mirror of
https://github.com/hanzoai/skills.git
synced 2026-11-19 00:47:36 +00:00
hanzoskills.com/llms.txt now opens with the install line, the API reference index and the company index, so an agent that lands on any one of the three can find the rest. robots.txt claimed 537 documents while llms.txt counts 920. The number is gone from robots.txt; llms.txt derives it. public/ is regenerated from scripts/site.py, which it had fallen behind: the copy-prompt menu and agent bar were in the generator and not in the committed pages. tests/site.mjs passes against the result. Co-authored-by: Hanzo Dev <[email protected]>
305 lines
24 KiB
HTML
305 lines
24 KiB
HTML
<!doctype html>
|
|
<html lang="en"><head>
|
|
<meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
|
|
<title>hip-0260</title>
|
|
<meta name="color-scheme" content="dark">
|
|
<style>
|
|
/* Committed to one visual world: a dark instrument panel. No light variant —
|
|
the dot field and the hairline rules only read on black, and a naive
|
|
inversion would wash both out. Every colour is painted explicitly so the page
|
|
never borrows a host background. */
|
|
@font-face{font-family:'Zen';src:url('/fonts/Zen-Variable.woff2') format('woff2');font-weight:100 900;font-style:normal;font-display:swap}
|
|
@font-face{font-family:'Zen';src:url('/fonts/Zen-Italic-Variable.woff2') format('woff2');font-weight:100 900;font-style:italic;font-display:swap}
|
|
@font-face{font-family:'Zen Mono';src:url('/fonts/ZenMono-Variable.woff2') format('woff2');font-weight:100 900;font-style:normal;font-display:swap}
|
|
@font-face{font-family:'Zen Mono';src:url('/fonts/ZenMono-Italic-Variable.woff2') format('woff2');font-weight:100 900;font-style:italic;font-display:swap}
|
|
@font-face{font-family:'Zen Pixel Square';src:url('/fonts/ZenPixel-Square.woff2') format('woff2');font-style:normal;font-display:swap}
|
|
|
|
:root{
|
|
--ink:#f4f4f2; --soft:#9c9c98; --faint:#6a6a66; --ground:#000; --panel:#0a0a0a;
|
|
--line:#232323; --line-hi:#3a3a3a; --accent:#f4f4f2;
|
|
--font-zen-sans:'Zen',-apple-system,BlinkMacSystemFont,"Segoe UI",Helvetica,Arial,sans-serif;
|
|
--font-zen-mono:'Zen Mono',ui-monospace,"SF Mono",SFMono-Regular,Menlo,Consolas,monospace;
|
|
--font-zen-pixel-square:'Zen Pixel Square',monospace;
|
|
--mono:var(--font-zen-mono);
|
|
--sans:var(--font-zen-sans);
|
|
/* Controls are pills and surfaces are soft, as on hanzo.ai. */
|
|
--r:999px; --r-soft:14px;
|
|
}
|
|
*{box-sizing:border-box}
|
|
html{-webkit-text-size-adjust:100%}
|
|
body{margin:0;background:var(--ground);color:var(--ink);font-family:var(--sans);
|
|
font-size:16px;line-height:1.65;-webkit-font-smoothing:antialiased}
|
|
.wrap{max-width:60rem;margin:0 auto;padding:0 1.75rem}
|
|
|
|
/* The dot field. A halftone made of one repeating radial-gradient, faded out
|
|
along the diagonal by a mask so it reads as a wave rather than wallpaper. */
|
|
/* No overflow clip here: the dot field is inset to the box already, and a clip
|
|
