigo-400 hover:text-indigo-300 flex items-center gap-2">145 <i class="fa-brands fa-huggingface"></i> Model Hub146 </a>147 </div>148 149 <!-- Hugging Face Link Requirement -->150 <a href="https://huggingface.co/spaces/akhaliq/anycoder" target="_blank" class="text-xs font-semibold text-gray-500 hover:text-indigo-400 transition flex items-center gap-1">151 <i class="fa-solid fa-link"></i> Built with anycoder152 </a>153 </div>154 </div>155 </nav>156 157 <!-- Main Content -->158 <main class="flex-grow container mx-auto px-4 py-8 space-y-12">159 160 <!-- Hero Section -->161 <section id="dashboard" class="text-center space-y-6 py-10">162 <div class="inline-block px-4 py-1 rounded-full bg-indigo-900/30 border border-indigo-500/30 text-indigo-300 text-sm font-medium mb-4">163 Model: GLM-4.7-Flash-Uncensored-Heretic-NEO-CODE164 </div>165 <h2 class="text-4xl md:text-6xl font-extrabold tracking-tight leading-tight">166 Run the <span class="text-gradient">Ultimate</span> Local AI167 </h2>168 <p class="text-xl text-gray-400 max-w-3xl mx-auto">169 A comprehensive dashboard to install, configure, and deploy the GLM-4.7 Flash model. 170 Optimized for GGUF quantization to run efficiently on consumer hardware or cloud instances.171 </p>172 173 <div class="flex flex-wrap justify-center gap-4 pt-4">174 <a href="#install" class="px-8 py-3 rounded-lg bg-indigo-600 hover:bg-indigo-700 text-white font-bold shadow-lg shadow-indigo-500/30 transition transform hover:scale-105 flex items-center gap-2">175 <i class="fa-solid fa-download"></i> Start Installation176 </a>177 <a href="#cloud" class="px-8 py-3 rounded-lg bg-gray-800 hover:bg-gray-700 text-white font-bold border border-gray-600 transition transform hover:scale-105 flex items-center gap-2">178 <i class="fa-solid fa-cloud"></i> Cloud Deployment179 </a>180 </div>181 </section>182 183 <!-- System Requirements -->184 <section class="grid md:grid-cols-3 gap-6">185 <div class="glass-panel p-6 rounded-xl feature-card">186 <div class="w-12 h-12 rounded-lg bg-blue-900/50 flex items-center justify-center mb-4 text-blue-400">187 <i class="fa-solid fa-memory text-2xl"></i>188 </div>189 <h3 class="text-lg font-bold mb-2">RAM Requirements</h3>190 <p class="text-sm text-gray-400">191 The <strong>Imatrix-MAX</strong> version is large. We recommend at least <span class="text-green-400 font-bold">32GB RAM</span> for smooth inference without swapping.192 </p>193 </div>194 <div class="glass-panel p-6 rounded-xl feature-card">195 <div class="w-12 h-12 rounded-lg bg-purple-900/50 flex items-center justify-center mb-4 text-purple-400">196 <i class="fa-solid fa-microchip text-2xl"></i>197 </div>198 <h3 class="text-lg font-bold mb-2">GPU Recommendation</h3>199 <p class="text-sm text-gray-400">200 An NVIDIA GPU with <span class="text-green-400 font-bold">8GB+ VRAM</span> (RTX 3060 or better) allows for acceleration via CUDA.201 </p>202 </div>203 <div class="glass-panel p-6 rounded-xl feature-card">204 <div class="w-12 h-12 rounded-lg bg-emerald-900/50 flex items-center justify-center mb-4 text-emerald-400">205 <i class="fa-solid fa-hard-drive text-2xl"></i>206 </div>207 <h3 class="text-lg font-bold mb-2">Storage</h3>208 <p class="text-sm text-gray-400">209 The GGUF model typically ranges between <span class="text-green-400 font-bold">20GB - 40GB</span> depending on the specific quantization (Q4_K_M vs Q8_0).210 </p>211 </div>212 </section>213 214 <!