Text Ranking
sentence-transformers
Safetensors
Korean
English
qwen3
cross-encoder
reranker
korean
english
Instructions to use nlpai-lab/KURE-Reranker-base with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- sentence-transformers
How to use nlpai-lab/KURE-Reranker-base with sentence-transformers:
from sentence_transformers import CrossEncoder model = CrossEncoder("nlpai-lab/KURE-Reranker-base") query = "Which planet is known as the Red Planet?" passages = [ "Venus is often called Earth's twin because of its similar size and proximity.", "Mars, known for its reddish appearance, is often referred to as the Red Planet.", "Jupiter, the largest planet in our solar system, has a prominent red spot.", "Saturn, famous for its rings, is sometimes mistaken for the Red Planet." ] scores = model.predict([(query, passage) for passage in passages]) print(scores) - Notebooks
- Google Colab
- Kaggle
Update README.md
Browse files
README.md
CHANGED
|
@@ -28,7 +28,6 @@ tags:
|
|
| 28 |
- **Strong Korean reranking:** 88.49 average nDCG@10 across nine Korean benchmarks, the highest average among the evaluated models with fewer than 4B parameters and within 1.1 points of Qwen3-Reranker-4B.
|
| 29 |
- **Long-document reranking:** 81.63 nDCG@10 on MultiLongDocRetrieval (MLDR), ranking second only to Qwen3-Reranker-8B (82.20) among the evaluated models.
|
| 30 |
- **Efficient:** 55.1 pairs per second averaged over nine benchmarks on an NVIDIA RTX A6000 48GB, about 2.3× Qwen3-Reranker-4B and 3.6× Qwen3-Reranker-8B under the measurement conditions described below.
|
| 31 |
-
- **Long-input support:** Up to **8,192 tokens** for the combined query–document input.
|
| 32 |
|
| 33 |
## Model Overview
|
| 34 |
|
|
@@ -36,7 +35,7 @@ tags:
|
|
| 36 |
|---|---|
|
| 37 |
| Model type | Cross-encoder reranker (causal LM with yes/no logit scoring) |
|
| 38 |
| Backbone | Qwen3-1.7B |
|
| 39 |
-
| Parameters |
|
| 40 |
| Languages | Korean and English |
|
| 41 |
| Maximum inference input length | 8,192 tokens, including instruction, query, document, and prompt template |
|
| 42 |
| Output | One scalar relevance score per query–document pair: logit("yes") − logit("no") |
|
|
@@ -140,7 +139,7 @@ Evaluation was conducted using the [reranker-simple-benchmark](https://github.co
|
|
| 140 |
|
| 141 |
<img src="./assets/pps_vs_ndcg9.png" width="700" alt="Reranking performance (mean nDCG@10) vs. throughput (mean PPS)">
|
| 142 |
|
| 143 |
-
The figure plots mean nDCG@10 against mean PPS across the nine Korean benchmarks, including MLDR. Dot color marks model size. Throughput uses each model's own input limit and its best measured batch size, so it measures neither speed at equal token lengths nor end-to-end retrieval latency.
|
| 144 |
|
| 145 |
### Results — MTEB-ko-retrieval (9 subsets)
|
| 146 |
<!-- **공식 9개 subset 을 모두 평가한 모델**의 9-subset mean NDCG@10, PPS -->
|
|
@@ -148,8 +147,8 @@ The figure plots mean nDCG@10 against mean PPS across the nine Korean benchmarks
|
|
| 148 |
|---|---|---|---|
|
| 149 |
| tomaarsen/Qwen3-Reranker-8B-seq-cls | 7.6B | 0.9004 | 15.1 |
|
| 150 |
| tomaarsen/Qwen3-Reranker-4B-seq-cls | 4.0B | 0.8956 | 24.3 |
|
| 151 |
-
| nlpai-lab/KURE-Reranker-base | 1.7B | 0.8849 | 55.1 |
|
| 152 |
-
| nlpai-lab/KURE-Reranker-nano | 149M | 0.8808 | 473.7 |
|
| 153 |
| zeroentropy/zerank-2-reranker | 4.0B | 0.8695 | 29.4 |
|
| 154 |
| lightonai/LightOn-rerank-PW-4B | 4.5B | 0.8664 | 15.0 |
|
| 155 |
| mixedbread-ai/mxbai-rerank-large-v2 | 1.5B | 0.8661 | 65.2 |
|
|
@@ -164,7 +163,7 @@ The figure plots mean nDCG@10 against mean PPS across the nine Korean benchmarks
|
|
| 164 |
| telepix/PIXIE-Spell-Reranker-Preview-0.6B | 596M | 0.7806 | 99.6 |
|
| 165 |
| cross-encoder/ettin-reranker-1b-v1 | 1.0B | 0.6901 | 49.7 |
|
| 166 |
|
| 167 |
-
PPS is measured on a single NVIDIA RTX A6000 48GB per model and averaged over the nine benchmarks.
