Qiu Huang
Qiu Huang
I foud that the "mean" of activations have not been used, how to define the "scaling factor" of activatons?
I can't load the 'part_1_cnn_genre.h5' via load_model function in keras.model ,but others can.Can you give me some tips?
### OpenVINO Version openvino 2024.3.0 + onnx 1.14.0 ### Operating System Ubuntu 20.04 (LTS) ### Device used for inference CPU ### Framework ONNX ### Model used resnet18 ### Issue description...
使用文档中,量化过程支持 "由深度学习框架 导出 量化模型, RKNN Toolkit 加载 并利用已有的 量化 信息 ,生成量化 RKNN模型。支持的深度学习框架(括号内为主要支持版本,请尽量使用对应版本生成的量化模 型): PyToch(v1.9.0)、 ONNX(Onnxruntime v1.5.1)、 Tensorflow、 TFLite。" 使用文档显示可以导入带有量化信息的模型,但在导入带有FakeQuantize的onnx模型时,导入失败。
untimeError: CUDA is available for PyTorch, but NNcF could not compile GPu quan tization extensions.
### 🐛 Describe the bug nccfv2.5.0 and torch1.9.1 can work normally on the PC, but do not work on the Linux system。 Are there any methods to quickly troubleshoot the...