-- Interactive Installation Script Generator -->215 <section id="install" class="glass-panel rounded-2xl overflow-hidden">216 <div class="p-6 border-b border-gray-700 bg-gray-900/50 flex justify-between items-center">217 <div>218 <h3 class="text-xl font-bold flex items-center gap-2">219 <i class="fa-solid fa-code"></i> Automated Setup Script220 </h3>221 <p class="text-sm text-gray-400">Generates the command line instructions for KoboldCpp & Ollama</p>222 </div>223 <button onclick="copyToClipboard()" id="copyBtn" class="px-4 py-2 rounded bg-gray-700 hover:bg-gray-600 text-white text-sm transition flex items-center gap-2">224 <i class="fa-regular fa-copy"></i> Copy to Clipboard225 </button>226 </div>227 228 <div class="p-8 grid lg:grid-cols-2 gap-8">229 <!-- Configuration Form -->230 <div class="space-y-6">231 <h4 class="text-lg font-semibold text-indigo-400 border-b border-gray-700 pb-2">Configuration</h4>232 233 <div>234 <label class="block text-sm font-medium text-gray-300 mb-2">Execution Backend</label>235 <select id="backendSelect" class="w-full bg-gray-800 border border-gray-700 rounded-lg p-3 text-white focus:ring-2 focus:ring-indigo-500 focus:outline-none">236 <option value="kobold">KoboldCpp (Local Web UI)</option>237 <option value="ollama">Ollama (System Service)</option>238 <option value="cmd">Direct CLI (llama.cpp)</option>239 </select>240 </div>241 242 <div>243 <label class="block text-sm font-medium text-gray-300 mb-2">GPU Acceleration</label>244 <div class="flex gap-4">245 <label class="flex items-center gap-2 text-gray-300 cursor-pointer">246 <input type="radio" name="gpu" value="cpu" checked class="text-indigo-600 focus:ring-indigo-500">247 <span>CPU Only (Slower)</span>248 </label>249 <label class="flex items-center gap-2 text-gray-300 cursor-pointer">250 <input type="radio" name="gpu" value="cuda" class="text-indigo-600 focus:ring-indigo-500">251 <span>NVIDIA (CUDA)</span>252 </label>253 <label class="flex items-center gap-2 text-gray-300 cursor-pointer">254 <input type="radio" name="gpu" value="metal" class="text-indigo-600 focus:ring-indigo-500">255 <span>Apple Silicon (Metal)</span>256 </label>257 </div>258 </div>259 260 <div>261 <label class="block text-sm font-medium text-gray-300 mb-2">Threads (CPU Cores)</label>262 <input type="range" id="threadSlider" min="1" max="32" value="8" class="w-full h-2 bg-gray-700 rounded-lg appearance-none cursor-pointer accent-indigo-500">263 <div class="flex justify-between text-xs text-gray-500 mt-1">264 <span>1 Thread</span>265 <span id="threadValue" class="text-indigo-400 font-bold">8 Threads</span>266 <span>32 Threads</span>267 </div>268 </div>269 270 <div class="bg-indigo-900/20 p-4 rounded-lg border border-indigo-500/20">271 <h5 class="text-sm font-bold text-indigo-300 mb-2"><i class="fa-solid fa-circle-info"></i> Note on Imatrix-MAX</h5>272 <p class="text-xs text-gray-400">273 The "Imatrix-MAX" version implies an optimized quantization matrix. Ensure you download the specific <code>.gguf</code> file from the HuggingFace link provided. The script below assumes standard GGUF loading.274 </p>275 </div>276 </div>277 278 <!