diff --git a/README.md b/README.md
index f005ee9..c26fd1d 100644
--- a/README.md
+++ b/README.md
@@ -29,9 +29,10 @@ All models use the backbone pretrained with [SoCo](https://github.com/hologerry/
 # Main Results
 | Model                  | Lr Schd   | AP<sup>bb</sup><sub>r</sub> | AP<sup>bb</sup><sub>c</sub> | AP<sup>bb</sup><sub>f</sub> | AP<sup>bb</sup> | AP<sup>mk</sup><sub>r</sub> | AP<sup>mk</sup><sub>c</sub>| AP<sup>mk</sup><sub>f</sub> | AP<sup>mk</sup> | Config | Prompt | Model |
 | ---------------------- | --------- | ---- | ---- | ---- | ---- | ------- | ---- | ---- | ---- | ------ | ------ | ----- |
-| ViLD                   | 20 epochs | 17.4 | 27.5 | 31.9 | 27.5 | 16.8 | 25.6 | 28.5 | 25.2 | [config](https://github.com/dyabel/detpro/blob/main/configs/lvis/detpro_ens_20e.py) |  [prompt](https://cloud.tsinghua.edu.cn/f/3f9017c3e217496ebc25/?dl=1) | [model](https://cloud.tsinghua.edu.cn/f/d57e11e2ebf24d509218/?dl=1)      |
+| ViLD*                   | 20 epochs | 17.4 | 27.5 | 31.9 | 27.5 | 16.8 | 25.6 | 28.5 | 25.2 | [config](https://github.com/dyabel/detpro/blob/main/configs/lvis/detpro_ens_20e.py) |  [prompt](https://cloud.tsinghua.edu.cn/f/3f9017c3e217496ebc25/?dl=1) | [model](https://cloud.tsinghua.edu.cn/f/d57e11e2ebf24d509218/?dl=1)      |
 | DetPro (Mask R-CNN)    | 20 epochs | 20.8 | 27.8 | 32.4 | 28.4 | 19.8 | 25.6 | 28.9 | 25.9 | [config](https://github.com/dyabel/detpro/blob/main/configs/lvis/detpro_ens_20e.py) | [prompt](https://cloud.tsinghua.edu.cn/f/0fceb9cae4c249188170/?dl=1) | [model](https://cloud.tsinghua.edu.cn/f/91cecd9ef97843339c79/?dl=1) |
-| DetPro (Cascade R-CNN) | 20 epochs | 21.6 | 29.8 | 35.1 | 30.5 | 19.8 | 26.8 | 30.3 | 26.9 | [config](https://github.com/dyabel/detpro/blob/main/configs/lvis/cascade_mask_rcnn_r50_fpn_sample1e-3_mstrain_20e_lvis_v1_pretrain_ens.py) | [prompt](https://cloud.tsinghua.edu.cn/f/0fceb9cae4c249188170/?dl=1) | [model](https://cloud.tsinghua.edu.cn/f/f75712011cd342bdb49e/?dl=1) |
+| DetPro (Cascade R-CNN) | 20 epochs | 21.6 | 29.8 | 35.1 | 30.5 | 19.8 | 26.8 | 30.3 | 26.9 | [config](https://github.com/dyabel/detpro/blob/main/configs/lvis/cascade_mask_rcnn_r50_fpn_sample1e-3_mstrain_20e_lvis_v1_pretrain_ens.py) | [prompt](https://cloud.tsinghua.edu.cn/f/0fceb9cae4c249188170/?dl=1) | [model](https://cloud.tsinghua.edu.cn/f/f75712011cd342bdb49e/?dl=1) | 
+We re-implement ViLD (denoted as ViLD*) with backbone pretrained using [SoCo](https://github.com/hologerry/SoCo). Our re-implementation version achieves comparable AP compared with the original implementation, while reducing the training epochs from 460 to 20.
 # Installation
 This repo is built on [mmdetection](https://github.com/open-mmlab/mmdetection), [CLIP](https://github.com/openai/CLIP.git) and [CoOP](https://github.com/kaiyangzhou/coop)