Search before asking

• I have searched the YOLOv5 issues and found no similar bug report.

YOLOv5 Component

Detection

Bug

Detection on edgetpu model returns the following error:
File "yolov5/models/common.py", line 449, in forward y[..., :4] *= [w, h, w, h] # xywh normalized to pixels UnboundLocalError: local variable 'y' referenced before assignment

I believe that the error from is caused by self.tflite not being set in the if/else branches in the forward method when edgetpu is true (common.py line 438). That should be cause by tflite &= not edgetpu # *.tflite in the model_type method.

@staticmethod
    def model_type(p='path/to/model.pt'):
        # Return model type from model path, i.e. path='path/to/model.onnx' -> type=onnx
        from export import export_formats
        suffixes = list(export_formats().Suffix) + ['.xml']  # export suffixes
        check_suffix(p, suffixes)  # checks
        p = Path(p).name  # eliminate trailing separators
        pt, jit, onnx, xml, engine, coreml, saved_model, pb, tflite, edgetpu, tfjs, xml2 = (s in p for s in suffixes)
        xml |= xml2  # *_openvino_model or *.xml
        tflite &= not edgetpu  # *.tflite
        return pt, jit, onnx, xml, engine, coreml, saved_model, pb, tflite, edgetpu, tfjs

tflite &= not edgetpu sets tflite to true only if tflite is true and edgetpu is false, it is set to false otherwise.

Environment

• OS: Ubuntu 20.04
• Python 3.8.10

Minimal Reproducible Example

I have exported the YoloV5s model to edgetpu using:
python export.py --weights yolov5s.pt --include edgetpu --img 320 --data data/coco128.yaml

Then run the detection with python detect.py --weights yolov5s-int8_edgetpu.tflite --img 320 --data data/coco128.yaml

Additional

No response

Are you willing to submit a PR?

• Yes I'd like to help by submitting a PR!