Dong fang

Results 6 comments of Dong fang

同一模型结构 使用cpu可以正常推理使用gpu会直接拉满内存程序卡死

> vkBeginCommandBuffer failed -2 有报这个错误码吗？ 错误码为 vkQueueSubmit failed -4 我换一个模型用gpu推理就没问题 这个模型也就15mb的参数。然后我重新训练一遍再转换一遍就没问题了

The same engine produces incorrect inference output using C++, but correct results using Python. TensorRT-8.6.1.6+cuda12.1

> can you share your python code? ``` import tensorrt as trt import pycuda.driver as cuda import pycuda.autoinit import numpy as np import cv2 # 加载TensorRT引擎 TRT_LOGGER = trt.Logger(trt.Logger.INFO) with...

The same engine produces incorrect inference output using C++, but correct results using Python. TensorRT-8.6.1.6+cuda12.1

> how about c++ code ``` #pragma once #include #include "NvInfer.h" #include "cuda_runtime_api.h" #include #include #include using namespace nvinfer1; class Cls { public: ~Cls(); bool Init(std::string engine_file_path); int Detect(cv::Mat img);...

The same engine produces incorrect inference output using C++, but correct results using Python. TensorRT-8.6.1.6+cuda12.1

> how about modify > > ``` > for (int y = 0; y < h; ++y) { > for (int x = 0; x < w; ++x) { >...

The same engine produces incorrect inference output using C++, but correct results using Python. TensorRT-8.6.1.6+cuda12.1

> how about modify > > ``` > for (int y = 0; y < h; ++y) { > for (int x = 0; x < w; ++x) { >...

请问为什么换个自己的模型就不能使用vulkan了？

使用int8 模型是 会报 set_vulkan_compute failed, network use_vulkan_compute disabled fp16 和fp32 就能正常使用 windows端使用的ncnn