|
| 168 |
|
| 169 |
`jinaai/jina-reranker-v3` and `jinaai/jina-reranker-v3.5` are listwise rerankers that cannot be compared under the same 8,192-token condition on MLDR, so they are excluded from the 9-subset table and their MLDR cells are left blank below.
|
| 170 |
|
|
@@ -174,8 +173,8 @@ PPS is measured on a single NVIDIA RTX A6000 48GB per model and averaged over th
|
|
| 174 |
|---|---|---|---|---|---|---|---|---|---|---|
|
| 175 |
| tomaarsen/Qwen3-Reranker-8B-seq-cls | 7.6B | 0.8679 | 0.9546 | 0.8893 | 0.9907 | 0.8490 | 0.8409 | 0.8220 | 0.9880 | 0.9014 |
|
| 176 |
| tomaarsen/Qwen3-Reranker-4B-seq-cls | 4.0B | 0.8733 | 0.9707 | 0.8685 | 0.9906 | 0.8533 | 0.8321 | 0.8105 | 0.9861 | 0.8752 |
|
| 177 |
-
| nlpai-lab/KURE-Reranker-base | 1.7B | 0.8548 | 0.9762 | 0.8716 | 0.9880 | 0.8371 | 0.7970 | 0.8163 | 0.9891 | 0.8343 |
|
| 178 |
-
| nlpai-lab/KURE-Reranker-nano | 149M | 0.8582 | 0.9741 | 0.8475 | 0.9830 | 0.8400 | 0.8011 | 0.7966 | 0.9895 | 0.8368 |
|
| 179 |
| jinaai/jina-reranker-v3.5 | 597M | 0.8539 | 0.9838 | 0.8094 | 0.9733 | 0.8565 | 0.8194 | — | 0.9887 | 0.8545 |
|
| 180 |
| jinaai/jina-reranker-v3 | 597M | 0.8553 | 0.9773 | 0.7960 | 0.9695 | 0.8449 | 0.8104 | — | 0.9859 | 0.8546 |
|
| 181 |
| zeroentropy/zerank-2-reranker | 4.0B | 0.8712 | 0.9436 | 0.8646 | 0.9846 | 0.8003 | 0.8027 | 0.7120 | 0.9791 | 0.8669 |
|
|
@@ -198,8 +197,8 @@ PPS is measured on a single NVIDIA RTX A6000 48GB per model and averaged over th
|
|
| 198 |
|---|---|---|---|---|---|---|---|---|---|---|
|
| 199 |
| tomaarsen/Qwen3-Reranker-8B-seq-cls | 7.6B | 16.6 | 7.3 | 17.5 | 19.4 | 24.9 | 26.5 | 0.7 | 10.6 | 12.3 |
|
| 200 |
| tomaarsen/Qwen3-Reranker-4B-seq-cls | 4.0B | 26.1 | 11.8 | 28.4 | 31.5 | 40.0 | 42.4 | 1.1 | 17.2 | 20.0 |
|
| 201 |
-
| nlpai-lab/KURE-Reranker-base | 1.7B | 59.2 | 27.3 | 64.6 | 71.1 | 89.1 | 97.1 | 2.6 | 39.1 | 45.7 |
|
| 202 |
-
| nlpai-lab/KURE-Reranker-nano | 149M | 475.4 | 233.2 | 569.7 | 630.9 | 765.8 | 855.2 | 16.5 | 330.8 | 386.0 |
|
| 203 |
| jinaai/jina-reranker-v3.5 | 597M | 141.0 | 38.6 | 148.6 | 174.6 | 277.3 | 295.7 | — | 68.4 | 86.7 |
|
| 204 |
| jinaai/jina-reranker-v3 | 597M | 113.8 | 25.0 | 121.8 | 169.2 | 244.6 | 261.6 | — | 48.7 | 64.0 |
|
| 205 |
| zeroentropy/zerank-2-reranker | 4.0B | 31.0 | 12.7 | 33.8 | 37.8 | 51.0 | 56.4 | 1.1 | 18.8 | 22.3 |
|
|
@@ -216,7 +215,7 @@ PPS is measured on a single NVIDIA RTX A6000 48GB per model and averaged over th
|
|
| 216 |
| telepix/PIXIE-Spell-Reranker-Preview-0.6B | 596M | 104.6 | 49.4 | 117.8 | 129.8 | 160.5 | 175.6 | 4.3 | 72.1 | 82.6 |
|
| 217 |
| cross-encoder/ettin-reranker-1b-v1 | 1.0B | 52.4 | 19.9 | 51.4 | 65.6 | 92.7 | 97.2 | 3.8 | 31.2 | 33.4 |
|
| 218 |
|
| 219 |
-
Batch size starts at 8 and doubles until an out-of-memory error. Samples are repeated to fill complete batches. After warmup, three full passes are timed with CUDA events. Throughput is the total number of processed pairs divided by the accumulated GPU forward time. The highest-throughput successful batch is reported. Tokenization, data loading, and CPU preprocessing are excluded. The batch-search approach is informed by the [Ettin reranker speed benchmark](https://huggingface.co/blog/ettin-reranker#speed)
|
| 220 |
|
| 221 |
## Citation
|
| 222 |
|
|
|
|
| 28 |
- **Strong Korean reranking:** 88.49 average nDCG@10 across nine Korean benchmarks, the highest average among the evaluated models with fewer than 4B parameters and within 1.1 points of Qwen3-Reranker-4B.
