Update README.md
Browse files
README.md
CHANGED
|
@@ -106,6 +106,69 @@ model-index:
|
|
| 106 |
- type: cosine_accuracy
|
| 107 |
value: 0.994
|
| 108 |
name: Cosine Accuracy
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 109 |
---
|
| 110 |
|
| 111 |
# SentenceTransformer based on answerdotai/ModernBERT-large
|
|
@@ -216,6 +279,16 @@ You can finetune this model on your own dataset.
|
|
| 216 |
|:--------------------|:----------|
|
| 217 |
| **cosine_accuracy** | **0.994** |
|
| 218 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 219 |
<!--
|
| 220 |
## Bias, Risks and Limitations
|
| 221 |
|
|
|
|
| 106 |
- type: cosine_accuracy
|
| 107 |
value: 0.994
|
| 108 |
name: Cosine Accuracy
|
| 109 |
+
- task:
|
| 110 |
+
type: retrieval
|
| 111 |
+
dataset:
|
| 112 |
+
name: SCIDOCS
|
| 113 |
+
type: SCIDOCS
|
| 114 |
+
split: test
|
| 115 |
+
metrics:
|
| 116 |
+
- type: ndcg@10
|
| 117 |
+
value: 0.15789
|
| 118 |
+
- task:
|
| 119 |
+
type: retrieval
|
| 120 |
+
dataset:
|
| 121 |
+
name: FiQA2018
|
| 122 |
+
type: FiQA2018
|
| 123 |
+
split: test
|
| 124 |
+
metrics:
|
| 125 |
+
- type: ndcg@10
|
| 126 |
+
value: 0.33974
|
| 127 |
+
- task:
|
| 128 |
+
type: retrieval
|
| 129 |
+
dataset:
|
| 130 |
+
name: HotpotQA
|
| 131 |
+
type: HotpotQA
|
| 132 |
+
split: test
|
| 133 |
+
metrics:
|
| 134 |
+
- type: ndcg@10
|
| 135 |
+
value: 0.51818
|
| 136 |
+
- task:
|
| 137 |
+
type: retrieval
|
| 138 |
+
dataset:
|
| 139 |
+
name: ArguAna
|
| 140 |
+
type: ArguAna
|
| 141 |
+
split: test
|
| 142 |
+
metrics:
|
| 143 |
+
- type: ndcg@10
|
| 144 |
+
value: 0.47797
|
| 145 |
+
- task:
|
| 146 |
+
type: retrieval
|
| 147 |
+
dataset:
|
| 148 |
+
name: NFCorpus
|
| 149 |
+
type: NFCorpus
|
| 150 |
+
split: test
|
| 151 |
+
metrics:
|
| 152 |
+
- type: ndcg@10
|
| 153 |
+
value: 0.28443
|
| 154 |
+
- task:
|
| 155 |
+
type: retrieval
|
| 156 |
+
dataset:
|
| 157 |
+
name: SciFact
|
| 158 |
+
type: SciFact
|
| 159 |
+
split: test
|
| 160 |
+
metrics:
|
| 161 |
+
- type: ndcg@10
|
| 162 |
+
value: 0.60626
|
| 163 |
+
- task:
|
| 164 |
+
type: retrieval
|
| 165 |
+
dataset:
|
| 166 |
+
name: TRECCOVID
|
| 167 |
+
type: TRECCOVID
|
| 168 |
+
split: test
|
| 169 |
+
metrics:
|
| 170 |
+
- type: ndcg@10
|
| 171 |
+
value: 0.77495
|
| 172 |
---
|
| 173 |
|
| 174 |
# SentenceTransformer based on answerdotai/ModernBERT-large
|
|
|
|
| 279 |
|:--------------------|:----------|
|
| 280 |
| **cosine_accuracy** | **0.994** |
|
| 281 |
|
| 282 |
+
#### Subset of BEIR
|
| 283 |
+
| dataset | ndcg@10 |
|
| 284 |
+
|:----------|----------:|
|
| 285 |
+
| SCIDOCS | 0.15789 |
|
| 286 |
+
| FiQA2018 | 0.33974 |
|
| 287 |
+
| HotpotQA | 0.51818 |
|
| 288 |
+
| ArguAna | 0.47797 |
|
| 289 |
+
| NFCorpus | 0.28443 |
|
| 290 |
+
| SciFact | 0.60626 |
|
| 291 |
+
| TRECCOVID | 0.77495 |
|
| 292 |
<!--
|
| 293 |
## Bias, Risks and Limitations
|
| 294 |
|