would cut off the copy-prompt menu. */
|
|
.hero{position:relative;border-bottom:1px solid var(--line)}
|
|
.hero::before{content:"";position:absolute;inset:0;
|
|
background-image:radial-gradient(var(--line-hi) 1px, transparent 1px);
|
|
background-size:7px 7px;
|
|
-webkit-mask-image:linear-gradient(115deg,#000 0%,rgba(0,0,0,.55) 38%,transparent 72%);
|
|
mask-image:linear-gradient(115deg,#000 0%,rgba(0,0,0,.55) 38%,transparent 72%);
|
|
pointer-events:none}
|
|
.hero .wrap{position:relative;padding-top:5.5rem;padding-bottom:4rem}
|
|
|
|
nav{border-bottom:1px solid var(--line)}
|
|
nav .in{max-width:60rem;margin:0 auto;padding:1.1rem 1.75rem;display:flex;
|
|
justify-content:space-between;align-items:center;gap:1rem}
|
|
.brand{font-family:var(--mono);font-size:.85rem;letter-spacing:.14em;
|
|
text-transform:uppercase;color:var(--ink);text-decoration:none;
|
|
display:inline-flex;align-items:center;gap:.6rem}
|
|
.brand i{font-style:normal;color:var(--faint)}
|
|
.mark{width:1.15em;height:1.15em;display:block;flex:0 0 auto}
|
|
nav a.up{font-family:var(--mono);font-size:.72rem;letter-spacing:.14em;
|
|
text-transform:uppercase;color:var(--soft);text-decoration:none}
|
|
nav a.up:hover{color:var(--ink)}
|
|
|
|
h1{font-family:var(--mono);font-size:clamp(2rem,5.4vw,3.4rem);line-height:1.12;
|
|
letter-spacing:-.02em;margin:0 0 1.1rem;font-weight:500;text-wrap:balance}
|
|
h1 em{font-style:normal;color:var(--faint)}
|
|
.lede{color:var(--soft);font-size:1.05rem;max-width:52ch;margin:0 0 2.25rem}
|
|
|
|
.btns{display:flex;flex-wrap:wrap;gap:.7rem;margin:0 0 .5rem;align-items:center}
|
|
.btn{display:inline-flex;align-items:center;gap:.5rem;font-family:var(--mono);
|
|
font-size:.72rem;letter-spacing:.14em;text-transform:uppercase;
|
|
padding:.8rem 1.4rem;border:1px solid var(--line-hi);border-radius:var(--r);
|
|
color:var(--ink);text-decoration:none;background:transparent;cursor:pointer;
|
|
line-height:1.2;font-weight:400}
|
|
.btn:hover{background:#141414;text-decoration:none;border-color:var(--soft)}
|
|
.btn.solid{background:var(--ink);color:#000;border-color:var(--ink)}
|
|
.btn.solid:hover{background:#d8d8d4}
|
|
.btn .chev{width:.7em;height:.7em;flex:0 0 auto;transition:transform .15s ease}
|
|
.btn[aria-expanded="true"] .chev{transform:rotate(180deg)}
|
|
|
|
/* Copy-prompt selector. Each row copies the prompt for its own target. */
|
|
.copy{position:relative;display:inline-flex}
|
|
.menu{position:absolute;top:calc(100% + .5rem);left:0;z-index:20;min-width:21rem;
|
|
background:var(--panel);border:1px solid var(--line-hi);border-radius:var(--r-soft);
|
|
padding:.4rem;display:flex;flex-direction:column;
|
|
box-shadow:0 12px 32px rgba(0,0,0,.75)}
|
|
.menu[hidden]{display:none}
|
|
.mi{display:flex;align-items:center;gap:.7rem;width:100%;text-align:left;
|
|
padding:.62rem .8rem;border:0;border-radius:calc(var(--r-soft) - .35rem);
|
|
background:transparent;color:var(--ink);font-family:var(--mono);font-size:.76rem;
|
|