-- Code Output -->279 <div class="relative">280 <div class="terminal-window">281 <div class="terminal-header">282 <div class="dot red"></div>283 <div class="dot yellow"></div>284 <div class="dot green"></div>285 <span class="text-xs text-gray-400 ml-2">root@server:~/glm-deploy</span>286 </div>287 <div class="terminal-body" id="codeOutput">288 <div class="mb-2">289 <span class="path-text">~</span>290 <span class="text-white"> $ </span>291 <span class="cmd-text"># Initializing GLM-4.7 Flash Setup...</span>292 </div>293 <div id="scriptContent">294 <div class="mb-2">295 <span class="path-text">~</span>296 <span class="text-white"> $ </span>297 <span class="cmd-text">git clone https://github.com/ggerganov/llama.cpp</span>298 </div>299 <div class="mb-2">300 <span class="path-text">~</span>301 <span class="text-white"> $ </span>302 <span class="cmd-text">cd llama.cpp && make</span>303 </div>304 <div class="mb-2">305 <span class="path-text">~/llama.cpp</span>306 <span class="text-white"> $ </span>307 <span class="cmd-text">./llama-server</span>308 <span class="text-gray-500"> \</span>309 </div>310 <div class="mb-2 ml-4">311 <span class="text-gray-500"> --model ./models/glm-4.7-flash-uncensored.Q4_K_M.gguf</span>312 </div>313 <div class="mb-2 ml-4">314 <span class="text-gray-500"> --n_ctx 4096 --n_gpu_layers 35</span>315 </div>316 <div class="mb-2 ml-4">317 <span class="text-gray-500"> --port 8080 --host 0.0.0.0</span>318 </div>319 <div class="mt-4">320 <span class="success-text">[INFO] Server starting on http://localhost:8080</span>321 </div>322 </div>323 </div>324 </div>325 <div class="mt-4 text-xs text-gray-500 text-center">326 * Adjust <code>--n_gpu_layers</code> based on your VRAM capacity.327 </div>328 </div>329 </div>330 </section>331 332 <!-- Cloud Deployment Guide -->333 <section id="cloud" class="space-y-8">334 <div class="text-center">335 <h2 class="text-3xl font-bold">Cloud Deployment Options</h2>336 <p class="text-gray-400 mt-2">Running this model in the cloud requires instances with high RAM and VRAM.</p>337 </div>338 339 <div class="grid md:grid-cols-3 gap-6">340 <!-- Option 1: RunPod -->341 <div class="glass-panel p-6 rounded-xl border-t-4 border-pink-500 flex flex-col justify-between hover:border-pink-400 transition">342 <div>343 <div class="w-10 h-10 rounded bg-pink-900/30 flex items-center justify-center text-pink-400 mb-4">344 <i class="fa-solid fa-cloud-bolt"></i>345 </div>346 <h3 class="text-xl font-bold mb-2">RunPod</h3>347 <p class="text-sm text-gray-400 mb-4">348 Best for short bursts. Use an A100 or H100 pod. Upload your GGUF file to Pod Storage and run the server.349 </p>350 <div class="bg-gray-900/50 p-3 rounded text-xs font-mono text-pink-300 mb-4">351 nvidia-a100-80gb352 </div>353 <div class="flex items-center gap-2 text-xs text-gray-500">354 <i class="fa-solid fa-check text-green-500"></i> Pay per second355 </div>356 </div>357 <a href="https://runpod.io" target="_blank" class="mt-4 w-full text-center py-2 rounded bg-pink-600 hover:bg-pink-700 text-white text-sm font-medium transition">358 Launch Pod359 </a>360 </div>361 362 <!-- Option 2: Vast.ai -->363 <div class="glass-panel p-6 rounded-xl border-t-4 border-orange-500 flex flex-col justify-between hover:border-orange-400 transition">364 <div>365 <div class="w-10 h-10 rounded bg-orange-900/30 flex items-center justify-center text-orange-400 mb-4">366 <i class="fa-solid fa-server"></i>367 </div>368 <h3 class="text-xl font-bold mb-2">Vast.ai</h3>369 <p class="text-sm text-gray-400 mb-4">370 Marketplaces for GPU instances. Look for "RTX 4090" or "A100 80GB" instances. Very cost-effective.371 </p>372 <div class="bg-gray-900/50 p-3 rounded text-xs font-mono text-orange-300 mb-4">373 96GB RAM + 24GB VRAM374 </div>375 <div class="flex items-center gap-2 text-xs text-gray-500">376 <i class="fa-solid fa-check text-green-500"></i> Bid/On-demand options377 </div>378 </div>379 <a href="https://vast.ai" target="_blank" class="mt-4 w-full text-center py-2 rounded bg-orange-600 hover:bg-orange-700 text-white text-sm font-medium transition">380 Browse Instances381 </a>382 </div>383 384 <!