|
| 29 |
- **Long-document reranking:** 81.63 nDCG@10 on MultiLongDocRetrieval (MLDR), ranking second only to Qwen3-Reranker-8B (82.20) among the evaluated models.
|
| 30 |
- **Efficient:** 55.1 pairs per second averaged over nine benchmarks on an NVIDIA RTX A6000 48GB, about 2.3× Qwen3-Reranker-4B and 3.6× Qwen3-Reranker-8B under the measurement conditions described below.
|
|
|
|
| 31 |
|
| 32 |
## Model Overview
|
| 33 |
|
|
|
|
| 35 |
|---|---|
|
| 36 |
| Model type | Cross-encoder reranker (causal LM with yes/no logit scoring) |
|
| 37 |
| Backbone | Qwen3-1.7B |
|
| 38 |
+
| Parameters | 1.7B |
|
| 39 |
| Languages | Korean and English |
|
| 40 |
| Maximum inference input length | 8,192 tokens, including instruction, query, document, and prompt template |
|
| 41 |
| Output | One scalar relevance score per query–document pair: logit("yes") − logit("no") |
|
|
|
|
| 139 |
|
| 140 |
<img src="./assets/pps_vs_ndcg9.png" width="700" alt="Reranking performance (mean nDCG@10) vs. throughput (mean PPS)">
|
| 141 |
|
| 142 |
+
The figure plots mean nDCG@10 against mean PPS across the nine Korean benchmarks, including MLDR. Dot color marks model size. Throughput uses each model's own input limit and its best measured batch size, so it measures neither speed at equal token lengths nor end-to-end retrieval latency.
|
| 143 |
|
| 144 |
### Results — MTEB-ko-retrieval (9 subsets)
|
| 145 |
<!-- **공식 9개 subset 을 모두 평가한 모델**의 9-subset mean NDCG@10, PPS -->
|
|
|
|
| 147 |
|---|---|---|---|
|
| 148 |
| tomaarsen/Qwen3-Reranker-8B-seq-cls | 7.6B | 0.9004 | 15.1 |
|
| 149 |
| tomaarsen/Qwen3-Reranker-4B-seq-cls | 4.0B | 0.8956 | 24.3 |
|
| 150 |
+
| **nlpai-lab/KURE-Reranker-base** | 1.7B | 0.8849 | 55.1 |
|
| 151 |
+
| **nlpai-lab/KURE-Reranker-nano** | 149M | 0.8808 | 473.7 |
|
| 152 |
| zeroentropy/zerank-2-reranker | 4.0B | 0.8695 | 29.4 |
|
| 153 |
| lightonai/LightOn-rerank-PW-4B | 4.5B | 0.8664 | 15.0 |
|
| 154 |
| mixedbread-ai/mxbai-rerank-large-v2 | 1.5B | 0.8661 | 65.2 |
|
|
|
|
| 163 |
| telepix/PIXIE-Spell-Reranker-Preview-0.6B | 596M | 0.7806 | 99.6 |
|
| 164 |
| cross-encoder/ettin-reranker-1b-v1 | 1.0B | 0.6901 | 49.7 |
|
| 165 |
|
| 166 |
+
PPS is measured on a single NVIDIA RTX A6000 48GB per model and averaged over the nine benchmarks.Inputs are length-sorted and dynamically padded within each batch.
|
| 167 |
|
| 168 |
`jinaai/jina-reranker-v3` and `jinaai/jina-reranker-v3.5` are listwise rerankers that cannot be compared under the same 8,192-token condition on MLDR, so they are excluded from the 9-subset table and their MLDR cells are left blank below.
