Wengang Cao

a easy way to deploy

1

a easy way to deploy and export onnx

command: ```` python3 main_imagenet.py --data_path image --arch mobilenetv2 --n_bits_w 2 --n_bits_a 4 --channel_wise --weight 0.1 --act_quant ```` result: ``` Full quantization (W2A4) accuracy: 0.1419999897480011 ``` How to reproduce mobilenetv2 w2a4...

MegEngine 量化文档内容改进收集 & 追踪

4

需要添加 1. 如何创建一个量化的 Tensor 2. 量化训练时支持哪些量化类型 3. dump quantized 模型时哪些 op 支持哪些量化类型

When the input is of some size, such as 270x270, the following error will occur.

2

y = y + self.fuse_layers[i][j](x[j]) RuntimeError: The size of tensor a (68) must match the size of tensor b (72) at non-singleton dimension 3