model names in PW_FT_detection changed for consistency with PytorchWi…

…ldlife

PW_FT_detection/README.md

@@ -69,19 +69,19 @@ Below you find the models that you can use for fine-tuning, along with their res
 
 |Model|Name|License|
 |---|---|---|
-|MegaDetectroV6-Ultralytics-YoloV9-Compact|MDV6-yolov9c|AGPL-3.0|
-|MegaDetectorV6-Ultralytics-YoloV9-Extra|MDV6-yolov9e|AGPL-3.0|
-|MegaDetectorV6-Ultralytics-YoloV10-Compact|MDV6-yolov10n|AGPL-3.0|
-|MegaDetectorV6-Ultralytics-YoloV10-Extra|MDV6-yolov10x|AGPL-3.0|
-|MegaDetectorV6-Ultralytics-RtDetr-l|MDV6-rtdetrl|AGPL-3.0|
+|MegaDetectroV6-Ultralytics-YoloV9-Compact|MDV6-yolov9-c|AGPL-3.0|
+|MegaDetectorV6-Ultralytics-YoloV9-Extra|MDV6-yolov9-e|AGPL-3.0|
+|MegaDetectorV6-Ultralytics-YoloV10-Compact|MDV6-yolov10-c|AGPL-3.0|
+|MegaDetectorV6-Ultralytics-YoloV10-Extra|MDV6-yolov10-e|AGPL-3.0|
+|MegaDetectorV6-Ultralytics-RtDetr-Compact|MDV6-rtdetr-c|AGPL-3.0|
 
 ## Configuration
 
 Before training your model, you need to configure the training and data parameters in the `config.yaml` file. Here's a brief explanation of the parameters to help both technical and non-technical users understand their purposes:
 
 - **General Parameters:**  
   - `model`: The type of model used (YOLO or RTDETR). Default: YOLO  
-  - `model_name`: The name of the model [see availbale names here](#detection-models-available-for-finetuning). Default: MDV6-yolov9e.pt  
+  - `model_name`: The name of the model [see availbale names here](#detection-models-available-for-finetuning). Default: MDV6-yolov9-e.pt  
   - `data`: Path to the dataset configuration file. Default: ./data/data_example.yaml  
   - `test_data`: Path to the test data directory. Default: ./data/data_example/images/test  
   - `task`: The task to perform (train, validation or inference). Default: train  

PW_FT_detection/config.yaml

@@ -1,10 +1,10 @@
 # General
-model: RTDETR #YOLO or RTDETR
-model_name: MDV6-rtdetrl
+model: YOLO #YOLO or RTDETR
+model_name: MDV6-yolov9-e
 data: ./data/data_example.yaml 
 test_data: ./data/data_example/images/test 
 task: train
-exp_name: MDV6-yolov9e
+exp_name: MDV6-yolov9-e
 
 # Train
 epochs: 20

PW_FT_detection/utils.py

@@ -4,23 +4,23 @@
 
 def get_model_path(model):
 
-    if model == "MDV6-yolov9c":           
+    if model == "MDV6-yolov9-c":           
         url = "https://zenodo.org/records/14567879/files/MDV6b-yolov9c.pt?download=1" 
         model_name = "MDV6b-yolov9c.pt"
-    elif model == "MDV6-yolov9e":
+    elif model == "MDV6-yolov9-e":
         url = "https://zenodo.org/records/14567879/files/MDV6-yolov9e.pt?download=1"
         model_name = "MDV6-yolov9e.pt"
-    elif model == "MDV6-yolov10n":
+    elif model == "MDV6-yolov10-c":
         url = "https://zenodo.org/records/14567879/files/MDV6-yolov10n.pt?download=1"
         model_name = "MDV6-yolov10n.pt"
-    elif model == "MDV6-yolov10x":
+    elif model == "MDV6-yolov10-e":
         url = "https://zenodo.org/records/14567879/files/MDV6-yolov10x.pt?download=1"
         model_name = "MDV6-yolov10x.pt"
-    elif model == "MDV6-rtdetrl":
+    elif model == "MDV6-rtdetr-c":
         url = "https://zenodo.org/records/14567879/files/MDV6b-rtdetrl.pt?download=1"
         model_name = "MDV6b-rtdetrl.pt"
     else:
-        raise ValueError('Select a valid model version: yolov9c, yolov9e, yolov10n, yolov10x or rtdetrl')
+        raise ValueError('Select a valid model version: MDV6-yolov9-c, MDV6-yolov9-e, MDV6-yolov10-c, MDV6-yolov10-e or MDV6-rtdetr-c')
 
     if not os.path.exists(os.path.join(torch.hub.get_dir(), "checkpoints", model_name)):
         os.makedirs(os.path.join(torch.hub.get_dir(), "checkpoints"), exist_ok=True)