|
| 169 |
|
|
|
|
| 173 |
|---|---|---|---|---|---|---|---|---|---|---|
|
| 174 |
| tomaarsen/Qwen3-Reranker-8B-seq-cls | 7.6B | 0.8679 | 0.9546 | 0.8893 | 0.9907 | 0.8490 | 0.8409 | 0.8220 | 0.9880 | 0.9014 |
|
| 175 |
| tomaarsen/Qwen3-Reranker-4B-seq-cls | 4.0B | 0.8733 | 0.9707 | 0.8685 | 0.9906 | 0.8533 | 0.8321 | 0.8105 | 0.9861 | 0.8752 |
|
| 176 |
+
| **nlpai-lab/KURE-Reranker-base** | 1.7B | 0.8548 | 0.9762 | 0.8716 | 0.9880 | 0.8371 | 0.7970 | 0.8163 | 0.9891 | 0.8343 |
|
| 177 |
+
| **nlpai-lab/KURE-Reranker-nano** | 149M | 0.8582 | 0.9741 | 0.8475 | 0.9830 | 0.8400 | 0.8011 | 0.7966 | 0.9895 | 0.8368 |
|
| 178 |
| jinaai/jina-reranker-v3.5 | 597M | 0.8539 | 0.9838 | 0.8094 | 0.9733 | 0.8565 | 0.8194 | — | 0.9887 | 0.8545 |
|
| 179 |
| jinaai/jina-reranker-v3 | 597M | 0.8553 | 0.9773 | 0.7960 | 0.9695 | 0.8449 | 0.8104 | — | 0.9859 | 0.8546 |
|
| 180 |
| zeroentropy/zerank-2-reranker | 4.0B | 0.8712 | 0.9436 | 0.8646 | 0.9846 | 0.8003 | 0.8027 | 0.7120 | 0.9791 | 0.8669 |
|
|
|
|
| 197 |
|---|---|---|---|---|---|---|---|---|---|---|
|
| 198 |
| tomaarsen/Qwen3-Reranker-8B-seq-cls | 7.6B | 16.6 | 7.3 | 17.5 | 19.4 | 24.9 | 26.5 | 0.7 | 10.6 | 12.3 |
|
| 199 |
| tomaarsen/Qwen3-Reranker-4B-seq-cls | 4.0B | 26.1 | 11.8 | 28.4 | 31.5 | 40.0 | 42.4 | 1.1 | 17.2 | 20.0 |
|
| 200 |
+
| **nlpai-lab/KURE-Reranker-base** | 1.7B | 59.2 | 27.3 | 64.6 | 71.1 | 89.1 | 97.1 | 2.6 | 39.1 | 45.7 |
|
| 201 |
+
| **nlpai-lab/KURE-Reranker-nano** | 149M | 475.4 | 233.2 | 569.7 | 630.9 | 765.8 | 855.2 | 16.5 | 330.8 | 386.0 |
|
| 202 |
| jinaai/jina-reranker-v3.5 | 597M | 141.0 | 38.6 | 148.6 | 174.6 | 277.3 | 295.7 | — | 68.4 | 86.7 |
|
| 203 |
| jinaai/jina-reranker-v3 | 597M | 113.8 | 25.0 | 121.8 | 169.2 | 244.6 | 261.6 | — | 48.7 | 64.0 |
|
| 204 |
| zeroentropy/zerank-2-reranker | 4.0B | 31.0 | 12.7 | 33.8 | 37.8 | 51.0 | 56.4 | 1.1 | 18.8 | 22.3 |
|
|
|
|
| 215 |
| telepix/PIXIE-Spell-Reranker-Preview-0.6B | 596M | 104.6 | 49.4 | 117.8 | 129.8 | 160.5 | 175.6 | 4.3 | 72.1 | 82.6 |
|
| 216 |
| cross-encoder/ettin-reranker-1b-v1 | 1.0B | 52.4 | 19.9 | 51.4 | 65.6 | 92.7 | 97.2 | 3.8 | 31.2 | 33.4 |
|
| 217 |
|
| 218 |
+
Batch size starts at 8 and doubles until an out-of-memory error. Samples are repeated to fill complete batches. After warmup, three full passes are timed with CUDA events. Throughput is the total number of processed pairs divided by the accumulated GPU forward time. The highest-throughput successful batch is reported. Tokenization, data loading, and CPU preprocessing are excluded. The batch-search approach is informed by the [Ettin reranker speed benchmark](https://huggingface.co/blog/ettin-reranker#speed).
|
| 219 |
|
| 220 |
## Citation
|
| 221 |
|