-- Option 3: Lambda Labs -->385 <div class="glass-panel p-6 rounded-xl border-t-4 border-blue-500 flex flex-col justify-between hover:border-blue-400 transition">386 <div>387 <div class="w-10 h-10 rounded bg-blue-900/30 flex items-center justify-center text-blue-400 mb-4">388 <i class="fa-solid faRocket"></i>389 </div>390 <h3 class="text-xl font-bold mb-2">Lambda Labs</h3>391 <p class="text-sm text-gray-400 mb-4">392 User-friendly interface. Good for A10s and H100s. Easy to set up SSH and transfer models.393 </p>394 <div class="bg-gray-900/50 p-3 rounded text-xs font-mono text-blue-300 mb-4">395 H100 80GB or A100 40/80GB396 </div>397 <div class="flex items-center gap-2 text-xs text-gray-500">398 <i class="fa-solid fa-check text-green-500"></i> Flat rate pricing399 </div>400 </div>401 <a href="https://lambdalabs.com" target="_blank" class="mt-4 w-full text-center py-2 rounded bg-blue-600 hover:bg-blue-700 text-white text-sm font-medium transition">402 Get API Key403 </a>404 </div>405 </div>406 407 <div class="glass-panel p-6 rounded-xl bg-gradient-to-r from-gray-900 to-gray-800 border border-gray-700">408 <h3 class="text-lg font-bold text-white mb-4"><i class="fa-solid fa-list-check"></i> Quick Cloud Setup Checklist</h3>409 <ul class="space-y-3">410 <li class="flex items-start gap-3">411 <div class="mt-1 w-5 h-5 rounded-full bg-green-900 flex items-center justify-center text-green-400 text-xs"><i class="fa-solid fa-check"></i></div>412 <div>413 <span class="font-bold text-gray-200">Select Instance:</span> 414 <span class="text-gray-400 text-sm"> Ensure instance has > 32GB System RAM AND > 16GB VRAM for the Imatrix-MAX version.</span>415 </div>416 </li>417 <li class="flex items-start gap-3">418 <div class="mt-1 w-5 h-5 rounded-full bg-green-900 flex items-center justify-center text-green-400 text-xs"><i class="fa-solid fa-check"></i></div>419 <div>420 <span class="font-bold text-gray-200">Download Model:</span> 421 <span class="text-gray-400 text-sm"> Use `wget` or `huggingface-cli` on the cloud instance to download the GGUF file directly to storage.</span>422 </div>423 </li>424 <li class="flex items-start gap-3">425 <div class="mt-1 w-5 h-5 rounded-full bg-green-900 flex items-center justify-center text-green-400 text-xs"><i class="fa-solid fa-check"></i></div>426 <div>427 <span class="font-bold text-gray-200">Start Server:</span> 428 <span class="text-gray-400 text-sm"> Run the `llama-server` command with `--host 0.0.0.0` to allow external web access.</span>429 </div>430 </li>431 </ul>432 </div>433 </section>434 435 <!-- Footer -->436 <footer class="border-t border-gray-800 pt-8 pb-4 text-center">437 <div class="flex items-center justify-center gap-2 mb-4 opacity-70">438 <i class="fa-solid fa-robot text-indigo-500"></i>439 <span class="font-semibold text-gray-400">GLM-4.7 Flash Deployer</span>440 </div>441 <p class="text-sm text-gray-600">442 This tool is a UI helper for deploying open-source AI models. 