Hugging Face's logo

HelloKKMe
/
GTA1-72B

Image-to-Text
Transformers
Safetensors
qwen2_5_vl
text-generation-inference
Model card Files Community
1
HelloKKMe commited on
Commit
674ce16
·
verified ·
1 Parent(s): 6b6cc53

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +5 -5
README.md CHANGED
@@ -24,9 +24,9 @@ We follow the standard evaluation protocol and benchmark our model on three chal
24
  | UGround-v1-72B | 72B | ✅ | — | 34.5 | — |
25
  | Qwen2.5-VL-72B-Instruct | 72B | ✅ | 94.00* | 53.3 | 62.2* |
26
  | UI-TARS | 72B | ✅ | 90.3 | 38.1 | — |
27
- | Grounding-R1 (Ours) | 7B | ✅ | 92.4 <sub>*(∆ +2.7)*</sub> | 50.1<sub>*(∆ +8.1)*</sub> | 67.7 <sub>*(∆ +3.5)*</sub> |
28
- | Grounding-R1 (Ours) | 32B | ✅ | 93.2 <sub>*(∆ +1.3)*</sub> | 53.6 <sub>*(∆ +5.6)*</sub> | 61.9<sub>*(∆ +2.3)*</sub> |
29
- | Grounding-R1 (Ours) | 72B | ✅ | 94.8<sub>*(∆ +0.8)*</sub> | 58.4 <sub>*(∆ +5.1)*</sub> | 66.7<sub>*(∆ +4.5)*</sub> |
30
 
31
 
32
  > **Note:**
@@ -63,7 +63,7 @@ def extract_coordinates(raw_string):
63
  return 0,0
64
 
65
  # Load model and processor
66
- model_path = "HelloKKMe/grounding-r1-72B"
67
  max_new_tokens = 32
68
 
69
  model = Qwen2_5_VLForConditionalGeneration.from_pretrained(
@@ -125,4 +125,4 @@ pred_y*=scale_y
125
  print(pred_x,pred_y)
126
  ```
127
 
128
- Refer to our [code](https://github.com/Yan98/Grounding-R1) for more details.
 
24
  | UGround-v1-72B | 72B | ✅ | — | 34.5 | — |
25
  | Qwen2.5-VL-72B-Instruct | 72B | ✅ | 94.00* | 53.3 | 62.2* |
26
  | UI-TARS | 72B | ✅ | 90.3 | 38.1 | — |
27
+ | GTA1 (Ours) | 7B | ✅ | 92.4 <sub>*(∆ +2.7)*</sub> | 50.1<sub>*(∆ +8.1)*</sub> | 67.7 <sub>*(∆ +3.5)*</sub> |
28
+ | GTA1 (Ours) | 32B | ✅ | 93.2 <sub>*(∆ +1.3)*</sub> | 53.6 <sub>*(∆ +5.6)*</sub> | 61.9<sub>*(∆ +2.3)*</sub> |
29
+ | GTA1 (Ours) | 72B | ✅ | 94.8<sub>*(∆ +0.8)*</sub> | 58.4 <sub>*(∆ +5.1)*</sub> | 66.7<sub>*(∆ +4.5)*</sub> |
30
 
31
 
32
  > **Note:**
 
63
  return 0,0
64
 
65
  # Load model and processor
66
+ model_path = "HelloKKMe/GTA1-72B"
67
  max_new_tokens = 32
68
 
69
  model = Qwen2_5_VLForConditionalGeneration.from_pretrained(
 
125
  print(pred_x,pred_y)
126
  ```
127
 
128
+ Refer to our [code](https://github.com/Yan98/GTA1) for more details.