提交 23a6f7ce 编写于 作者: H hanbuhe

code format

上级 0062bf52
......@@ -46,10 +46,8 @@ static Dtype find_max(Dtype* data, int num) {
return max;
}
// template <typename Dtype>
void quantify_filter(framework::Tensor* filter) {
DLOG << "quantilize_filter........";
float scale = 0;
......@@ -87,7 +85,6 @@ void quantify_filter(framework::Tensor* filter) {
chw_to_hwc<int8_t>(tmp_data, int_data, batch_size, channel, height, width);
delete tmp_data;
*(filter->fpga_args().scale_pointer()) = scale;
}
} // namespace fpga
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册