Space: 1.03 card caption - shared protocol metrics for step 7150
Browse files- index.html +1 -1
index.html
CHANGED
|
@@ -466,7 +466,7 @@ tbody tr:hover { background: #19291f80; }
|
|
| 466 |
<div class="model-card__head"><div><p class="section-kicker">Slayer1B</p><h3 class="model-name">1.03</h3></div><span class="status-badge status-released"><span class="l-en" lang="en">Released</span><span class="l-pl" lang="pl">Wydany</span></span></div>
|
| 467 |
<p class="model-description"><span class="l-en" lang="en">A completed base-model release. Its older lm-eval report is separate from the common comparison.</span><span class="l-pl" lang="pl">Zakończone wydanie modelu bazowego. Starszy raport lm-eval jest oddzielony od wspólnego porównania.</span></p>
|
| 468 |
<dl class="model-facts"><div><dt><span class="l-en" lang="en">Checkpoint</span><span class="l-pl" lang="pl">Checkpoint</span></dt><dd>7150</dd></div><div><dt><span class="l-en" lang="en">Training tokens</span><span class="l-pl" lang="pl">Tokeny treningu</span></dt><dd><span class="l-en" lang="en">7.5 B</span><span class="l-pl" lang="pl">7,5 mld</span></dd></div></dl>
|
| 469 |
-
<dl class="model-metrics"><div><dt class="metric-label">BLiMP EN</dt><dd class="metric-value" data-eval-model="1.03" data-eval-step="7150" data-eval-field="blimp"><span class="l-en" lang="en">—</span><span class="l-pl" lang="pl">—</span></dd></div><div><dt class="metric-label">ARC-Easy EN</dt><dd class="metric-value" data-eval-model="1.03" data-eval-step="7150" data-eval-field="arc_easy"><span class="l-en" lang="en">—</span><span class="l-pl" lang="pl">—</span></dd></div><div><dt class="metric-label">BytePPL ↓</dt><dd class="metric-value" data-eval-model="1.03" data-eval-step="7150" data-eval-field="wikitext2_byte_ppl"><span class="l-en" lang="en">—</span><span class="l-pl" lang="pl">—</span></dd></div></dl><p class="method-note"><span class="l-en" lang="en">
|
| 470 |
<div class="model-links"><a href="https://huggingface.co/SlayerLab/Slayer1B-1.03/tree/main"><span class="l-en" lang="en">Weight files</span><span class="l-pl" lang="pl">Pliki wag</span></a><a href="https://huggingface.co/SlayerLab/Slayer1B-1.03-checkpoints"><span class="l-en" lang="en">Checkpoints</span><span class="l-pl" lang="pl">Checkpointy</span></a><a href="https://huggingface.co/datasets/SlayerLab/slayer-pl-8x3b"><span class="l-en" lang="en">Data</span><span class="l-pl" lang="pl">Dane</span></a><a href="https://huggingface.co/SlayerLab/Slayer1B-1.03"><span class="l-en" lang="en">Model card</span><span class="l-pl" lang="pl">Karta</span></a><a href="https://track.fabryka.ai/run/7ca4f84f-3b07-4525-8cde-f3a936dc3425"><span class="l-en" lang="en">Run</span><span class="l-pl" lang="pl">Bieg</span></a></div></article><article class="model-card" data-model-card data-version="1.02" data-status="released" data-search="slayer1b 1.02 released wydany base polish polski">
|
| 471 |
<div class="model-card__head"><div><p class="section-kicker">Slayer1B</p><h3 class="model-name">1.02</h3></div><span class="status-badge status-released"><span class="l-en" lang="en">Released</span><span class="l-pl" lang="pl">Wydany</span></span></div>
|
| 472 |
<p class="model-description"><span class="l-en" lang="en">Continuation at a constant learning rate, without a final decay.</span><span class="l-pl" lang="pl">Kontynuacja przy stałym tempie uczenia, bez końcowego wygaszania.</span></p>
|
|
|
|
| 466 |
