facefusion-onnxrun icon indicating copy to clipboard operation
facefusion-onnxrun copied to clipboard

使用onnxruntime部署facefusion换脸,包含C++和Python两个版本的程序

Results 9 facefusion-onnxrun issues
Sort by recently updated
recently updated
newest added

this->input_image.resize(3 * image_area); size_t single_chn_size = image_area * sizeof(float); memcpy(this->input_image.data(), (float *)bgrChannels[0].data, single_chn_size); //明显的越界 memcpy(this->input_image.data() + image_area, (float *)bgrChannels[1].data, single_chn_size); memcpy(this->input_image.data() + image_area * 2, (float *)bgrChannels[2].data, single_chn_size);

Hi, Im from Brasil and never reach the SMS from Baidu. Can you share the models/weights on Google Drive, please? Thanks.

./build/facefusion-onnxrun --weights ./weights/ --out ./sample_out/ --source ./images/1.jpg --target ./images/2.jpg weights_dir: /Data2/wyw/facefusion-onnxrun-add_cmake/weights out_dir: /Data2/wyw/facefusion-onnxrun-add_cmake/sample_out Yolov8Face input name: images Yolov8Face output name: output0 Face68Landmarks input name: input Face68Landmarks ouput name: landmarks_xyscore Face68Landmarks...

Hi @hpc203, 这个项目真的是太棒了,非常感谢您的项目。 我在使用C++版本代码,在高版本onnxruntime api上发现存在部分代码笔误以及内存越界问题,因此我对代码进行了修复,现在它可以很好的运行在我的linux和macos系统上,因此我将我的代码贡献到上游,也希望能帮助到更多人。

换脸后口型会张不开,幅度特别小,应该怎么处理

cv::Point2f pt; pt.x = pdata[5 * num_box + i] * ratio_width; pt.y = pdata[6 * num_box + i] * ratio_height; boxAndMarks.landmarks.push_back(pt); pt.x = pdata[8 * num_box + i] * ratio_width;...