letter-spacing:.04em;text-decoration:none;cursor:pointer;line-height:1.3;
|
|
white-space:nowrap}
|
|
.mi:hover,.mi:focus-visible{background:#181818;outline:none}
|
|
.mi .mark{width:1.05rem;height:1.05rem;color:var(--soft)}
|
|
.mi:hover .mark{color:var(--ink)}
|
|
.mi u{margin-left:auto;color:var(--faint);text-decoration:none;font-size:.68rem;
|
|
letter-spacing:.1em;text-transform:uppercase}
|
|
|
|
/* Works-with bar: one scrolling row, name and mark per agent. */
|
|
.strip{border-bottom:1px solid var(--line);background:var(--panel)}
|
|
.strip .in{max-width:60rem;margin:0 auto;padding:.9rem 1.75rem;display:flex;
|
|
align-items:center;gap:1.25rem;font-family:var(--mono);font-size:.72rem;
|
|
letter-spacing:.1em;text-transform:uppercase;color:var(--faint)}
|
|
.strip b{color:var(--soft);font-weight:400;flex:0 0 auto;
|
|
border-right:1px solid var(--line);padding-right:1.25rem;align-self:stretch;
|
|
display:flex;align-items:center}
|
|
.bar{display:flex;align-items:center;gap:.5rem;flex:1 1 auto;min-width:0;
|
|
overflow-x:auto;scrollbar-width:none;scroll-snap-type:x proximity;
|
|
-webkit-overflow-scrolling:touch;
|
|
-webkit-mask-image:linear-gradient(90deg,#000 calc(100% - 3rem),transparent);
|
|
mask-image:linear-gradient(90deg,#000 calc(100% - 3rem),transparent)}
|
|
.bar::-webkit-scrollbar{display:none}
|
|
.agent{display:inline-flex;align-items:center;gap:.45rem;flex:0 0 auto;
|
|
padding:.4rem .85rem;border:1px solid var(--line);border-radius:var(--r);
|
|
color:var(--soft);white-space:nowrap;scroll-snap-align:start;
|
|
text-decoration:none;background:var(--ground)}
|
|
.agent:hover{color:var(--ink);border-color:var(--line-hi);text-decoration:none}
|
|
.agent .mark{width:.95rem;height:.95rem}
|
|
|
|
h2{font-family:var(--mono);font-size:1.1rem;letter-spacing:-.01em;margin:2.75rem 0 .75rem;font-weight:500}
|
|
h3{font-size:1rem;margin:2rem 0 .5rem;font-weight:600}
|
|
h4,h5,h6{font-size:.94rem;margin:1.5rem 0 .5rem;font-weight:600}
|
|
p{margin:0 0 1rem;max-width:70ch}
|
|
ul,ol{margin:0 0 1rem;padding-left:1.35rem;max-width:70ch}
|
|
li{margin:.22rem 0}
|
|
a{color:var(--ink);text-decoration:underline;text-underline-offset:3px;
|
|
text-decoration-color:var(--line-hi)}
|
|
a:hover{text-decoration-color:var(--ink)}
|
|
a:focus-visible{outline:1px solid var(--ink);outline-offset:3px}
|
|
code{font-family:var(--mono);font-size:.85em;background:var(--panel);
|
|
border:1px solid var(--line);border-radius:5px;padding:.1em .36em}
|
|
pre{background:var(--panel);border:1px solid var(--line);padding:1rem 1.1rem;
|
|
border-radius:var(--r-soft);overflow-x:auto;margin:0 0 1.25rem}
|
|
pre code{background:none;border:0;padding:0;font-size:.82rem;line-height:1.7}
|
|
blockquote{margin:0 0 1rem;padding-left:1rem;border-left:1px solid var(--line-hi);color:var(--soft)}
|
|
hr{border:0;border-top:1px solid var(--line);margin:2.5rem 0}
|
|
table{border-collapse:collapse;width:100%;margin:0 0 1.25rem;font-size:.92rem;
|
|
display:block;overflow-x:auto}
|
