Update README.md
Browse files
README.md
CHANGED
|
@@ -15,7 +15,7 @@ Current Leaderboard (You are welcome to test your models on SpatialScore!):
|
|
| 15 |
</div>
|
| 16 |
|
| 17 |
## Some Information
|
| 18 |
-
[Project Page](https://haoningwu3639.github.io/SpatialScore/)
|
| 19 |
|
| 20 |
## News
|
| 21 |
- [2026.5] We have updated our up-to-date code and data!
|
|
@@ -61,27 +61,6 @@ huggingface-cli download --resume-download --repo-type dataset haoningwu/Spatial
|
|
| 61 |
unzip SpatialScore_benchmark.zip
|
| 62 |
```
|
| 63 |
|
| 64 |
-
## SFT with SpatialCorpus
|
| 65 |
-
Please check out [SpaitalCorpus](https://huggingface.co/datasets/haoningwu/SpatialCorpus) to download our introduced training resources (`SpatialCorpus`).
|
| 66 |
-
|
| 67 |
-
You can follow the commands below to prepare the data:
|
| 68 |
-
|
| 69 |
-
```
|
| 70 |
-
huggingface-cli download --resume-download --repo-type dataset haoningwu/SpatialCorpus --local-dir ./ --local-dir-use-symlinks False
|
| 71 |
-
unzip SpatialCorpus.zip
|
| 72 |
-
```
|
| 73 |
-
|
| 74 |
-
Our codebase integrates the training script for Qwen3-VL.
|
| 75 |
-
Once you have prepared the dataset and configured the corresponding data paths, you can use the following script to conduct SFT on the Qwen3-VL model using SpatialCorpus:
|
| 76 |
-
|
| 77 |
-
```
|
| 78 |
-
cd Qwen3-VL/qwen-vl-finetune
|
| 79 |
-
bash ./scripts/sft_qwen3_8b.sh
|
| 80 |
-
```
|
| 81 |
-
|
| 82 |
-
For further details, please refer to the official [Qwen3-VL](https://github.com/qwenlm/qwen3-vl) tutorial.
|
| 83 |
-
|
| 84 |
-
|
| 85 |
## Evaluation
|
| 86 |
Considering the current mainstream model architectures, we have prioritized support for the Qwen2.5-VL and Qwen3-VL series models.
|
| 87 |
You can evaluate them on SpatialScore using the following commands:
|
|
|
|
| 15 |
</div>
|
| 16 |
|
| 17 |
## Some Information
|
| 18 |
+
[Project Page](https://haoningwu3639.github.io/SpatialScore/) 路 [Paper](https://arxiv.org/abs/2505.17012/) 路 [SpatialScore_Benchmark](https://huggingface.co/datasets/haoningwu/SpatialScore) 路 [SpatialCorpus](https://huggingface.co/datasets/haoningwu/SpatialCorpus) 路 [Model](https://huggingface.co/haoningwu/SpatialScore)
|
| 19 |
|
| 20 |
## News
|
| 21 |
- [2026.5] We have updated our up-to-date code and data!
|
|
|
|
| 61 |
unzip SpatialScore_benchmark.zip
|
| 62 |
```
|
| 63 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 64 |
## Evaluation
|
| 65 |
Considering the current mainstream model architectures, we have prioritized support for the Qwen2.5-VL and Qwen3-VL series models.
|
| 66 |
You can evaluate them on SpatialScore using the following commands:
|