443 Please ensure you comply with the model's license agreement.444 </p>445 <div class="mt-4 text-xs text-gray-700">446 © <span id="year"></span> AI Deployment Tools. All rights reserved.447 </div>448 </footer>449 450 </main>451 452 <!-- Logic -->453 <script>454 // Update Year455 document.getElementById('year').textContent = new Date().getFullYear();456 457 // Thread Slider Logic458 const slider = document.getElementById('threadSlider');459 const threadValue = document.getElementById('threadValue');460 461 slider.addEventListener('input', function() {462 threadValue.textContent = this.value + " Threads";463 updateScript();464 });465 466 // Backend Selection Logic467 const backendSelect = document.getElementById('backendSelect');468 469 backendSelect.addEventListener('change', updateScript);470 471 // GPU Selection Logic472 const gpuRadios = document.querySelectorAll('input[name="gpu"]');473 gpuRadios.forEach(radio => {474 radio.addEventListener('change', updateScript);475 });476 477 function updateScript() {478 const backend = backendSelect.value;479 const threads = slider.value;480 const gpu = document.querySelector('input[name="gpu"]:checked').value;481 const outputDiv = document.getElementById('scriptContent');482 483 let scriptHTML = '';484 485 if (backend === 'kobold') {486 scriptHTML = `487 <div class="mb-2">488 <span class="path-text">~</span>489 <span class="text-white"> $ </span>490 <span class="cmd-text">git clone https://github.com/LostRuins/koboldcpp.git</span>491 </div>492 <div class="mb-2">493 <span class="path-text">~</span>494 <span class="text-white"> $ </span>495 <span class="cmd-text">cd koboldcpp</span>496 </div>497 <div class="mb-2">498 <span class="path-text">~/koboldcpp</span>499 <span class="text-white"> $ </span>500 <span class="cmd-text">mkdir -p models</span>501 </div>502 <div class="mb-2">503 <span class="path-text">~/koboldcpp/models</span>504 <span class="text-white"> $ </span>505 <span class="cmd-text"># Place your GLM-4.7-Flash-Q4_K_M.gguf here</span>506 </div>507 <div class="mb-2">508 <span class="path-text">~/koboldcpp</span>509 <span class="text-white"> $ </span>510 <span class="cmd-text">python3 koboldcpp.py</span>511 <span class="text-gray-500"> \</span>512 </div>513 <div class="mb-2 ml-4">514 <span class="text-gray-500"> --model models/glm-4.7-flash-uncensored.Q4_K_M.gguf</span>515 </div>516 <div class="mb-2 ml-4">517 <span class="text-gray-500"> --contextsize ${threads * 512} --threads ${threads}</span>518 </div>519 <div class="mb-2 ml-4">520 <span class="text-gray-500"> --port 5001 --host 0.0.0.0</span>521 </div>522 <div class="mt-4">523 <span class="success-text">[INFO] KoboldAI UI ready at http://localhost:5001</span>524 </div>525 `;526 } 527 else if (backend === 'ollama') {528 scriptHTML = `529 <div class="mb-2">530 <span class="path-text">~</span>531 <span class="text-white"> $ </span>532 <span class="cmd-text">curl -fsSL https://ollama.com/install.sh | sh</span>533 </div>534 <div class="mb-2">535 <span class="path-text">~</span>536 <span class="text-white"> $ </span>537 <span class="cmd-text">ollama pull nomic-ai/gpt4all-falcon</span> 538 <span class="text-gray-500">/* Fallback if direct GLM pull fails */</span>539 </div>540 <div class="mb-2">541 <span class="path-text">~</span>542 <span class="text-white"> $ </span>543 <span class="cmd-text">ollama create