Load fp16 weight from onnx

microsoft / nnfusion

A flexible and efficient deep neural network (DNN) compiler that generates high-performance executable from a DNN model description.

MIT License

958 stars 162 forks source link

Closed mzmssg closed 2 years ago