<div class="model-card__head"><div><p class="section-kicker">Slayer1B</p><h3 class="model-name">1.03</h3></div><span class="status-badge status-released"><span class="l-en" lang="en">Released</span><span class="l-pl" lang="pl">Wydany</span></span></div>
|
| 467 |
<p class="model-description"><span class="l-en" lang="en">A completed base-model release. Its older lm-eval report is separate from the common comparison.</span><span class="l-pl" lang="pl">Zakończone wydanie modelu bazowego. Starszy raport lm-eval jest oddzielony od wspólnego porównania.</span></p>
|
| 468 |
<dl class="model-facts"><div><dt><span class="l-en" lang="en">Checkpoint</span><span class="l-pl" lang="pl">Checkpoint</span></dt><dd>7150</dd></div><div><dt><span class="l-en" lang="en">Training tokens</span><span class="l-pl" lang="pl">Tokeny treningu</span></dt><dd><span class="l-en" lang="en">7.5 B</span><span class="l-pl" lang="pl">7,5 mld</span></dd></div></dl>
|
| 469 |
+
<dl class="model-metrics"><div><dt class="metric-label">BLiMP EN</dt><dd class="metric-value" data-eval-model="1.03" data-eval-step="7150" data-eval-field="blimp"><span class="l-en" lang="en">—</span><span class="l-pl" lang="pl">—</span></dd></div><div><dt class="metric-label">ARC-Easy EN</dt><dd class="metric-value" data-eval-model="1.03" data-eval-step="7150" data-eval-field="arc_easy"><span class="l-en" lang="en">—</span><span class="l-pl" lang="pl">—</span></dd></div><div><dt class="metric-label">BytePPL ↓</dt><dd class="metric-value" data-eval-model="1.03" data-eval-step="7150" data-eval-field="wikitext2_byte_ppl"><span class="l-en" lang="en">—</span><span class="l-pl" lang="pl">—</span></dd></div></dl><p class="method-note"><span class="l-en" lang="en">Metrics above use the shared evaluation protocol at step 7150. Older lm-eval results are reported separately.</span><span class="l-pl" lang="pl">Powyższe metryki pochodzą ze wspólnego protokołu ewaluacji dla kroku 7150. Starsze wyniki lm-eval są w osobnym raporcie.</span></p>
|
| 470 |
<div class="model-links"><a href="https://huggingface.co/SlayerLab/Slayer1B-1.03/tree/main"><span class="l-en" lang="en">Weight files</span><span class="l-pl" lang="pl">Pliki wag</span></a><a href="https://huggingface.co/SlayerLab/Slayer1B-1.03-checkpoints"><span class="l-en" lang="en">Checkpoints</span><span class="l-pl" lang="pl">Checkpointy</span></a><a href="https://huggingface.co/datasets/SlayerLab/slayer-pl-8x3b"><span class="l-en" lang="en">Data</span><span class="l-pl" lang="pl">Dane</span></a><a href="https://huggingface.co/SlayerLab/Slayer1B-1.03"><span class="l-en" lang="en">Model card</span><span class="l-pl" lang="pl">Karta</span></a><a href="https://track.fabryka.ai/run/7ca4f84f-3b07-4525-8cde-f3a936dc3425"><span class="l-en" lang="en">Run</span><span class="l-pl" lang="pl">Bieg</span></a></div></article><article class="model-card" data-model-card data-version="1.02" data-status="released" data-search="slayer1b 1.02 released wydany base polish polski">
|
| 471 |
<div class="model-card__head"><div><p class="section-kicker">Slayer1B</p><h3 class="model-name">1.02</h3></div><span class="status-badge status-released"><span class="l-en" lang="en">Released</span><span class="l-pl" lang="pl">Wydany</span></span></div>
|
| 472 |
<p class="model-description"><span class="l-en" lang="en">Continuation at a constant learning rate, without a final decay.</span><span class="l-pl" lang="pl">Kontynuacja przy stałym tempie uczenia, bez końcowego wygaszania.</span></p>
|