Update README.md
Browse files
README.md
CHANGED
|
@@ -384,6 +384,17 @@ model-index:
|
|
| 384 |
|
| 385 |
This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [Qwen/Qwen3-VL-Embedding-2B](https://huggingface.co/Qwen/Qwen3-VL-Embedding-2B). It maps sentences & paragraphs to a 2048-dimensional dense vector space and can be used for retrieval.
|
| 386 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 387 |
## Model Details
|
| 388 |
|
| 389 |
### Model Description
|
|
|
|
| 384 |
|
| 385 |
This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [Qwen/Qwen3-VL-Embedding-2B](https://huggingface.co/Qwen/Qwen3-VL-Embedding-2B). It maps sentences & paragraphs to a 2048-dimensional dense vector space and can be used for retrieval.
|
| 386 |
|
| 387 |
+
> [!NOTE]
|
| 388 |
+
> The evaluation results reported in this README were obtained with `max_pixels = 768 * 32 * 32`. When evaluated on **KoViDoRe v2** with `max_pixels = 1800 * 32 * 32`, the NDCG@10 scores are as follows:
|
| 389 |
+
>
|
| 390 |
+
> | Domain | NDCG@10 |
|
| 391 |
+
> |---|---|
|
| 392 |
+
> | Cybersecurity | 0.7073 |
|
| 393 |
+
> | Energy | 0.6035 |
|
| 394 |
+
> | Hr | 0.4107 |
|
| 395 |
+
> | Economic | 0.2404 |
|
| 396 |
+
> | **Average** | **0.4905** |
|
| 397 |
+
|
| 398 |
## Model Details
|
| 399 |
|
| 400 |
### Model Description
|