Spaces:
Running
Running
Update benchmark report
Browse files- index.html +6 -4
index.html
CHANGED
|
@@ -519,7 +519,7 @@ td.rank-cell{font-size:15px;font-weight:700;width:38px;text-align:center;}
|
|
| 519 |
</div>
|
| 520 |
|
| 521 |
<p class="hero-desc lang-en">
|
| 522 |
-
<strong>cheRAGh (چراغ)</strong> is a unified benchmarking suite for Persian Retrieval-Augmented Generation (RAG) systems, covering embedding models, rerankers, retrieval quality, tool calling, and large language model performance across diverse Persian-language datasets from
|
| 523 |
</p>
|
| 524 |
<p class="hero-desc lang-en">
|
| 525 |
This report presents the evaluation results of embedding models in the <strong>cheRAGh</strong> benchmark. Retrieval performance is measured using <strong>Recall@5</strong> and <strong>MRR</strong>.
|
|
@@ -612,7 +612,7 @@ td.rank-cell{font-size:15px;font-weight:700;width:38px;text-align:center;}
|
|
| 612 |
<div class="table-toolbar">
|
| 613 |
<span class="sort-label" id="sort-label">Sorted by <strong>Average MRR ↓</strong></span>
|
| 614 |
<button class="btn-ghost active" id="btn-info" onclick="toggleInfo()">
|
| 615 |
-
<span class="lang-en">
|
| 616 |
</button>
|
| 617 |
</div>
|
| 618 |
<div class="table-wrap">
|
|
@@ -739,7 +739,7 @@ td.rank-cell{font-size:15px;font-weight:700;width:38px;text-align:center;}
|
|
| 739 |
</div>
|
| 740 |
|
| 741 |
<div class="footer">
|
| 742 |
-
<div>cheRAGh · <a href="https://huggingface.co/QomSSLab" target="_blank">HuggingFace</a> · 2026-06-30
|
| 743 |
<div>
|
| 744 |
<span style="color:var(--high)">● ≥ 0.75</span>  
|
| 745 |
<span style="color:var(--mid)">● 0.50–0.74</span>  
|
|
@@ -882,7 +882,7 @@ function renderChips(){
|
|
| 882 |
|
| 883 |
/* LEADERBOARD */
|
| 884 |
let sortState={by:'avg_mrr',dir:'desc',ds:null};
|
| 885 |
-
let infoVisible=
|
| 886 |
function renderTable(){
|
| 887 |
const filtered=MODELS.filter(passesFilters);
|
| 888 |
const sorted=[...filtered].sort((a,b)=>{
|
|
@@ -990,6 +990,8 @@ function rebuildCharts(){buildBar('chart-mrr-bar',MRR_DATA,'MRR');buildBar('char
|
|
| 990 |
|
| 991 |
/* INIT */
|
| 992 |
renderTable();buildHeatmap();
|
|
|
|
|
|
|
| 993 |
buildBar('chart-mrr-bar',MRR_DATA,'MRR');buildBar('chart-recall-bar',REC_DATA,RECALL_LABEL);
|
| 994 |
buildRadar('chart-radar-mrr',MRR_DATA);buildRadar('chart-radar-recall',REC_DATA);
|
| 995 |
</script>
|
|
|
|
| 519 |
</div>
|
| 520 |
|
| 521 |
<p class="hero-desc lang-en">
|
| 522 |
+
<strong>cheRAGh (چراغ)</strong> is a unified benchmarking suite for Persian Retrieval-Augmented Generation (RAG) systems, covering embedding models, rerankers, retrieval quality, tool calling, and large language model performance across diverse Persian-language datasets from General, Scientific, Education, Legal, and Religious domains.
|
| 523 |
</p>
|
| 524 |
<p class="hero-desc lang-en">
|
| 525 |
This report presents the evaluation results of embedding models in the <strong>cheRAGh</strong> benchmark. Retrieval performance is measured using <strong>Recall@5</strong> and <strong>MRR</strong>.
|
|
|
|
| 612 |
<div class="table-toolbar">
|
| 613 |
<span class="sort-label" id="sort-label">Sorted by <strong>Average MRR ↓</strong></span>
|
| 614 |
<button class="btn-ghost active" id="btn-info" onclick="toggleInfo()">
|
| 615 |
+
<span class="lang-en">Show model info</span><span class="lang-fa" style="display:none">نمایش اطلاعات مدل</span>
|
| 616 |
</button>
|
| 617 |
</div>
|
| 618 |
<div class="table-wrap">
|
|
|
|
| 739 |
</div>
|
| 740 |
|
| 741 |
<div class="footer">
|
| 742 |
+
<div>cheRAGh · <a href="https://huggingface.co/QomSSLab" target="_blank">HuggingFace</a> · 2026-06-30 11:41</div>
|
| 743 |
<div>
|
| 744 |
<span style="color:var(--high)">● ≥ 0.75</span>  
|
| 745 |
<span style="color:var(--mid)">● 0.50–0.74</span>  
|
|
|
|
| 882 |
|
| 883 |
/* LEADERBOARD */
|
| 884 |
let sortState={by:'avg_mrr',dir:'desc',ds:null};
|
| 885 |
+
let infoVisible=false;
|
| 886 |
function renderTable(){
|
| 887 |
const filtered=MODELS.filter(passesFilters);
|
| 888 |
const sorted=[...filtered].sort((a,b)=>{
|
|
|
|
| 990 |
|
| 991 |
/* INIT */
|
| 992 |
renderTable();buildHeatmap();
|
| 993 |
+
document.querySelectorAll('.info-col-th,.model-info-col').forEach(el=>el.style.display='none');
|
| 994 |
+
document.getElementById('btn-info').classList.remove('active');
|
| 995 |
buildBar('chart-mrr-bar',MRR_DATA,'MRR');buildBar('chart-recall-bar',REC_DATA,RECALL_LABEL);
|
| 996 |
buildRadar('chart-radar-mrr',MRR_DATA);buildRadar('chart-radar-recall',REC_DATA);
|
| 997 |
</script>
|