|
th,td{border:1px solid var(--line);padding:.5rem .7rem;text-align:left}
|
|
|
|
.curl{display:block;background:var(--panel);border:1px solid var(--line);
|
|
border-radius:var(--r-soft);padding:.95rem 1.1rem;font-family:var(--mono);
|
|
font-size:.8rem;color:var(--soft);overflow-x:auto;margin:0 0 2.5rem;
|
|
white-space:pre;line-height:1.9}
|
|
.curl b{color:var(--ink);font-weight:400}
|
|
.curl i{font-style:normal;color:var(--faint)}
|
|
|
|
/* index */
|
|
#q{width:100%;padding:.85rem 1.2rem;font-family:var(--mono);font-size:.85rem;color:var(--ink);
|
|
background:var(--panel);border:1px solid var(--line);border-radius:var(--r);margin:0 0 .6rem}
|
|
#q::placeholder{color:var(--faint);text-transform:uppercase;letter-spacing:.1em;font-size:.72rem}
|
|
#q:focus{outline:none;border-color:var(--soft)}
|
|
#count{color:var(--faint);font-family:var(--mono);font-size:.72rem;letter-spacing:.1em;
|
|
text-transform:uppercase;margin:0 0 2rem}
|
|
.cat{margin:2.5rem 0 .5rem;font-family:var(--mono);font-size:.7rem;letter-spacing:.16em;
|
|
text-transform:uppercase;color:var(--faint);border-bottom:1px solid var(--line);
|
|
padding-bottom:.45rem}
|
|
.row{display:block;padding:.6rem .9rem;margin:0 -.9rem;border-bottom:1px solid var(--line);
|
|
border-radius:var(--r-soft);text-decoration:none}
|
|
.row:hover{background:#0a0a0a}
|
|
.nm{font-family:var(--mono);font-size:.84rem;color:var(--ink);display:block}
|
|
.ds{color:var(--soft);font-size:.9rem;display:block;margin-top:.12rem}
|
|
footer{margin-top:4rem;padding:2rem 0 4rem;border-top:1px solid var(--line);
|
|
color:var(--faint);font-size:.85rem}
|
|
footer a{color:var(--soft)}
|
|
.fmap{display:grid;gap:1.75rem;grid-template-columns:repeat(auto-fit,minmax(11rem,1fr));
|
|
margin:0 0 2rem;font-size:.85rem}
|
|
.fmap b{display:block;font-family:var(--mono);font-size:.68rem;letter-spacing:.16em;
|
|
text-transform:uppercase;color:var(--faint);font-weight:400;margin:0 0 .6rem;
|
|
border-bottom:1px solid var(--line);padding-bottom:.4rem}
|
|
.fmap a{display:inline-block;color:var(--ink);text-decoration:none;margin:.3rem .35rem 0 0}
|
|
.fmap a:hover{text-decoration:underline}
|
|
.fmap div{color:var(--faint);line-height:1.5}
|
|
.fend{border-top:1px solid var(--line);padding-top:1.25rem;font-size:.82rem}
|
|
|
|
@media (max-width:640px){
|
|
.hero .wrap{padding-top:3.25rem;padding-bottom:2.75rem}
|
|
.btns{gap:.5rem}
|
|
.btn{padding:.75rem 1.15rem;font-size:.68rem}
|
|
.copy,.copy .btn{width:100%;justify-content:center}
|
|
.menu{left:0;right:0;min-width:0}
|
|
.strip .in{flex-direction:column;align-items:flex-start;gap:.6rem;padding:.9rem 1.75rem}
|
|
.strip b{border-right:0;padding-right:0}
|
|
.bar{width:100%}
|
|
}
|
|
</style>
|
|
<link rel="alternate" type="text/markdown" href="/hip-0260.md">
|
|
|
|
<link rel="icon" href="/favicon.svg">
|
|
<link rel="canonical" href="https://hanzoskills.com/hip-0260">
|
|
</head>
|
|
<!-- Reading this with a machine? Every page here is also plain markdown.
