version 1.5

README.md

@@ -12,7 +12,7 @@
 **OmniParser** is a comprehensive method for parsing user interface screenshots into structured and easy-to-understand elements, which significantly enhances the ability of GPT-4V to generate actions that can be accurately grounded in the corresponding regions of the interface. 
 
 ## News
-- [2024/11] OmniParser V1.5 is out! It features: 1) new icon detection model trained on cleaned data, and 2) improved bbox merging logic. 
+- [2024/11/26] We release an updated version, OmniParser V1.5 which features more fine grained/small icon detection. Examples in the demo.ipynb. 
 - [2024/10] OmniParser was the #1 trending model on huggingface model hub (starting 10/29/2024). 
 - [2024/10] Feel free to checkout our demo on [huggingface space](https://huggingface.co/spaces/microsoft/OmniParser)! (stay tuned for OmniParser + Claude Computer Use)
 - [2024/10] Both Interactive Region Detection Model and Icon functional description model are released! [Hugginface models](https://huggingface.co/microsoft/OmniParser)

demo.ipynb

@@ -20,9 +20,6 @@
     "from PIL import Image\n",
     "device = 'cuda'\n",
     "\n",
-    "# som_model = get_yolo_model(model_path='weights/icon_detect/best.pt')\n",
-    "# som_model = get_yolo_model(model_path='/home/yadonglu/sandbox/data/yolo/runs/detect/yolo11l_som_detection_seq_10ep_b32_filter5more4/weights/best.pt')\n",
-    "# som_model = get_yolo_model('/home/yadonglu/sandbox/data/yolo/runs/detect/yolo11l_som_detection_seq_10ep_b24_filter5more1280/weights/best.pt')\n",
     "som_model = get_yolo_model(model_path='weights/icon_detect_v1_5/best.pt')\n",
     "\n",
     "som_model.to(device)\n",

weights/convert_safetensor_to_pt.py

@@ -16,7 +16,7 @@ if args.version == 'v1':
     torch.save({'model':model}, 'weights/icon_detect/best.pt')
 elif args.version == 'v1_5':
     print("Converting v1_5")
-    tensor_dict = load_file("weights/icon_detect_v1_5/model.safetensors")
+    tensor_dict = torch.load("weights/icon_detect_v1_5/model.safetensors")
     model = DetectionModel('weights/icon_detect_v1_5/model.yaml')
     model.load_state_dict(tensor_dict)
     save_dict = {'model':model}