Error when I use torch_model_2_trt.py to convert yolov3-spp.weights

cmcamdy / YOLOV3_Tensorrt_Rectangular_Inference

After use TensorRT model & rectangular Inference , the time forward is much lower then before.

MIT License

21 stars 4 forks source link

Error when I use torch_model_2_trt.py to convert yolov3-spp.weights #2

Open swim010 opened 4 years ago

swim010 commented 4 years ago

The error is:

RuntimeError: shape '[42, 256, 1, 1]' is invalid for input of size 10751.

Can you help me?

cmcamdy commented 4 years ago

hi，i find that 42x256x1x1=10752>10751 that means u may consider whether the input shape is right,or whether this project based on eriklindernoren is suitable for your model. if your model is based on ultralytics ,that may not work out.