|
|
This one: https://hanzoskills.com/hip-0260.md
|
|
Index: https://hanzoskills.com/llms.txt
|
|
Corpus: https://hanzoskills.com/all.md
|
|
Append .md to any path for the markdown twin. -->
|
|
<body>
|
|
<nav><div class="in">
|
|
<a class="brand" href="/"><svg class="mark" viewBox="0 0 67 67" fill="currentColor" aria-hidden="true"><path d="M22.21 67V44.6369H0V67H22.21Z"/><path d="M66.7038 22.3184H22.2534L0.0878906 44.6367H44.4634L66.7038 22.3184Z"/><path d="M22.21 0H0V22.3184H22.21V0Z"/><path d="M66.7198 0H44.5098V22.3184H66.7198V0Z"/><path d="M66.7198 67V44.6369H44.5098V67H66.7198Z"/></svg> HANZO<i>SKILLS</i></a>
|
|
<a class="up" href="/">All skills</a>
|
|
</div></nav>
|
|
|
|
<div class="wrap">
|
|
<h1>hip-0260</h1><p class="lede">HIP-260: Efficient Model Practices. Status Living. Hanzo architectural specification.</p><h1>HIP-0260: Efficient Model Practices</h1>
|
|
<h2>Abstract</h2>
|
|
<p>This HIP establishes best practices for developing and deploying energy-efficient AI models at Hanzo AI. It covers architecture decisions, training optimizations, inference efficiency, and operational practices that reduce environmental impact while maintaining capability.</p>
|
|
<h2>Efficiency Principles</h2>
|
|
<h3>Guiding Principles</h3>
|
|
<ol><li><strong>Right-size models</strong>: Use smallest model that meets requirements</li><li><strong>Optimize before scale</strong>: Efficiency improvements before scaling compute</li><li><strong>Measure continuously</strong>: Track energy and efficiency metrics</li><li><strong>Share learnings</strong>: Document and share efficiency improvements</li><li><strong>Balance trade-offs</strong>: Consider efficiency in capability decisions</li></ol>
|
|
<h3>Efficiency Hierarchy</h3>
|
|
<pre><code>1. Avoid unnecessary computation
|
|
2. Reduce computation needed
|
|
3. Make computation more efficient
|
|
4. Use clean energy for remaining computation
|
|
5. Offset residual emissions</code></pre>
|
|
<h2>Training Efficiency</h2>
|
|
<h3>Architecture Design</h3>
|
|
<h4>Efficient Architectures</h4>
|
|
<p>| Technique | Benefit | Trade-off | |-----------|---------|-----------| | <strong>Sparse attention</strong> | O(n) vs O(n²) | Some quality loss | | <strong>Linear attention</strong> | Lower complexity | Limited context | | <strong>Mixture of Experts</strong> | Conditional compute | Complexity | | <strong>Parameter sharing</strong> | Smaller models | Some quality loss |</p>
|
|
<h4>Architecture Guidelines</h4>
|
|
<p>| Guideline | Rationale | |-----------|-----------| | Start small | Prove approach before scaling | | Test efficiency | Measure before committing | | Consider alternatives | Evaluate efficient variants | | Document choices | Record efficiency trade-offs |</p>
|
|
<h3>Training Optimizations</h3>
|
|
<h4>Compute Efficiency</h4>
|
|
<p>| Technique | Implementation | Benefit | |-----------|----------------|---------| | <strong>Mixed precision (BF16/FP16)</strong> | Default for all training | 2x memory, ~1.5x speed | | <strong>Gradient checkpointing</strong> | For memory-limited | 3-4x memory reduction | | <strong>Flash Attention</strong> | Default for transformers | 2-4x attention speedup | | <strong>Fused kernels</strong> | Use optimized libraries | 10-30% speedup |</p>
|
|
<h4>Data Efficiency</h4>
|
|
