TNN: developed by Tencent Youtu Lab and Guangying Lab, a uniform deep learning inference framework for mobile、desktop and server. TNN is distinguished by several outstanding features, including its cross-platform capability, high performance, model compression and code pruning. Based on ncnn and Rapidnet, TNN further strengthens the support and performance optimization for mobile devices, and also draws on the advantages of good extensibility and high performance from existed open source efforts. TNN has been deployed in multiple Apps from Tencent, such as Mobile QQ, Weishi, Pitu, etc. Contributions are welcome to work in collaborative with us and make TNN a better framework.
1. 使用环境(environment)
OS Version: latest
2. Github版本
commit(optional):
3. 详细描述bug 情况 (Describe the bug) A clear and concise description of what the bug is.
有gridsample的模型转tnn后与onnx对不齐,去掉gridsample后,tnn和onnx可以对齐,并通过model check,所以可以确定是gridsample 的实现问题。 (如果可以的话,请上传原始的模型文件)
4. 日志(Log)
5. 截图(Screenshots) If applicable, add screenshots to help explain your problem. 插帧前的数据: 插帧后的数据: 插帧后的数据边界一行像素异常,且其他像素整体也像内部缩了一圈,onnx模型的结果是正常的。