glm-4-flash -f Modelfile</span>544 </div>545 <div class="mb-2 bg-gray-800 p-3 rounded border-l-4 border-yellow-500">546 <span class="text-xs text-gray-400 block mb-1"># Create a Modelfile:</span>547 <span class="text-xs text-white">FROM glm-4.7-flash-uncensored.Q4_K_M.gguf</span>548 <span class="text-xs text-white">PARAMETER num_thread ${threads}</span>549 </div>550 <div class="mt-4">551 <span class="success-text">[SUCCESS] Model loaded into Ollama. Run 'ollama run glm-4-flash'</span>552 </div>553 `;554 } 555 else {556 // Direct CLI557 scriptHTML = `558 <div class="mb-2">559 <span class="path-text">~</span>560 <span class="text-white"> $ </span>561 <span class="cmd-text">git clone https://github.com/ggerganov/llama.cpp</span>562 </div>563 <div class="mb-2">564 <span class="path-text">~</span>565 <span class="text-white"> $ </span>566 <span class="cmd-text">cd llama.cpp && make</span>567 </div>568 <div class="mb-2">569 <span class="path-text">~/llama.cpp</span>570 <span class="text-white"> $ </span>571 <span class="cmd-text">mkdir -p models</span>572 </div>573 <div class="mb-2">574 <span class="path-text">~/llama.cpp/models</span>575 <span class="text-white"> $ </span>576 <span class="cmd-text"># Download GLM-4.7-Flash-Q4_K_M.gguf here</span>577 </div>578 <div class="mb-2">579 <span class="path-text">~/llama.cpp</span>580 <span class="text-white"> $ </span>581 <span class="cmd-text">./main</span>582 <span class="text-gray-500"> \</span>583 </div>584 <div class="mb-2 ml-4">585 <span class="text-gray-500"> -m models/glm-4.7-flash-uncensored.Q4_K_M.gguf</span>586 </div>587 <div class="mb-2 ml-4">588 <span class="text-gray-500"> -n 512 --repeat_penalty 1.1</span>589 </div>590 <div class="mb-2 ml-4">591 <span class="text-gray-500"> -c ${threads * 512} -t ${threads} -b 1</span>592 </div>593 <div class="mt-4">594 <span class="success-text">[INFO] Entering interactive mode. Type 'exit' to quit.</span>595 </div>596 `;597 }598 599 // GPU Specific Flags600 if (gpu === 'cuda') {601 // Insert GPU flag logic roughly by appending to relevant lines or adding a new block602 // For simplicity in this template, we just update the text in the script603 const tempDiv = document.createElement('div');604 tempDiv.innerHTML = scriptHTML;605 // Simple replacement logic based on backend would be complex, so we stick to generic output606 // In a real app, we'd inject --n_gpu_layers based on backend607 if (backend === 'kobold') {608 scriptHTML = scriptHTML.replace('--port 5001', '--port 5001 --gpu_layer_count 35');609 } else if (backend === 'cmd') {610 scriptHTML = scriptHTML.replace('-t ' + threads, `-t ${threads} -ngl 35`);611 }612 }613 614 outputDiv.innerHTML = scriptHTML;615 }616 617 function copyToClipboard() {618 const codeElement = document.getElementById('scriptContent');619 const textToCopy = codeElement.innerText;620 621 navigator.clipboard.writeText(textToCopy).then(() => {622 const btn = document.getElementById('copyBtn');623 const originalText = btn.innerHTML;624 btn.innerHTML = '<i class="fa-solid fa-check"></i> Copied!';625 btn.classList.remove('bg-gray-700');626 btn.classList.add('bg-green-600');627 628 setTimeout(() => {629 btn.innerHTML = originalText;630 btn.classList.add('bg-gray-700');631 btn.classList.remove('bg-green-600');632 }, 3000);633 }).catch(err => {634 console.error('Failed to copy: ', err);635 });636 }637 </script>638</body>639</html>