<p>| Technique | Implementation | Benefit | |-----------|----------------|---------| | <strong>Data deduplication</strong> | Preprocessing | Better quality per token | | <strong>Quality filtering</strong> | Curation pipeline | Fewer tokens needed | | <strong>Curriculum learning</strong> | Easy to hard | Faster convergence | | <strong>Active learning</strong> | Targeted data collection | Less data needed |</p>
|
|
<h4>Training Strategies</h4>
|
|
<p>| Strategy | Implementation | Benefit | |----------|----------------|---------| | <strong>Learning rate scheduling</strong> | Cosine with warmup | Faster convergence | | <strong>Early stopping</strong> | Validation monitoring | Avoid overtraining | | <strong>Checkpoint averaging</strong> | Average best checkpoints | Better final model | | <strong>Hyperparameter tuning</strong> | Systematic search | Optimal efficiency |</p>
|
|
<h3>Training Process Requirements</h3>
|
|
<h4>Pre-Training Checklist</h4>
|
|
<p>| Item | Verification | |------|--------------| | ☐ Baseline efficiency established | Measured baseline metrics | | ☐ Efficiency techniques applied | All applicable techniques | | ☐ Hardware utilization planned | GPU utilization >80% target | | ☐ Energy tracking configured | Monitoring in place |</p>
|
|
<h4>During Training</h4>
|
|
<p>| Metric | Target | Action if Below | |--------|--------|-----------------| | GPU utilization | >80% | Optimize batching | | Memory utilization | >70% | Adjust batch size | | Training loss curve | Expected descent | Investigate, adjust |</p>
|
|
<h2>Inference Efficiency</h2>
|
|
<h3>Model Optimization</h3>
|
|
<h4>Quantization</h4>
|
|
<p>| Level | Format | Use Case | Quality Impact | |-------|--------|----------|----------------| | <strong>FP16</strong> | Half precision | Default deployment | Minimal | | <strong>INT8</strong> | 8-bit integer | Production | 0-2% quality loss | | <strong>INT4</strong> | 4-bit integer | Edge/cost-sensitive | 2-5% quality loss | | <strong>GPTQ/AWQ</strong> | Advanced quant | Best quality at low bits | <2% typically |</p>
|
|
<h4>Model Compression</h4>
|
|
<p>| Technique | Reduction | Quality Impact | |-----------|-----------|----------------| | <strong>Pruning</strong> | 30-50% parameters | 1-3% quality loss | | <strong>Knowledge distillation</strong> | 2-10x smaller | Variable | | <strong>Low-rank factorization</strong> | 20-40% reduction | 1-2% quality loss |</p>
|
|
<h3>Inference Optimizations</h3>
|
|
<h4>Batching</h4>
|
|
<p>| Strategy | Use Case | Benefit | |----------|----------|---------| | <strong>Dynamic batching</strong> | API serving | Better utilization | | <strong>Continuous batching</strong> | LLM serving | Higher throughput | | <strong>Request coalescing</strong> | Similar requests | Efficiency gain |</p>
|
|
<h4>Caching</h4>
|
|
<p>| Cache Type | Implementation | Benefit | |------------|----------------|---------| | <strong>KV cache</strong> | Standard for LLMs | Required for efficiency | | <strong>Response cache</strong> | Exact match cache | Avoid recomputation | | <strong>Semantic cache</strong> | Similar query cache | Reduce redundant work |</p>
|
|
<h4>Speculative Decoding</h4>
|
|
<p>| Technique | Implementation | Benefit | |-----------|----------------|---------| | <strong>Draft model</strong> | Small model proposes | 2-3x speedup | | <strong>Self-speculative</strong> | Same model, different depth | 1.5-2x speedup | | <strong>Medusa heads</strong> | Multiple prediction heads | 2-3x speedup |</p>
|
|
<h3>Serving Infrastructure</h3>
|
|
<h4>Request Routing</h4>
|
|
