linghu8812 / tensorrt_inference

702 stars 206 forks source link

Yolov4-trt demo error : Weights are outside of fp16 range #58

Closed IGnoredBird closed 3 years ago

IGnoredBird commented 3 years ago

[03/03/2021-17:23:42] [W] [TRT] onnx2trt_utils.cpp:220: Your ONNX model has been generated with INT64 weights, while TensorRT does not natively support INT64. Attempting to cast down to INT32. start building engine [03/03/2021-17:24:04] [I] [TRT] Some tactics do not have sufficient workspace memory to run. Increasing workspace size may increase performance, please check verbose output. [03/03/2021-17:24:05] [E] [TRT] ../builder/cudnnBuilderWeightConverters.cpp (541) - Misc Error in operator(): 1 (Weights are outside of fp16 range. A possible fix is to retrain the model with regularization to bring the magnitude of the weights down.) [03/03/2021-17:24:05] [E] [TRT] ../builder/cudnnBuilderWeightConverters.cpp (541) - Misc Error in operator(): 1 (Weights are outside of fp16 range. A possible fix is to retrain the model with regularization to bring the magnitude of the weights down.) build engine done Yolov4_trt: /harddisk/standard/tensorrt_inference/Yolov4/../includes/common/common.hpp:138: void onnxToTRTModel(const string&, const string&, nvinfer1::ICudaEngine*&, const int&): Assertion `engine' failed. Aborted (core dumped)