HeungJunKim

Re-parameterization.ipynb >> 'Model' object has no attribute 'nc'

1

I trained 300 epochs and chose the last model. model : YOLOv7-X weight file name : best_299.pt but, An error occurs at this location. in Re-parameterization.ipynb model.nc = ckpt['model'].nc ![스크린샷,...

Reparameter Problems : (model.nc+5) x 3 or (model.nc+5) x anchors

2

custom train model : yolov7-w6 model nc : 4 yolov7-w6 anchors : 4 so.. try for i in range((4+5)*anchors): # 9x4 36 >> IndexError: index 27 is out of bounds...

convert trt model & inference trt error >> failed.Cannot deserialize

2

try python/c++ example. same error occurs. step1. python export.py --weights yolov7x_reparam.pt --grid --end2end --simplify --topk-all 100 --iou-thres 0.45 --conf-thres 0.05 --img-size 640 640 step2. python ./tensorrt-python/export.py -o yolov7x_reparam.onnx -e yolov7x-nms.trt...

custom yolov7 model works in tensorrt, but not in deepstream. > deserializeEngineAndBackend()

3

nvinfer gstnvinfer.cpp:638:gst_nvinfer_logger: NvDsInferContext[UID 1]: Info from NvDsInferContextImpl::deserializeEngineAndBackend() [UID = 1]: deserialized trt engine from :/home/nvidia/workspace/deepstream_sdk/DeepStream-Yolo/yolov7x.engine deepstream-app error >> deserializeEngineAndBackend() It looks similar to the error in tensorrt [07/28/2022-15:02:55] [TRT] [E]...