<p>| Strategy | Implementation | Benefit | |----------|----------------|---------| | <strong>Model selection</strong> | Route to appropriate model | Use smallest sufficient | | <strong>Complexity estimation</strong> | Assess request complexity | Match model to need | | <strong>Load balancing</strong> | Efficient distribution | Better utilization |</p>
|
|
<h4>Scaling</h4>
|
|
<p>| Strategy | Implementation | Benefit | |----------|----------------|---------| | <strong>Horizontal scaling</strong> | Add instances | Handle load | | <strong>Vertical scaling</strong> | Better hardware | Efficiency per request | | <strong>Auto-scaling</strong> | Demand-based | Avoid idle compute |</p>
|
|
<h2>Operational Efficiency</h2>
|
|
<h3>Compute Scheduling</h3>
|
|
<h4>Time-Based Scheduling</h4>
|
|
<p>| Strategy | Implementation | Benefit | |----------|----------------|---------| | <strong>Off-peak training</strong> | Schedule for low-carbon hours | Lower emissions | | <strong>Batch processing</strong> | Aggregate non-urgent work | Better utilization | | <strong>Preemptible instances</strong> | Use spot/preemptible | Lower cost/emissions |</p>
|
|
<h4>Location-Based Scheduling</h4>
|
|
<p>| Strategy | Implementation | Benefit | |----------|----------------|---------| | <strong>Green region preference</strong> | Route to clean grids | Lower emissions | | <strong>Carbon-aware scheduling</strong> | Real-time carbon intensity | Optimal timing | | <strong>Follow-the-sun</strong> | Move work to clean regions | Maximize renewables |</p>
|
|
<h3>Hardware Efficiency</h3>
|
|
<h4>Hardware Selection</h4>
|
|
<p>| Factor | Consideration | |--------|---------------| | <strong>Latest generation</strong> | 20-50% efficiency gain per generation | | <strong>Right-sized</strong> | Match hardware to workload | | <strong>Utilization</strong> | Shared resources where appropriate |</p>
|
|
<h4>Hardware Lifecycle</h4>
|
|
<p>| Practice | Implementation | |----------|----------------| | <strong>Refresh cycles</strong> | Plan efficient hardware upgrades | | <strong>Utilization targets</strong> | Maintain >70% average utilization | | <strong>End-of-life</strong> | Responsible recycling/resale |</p>
|
|
<h3>Development Practices</h3>
|
|
<h4>Experiment Efficiency</h4>
|
|
<p>| Practice | Implementation | |----------|----------------| | <strong>Small-scale first</strong> | Test on small data/models first | | <strong>Ablation studies</strong> | Systematic, efficient experiments | | <strong>Negative result tracking</strong> | Avoid repeating failed experiments | | <strong>Experiment tracking</strong> | Log all runs to avoid duplicates |</p>
|
|
<h4>Code Efficiency</h4>
|
|
<p>| Practice | Implementation | |----------|----------------| | <strong>Profiling</strong> | Identify bottlenecks | | <strong>Optimized libraries</strong> | Use best implementations | | <strong>Batch operations</strong> | Vectorize where possible | | <strong>Memory management</strong> | Avoid unnecessary allocations |</p>
|
|
<h2>Metrics & Monitoring</h2>
|
|
<h3>Efficiency Metrics</h3>
|
|
<h4>Training Metrics</h4>
|
|
<p>| Metric | Definition | Target | |--------|------------|--------| | <strong>FLOPS/token</strong> | Compute per token | Track and reduce | | <strong>Samples/GPU-hour</strong> | Training throughput | Maximize | | <strong>GPU utilization</strong> | Compute usage | >80% | | <strong>Time to result</strong> | Training duration | Minimize |</p>
|
|
<h4>Inference Metrics</h4>
|
|
