LZXzju commited on
Commit
1fecc98
·
verified ·
1 Parent(s): 4a78521

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +29 -0
README.md CHANGED
@@ -32,6 +32,35 @@ Project page: https://github.com/lll6gg/UI-R1
32
  | GUI-R1-3B | w/ thinking | 114 | 26.6 |
33
  | UI-R1-3B (v2) | w/ thinking | 129 | 29.8 |
34
  | **UI-R1-E-3B** | w/o thinking | **28** | **33.5** |
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
35
  ## Evaluation Code for GUI Grounding
36
 
37
  1. Generation for UI-R1-E-3B:
 
32
  | GUI-R1-3B | w/ thinking | 114 | 26.6 |
33
  | UI-R1-3B (v2) | w/ thinking | 129 | 29.8 |
34
  | **UI-R1-E-3B** | w/o thinking | **28** | **33.5** |
35
+ ## Leaderboard: UI-I2E-Bench
36
+ | Model | ScreenSpot | UI-I2E-Bench Avg | ScreenSpot-Pro | Avg |
37
+ | :------------: | :--------: | :--------------: | :------------: | :--: |
38
+ | UI-TARS-1.5-7B | 88.1 | 73.2 | 42.2 | 67.8 |
39
+ | Uground-V1-72B | 89.7 | 76.3 | 34.3 | 66.8 |
40
+ | UI-TARS-72B | 88.4 | 73.7 | 38.1 | 66.7 |
41
+ | **UI-R1-E-3B** | 89.2 | 69.1 | 33.5 | 63.9 |
42
+ | Uground-V1-7B | 87.1 | 70.3 | 31.1 | 62.8 |
43
+ | InfiGUI-R1 | 87.5 | 69.7 | 29.6 | 62.3 |
44
+ | UI-TARS-7B | 89.5 | 61.4 | 35.7 | 62.2 |
45
+ | Qwen2.5-VL-72B | 87.1 | 51.4 | 43.6 | 60.7 |
46
+ | UI-I2E-VLM-7B | 82.5 | 69.5 | 23.6 | 58.5 |
47
+ | UI-TARS-2B | 82.3 | 62 | 27.7 | 57.3 |
48
+ | Qwen2.5-VL-7B | 84.7 | 53.8 | 29 | 55.8 |
49
+ | OmniParser-V2 | 72 | 54.8 | 39.6 | 55.5 |
50
+ | Uground-V1-2B | 78.8 | 57.4 | 26.6 | 54.3 |
51
+ | OS-Atlas-7B | 82.5 | 58.6 | 18.9 | 53.3 |
52
+ | UI-R1 | 78.6 | 58.5 | 17.8 | 51.6 |
53
+ | UGround-7B | 74.1 | 54.2 | 16.5 | 48.3 |
54
+ | UI-I2E-VLM-4B | 70.4 | 53.4 | 12.2 | 45.3 |
55
+ | OmniParser | 73.9 | 53.1 | 8.3 | 45.1 |
56
+ | ShowUI-2B | 76.8 | 41.5 | 7.7 | 42 |
57
+ | Qwen2.5-VL-3B | 55.5 | 41.7 | 23.9 | 41.3 |
58
+ | Aguvis-7B | 84.4 | 53.2 | 22.9 | 40.4 |
59
+ | OS-Atlas-4B | 70.1 | 44.3 | 3.7 | 39.4 |
60
+ | Qwen2-VL-7B | 42.6 | 48.7 | 1.6 | 31 |
61
+ | Seeclick | 55.8 | 26.4 | 1.1 | 27.8 |
62
+ | InternVL2-4B | 4.2 | 0.9 | 0.3 | 1.8 |
63
+
64
  ## Evaluation Code for GUI Grounding
65
 
66
  1. Generation for UI-R1-E-3B: