KeyError: 'cv3'

[zzff-sys]

My training weights were trained by adding the attention mechanism to YOLOv5's network, and there was an error when I used the code to export the thermal map： What is the problem and how to solve it，Thank you very much

[pooya-mohammadi]

@zzff-sys Checked the model and works fine. Please check it again on colab

[jccmoing-bit]

@zzff-sys Checked the model and works fine. Please check it again on colab

I added an attention mechanism to the model and added related modules to common.py, but I also got an error in the end: KeyError: 'cv3'

[jccmoing-bit]

My training weights were trained by adding the attention mechanism to YOLOv5's network, and there was an error when I used the code to export the thermal map： What is the problem and how to solve it，Thank you very much

how did you solve this problem？

[pooya-mohammadi]

@jccmoing-bit

1. Your weights are trained on a custom dataset or you are using pretrained weights of yolov5?
2. Secondly, have installed the requirements?
3. Finally, have tried colab notebook provided for inferencing?

[dayu178]

I have the same problem

[dayu178]

Use yolov5 custom training with added attention mechanism

[SolitudeWX]

I have found the reason. See these lines in main.py: parser.add_argument('--target-layer', type=str, default='model_23_cv3_act', help='The layer hierarchical address to which gradcam will applied,' ' the names should be separated by underline') means what layer you want to find. The names after default should be the layer name in your model, but to see the name of layer, you should debug the code.