<p>| Metric | Definition | Target | |--------|------------|--------| | <strong>Tokens/second/GPU</strong> | Throughput | Maximize | | <strong>Latency (p50, p99)</strong> | Response time | Per SLA | | <strong>CO2e/1K tokens</strong> | Carbon intensity | Minimize | | <strong>Requests/watt</strong> | Energy efficiency | Track and improve |</p>
|
|
<h3>Monitoring Dashboard</h3>
|
|
<p>| Panel | Contents | |-------|----------| | <strong>Efficiency overview</strong> | Key efficiency metrics | | <strong>Training efficiency</strong> | Current training jobs | | <strong>Inference efficiency</strong> | Serving metrics | | <strong>Carbon intensity</strong> | Real-time carbon metrics | | <strong>Trends</strong> | Efficiency over time |</p>
|
|
<h3>Reporting</h3>
|
|
<p>| Report | Frequency | Contents | |--------|-----------|----------| | <strong>Efficiency digest</strong> | Weekly | Key metrics, anomalies | | <strong>Optimization opportunities</strong> | Monthly | Identified improvements | | <strong>Efficiency review</strong> | Quarterly | Progress, initiatives |</p>
|
|
<h2>Implementation Requirements</h2>
|
|
<h3>New Model Development</h3>
|
|
<p>| Phase | Efficiency Requirement | |-------|------------------------| | <strong>Design</strong> | Efficiency consideration in architecture | | <strong>Training</strong> | Efficiency techniques applied | | <strong>Evaluation</strong> | Efficiency metrics measured | | <strong>Deployment</strong> | Optimization before deployment |</p>
|
|
<h3>Model Deployment</h3>
|
|
<p>| Requirement | Verification | |-------------|--------------| | Quantization evaluated | Documented quality vs. efficiency trade-off | | Serving optimized | Batching, caching implemented | | Monitoring configured | Efficiency metrics tracked | | Right-sized deployment | Hardware matches workload |</p>
|
|
<h3>Continuous Improvement</h3>
|
|
<p>| Activity | Frequency | |----------|-----------| | Efficiency benchmarking | Monthly | | Technique evaluation | Quarterly | | Hardware assessment | Annual | | Process review | Annual |</p>
|
|
<h2>Related HIPs</h2>
|
|
<ul><li><strong>HIP-200</strong>: Responsible AI Principles</li><li><strong>HIP-250</strong>: Sustainability Standards Alignment</li><li><strong>HIP-251</strong>: AI Compute Carbon Footprint</li><li><strong>HIP-270</strong>: AI Supply Chain Responsibility</li></ul>
|
|
<h2>Changelog</h2>
|
|
<p>| Version | Date | Changes | |---------|------|---------| | 1.0 | 2025-12-17 | Initial draft |</p>
|
|
<h2>Copyright</h2>
|
|
<p>Copyright and related rights waived via <a href="https://creativecommons.org/publicdomain/zero/1.0/">CC0</a>.</p>
|
|
<footer>
|
|
<div class="fmap">
|
|
<div><b>Build</b>
|
|
<a href="https://hanzo.app">hanzo.app</a> app builder
|
|
<a href="https://hanzo.codes">hanzo.codes</a> code with an agent
|
|
<a href="https://hanzoskills.com">hanzoskills.com</a> what an agent reads</div>
|
|
<div><b>Run</b>
|
|
<a href="https://console.hanzo.ai">console.hanzo.ai</a> cloud console
|
|
<a href="https://api.hanzo.ai">api.hanzo.ai</a> the API</div>
|
|
<div><b>Account</b>
|
|
<a href="https://hanzo.id">hanzo.id</a> sign in
|
|
<a href="https://pay.hanzo.ai">pay.hanzo.ai</a> top up
|
|
<a href="https://billing.hanzo.ai">billing.hanzo.ai</a> invoices and usage</div>
|
|
<div><b>Company</b>
|
|
<a href="https://hanzo.ai">hanzo.ai</a> canonical
|
|
<a href="https://hanzo.ventures">hanzo.ventures</a> ventures</div>
|
|
</div>
|
|
<div class="fend"><a href="/hip-0260.md">This page as plain text</a> · <a href="/llms.txt">Index for agents</a></div>
|
|
</footer>
|
|
</div></body></html> |