When the kernel size is 7×7, as with convolution where the kernel size is 3×3, the two outputs of MB are not fully pipelined. These two outputs need to accumulate 6 and 2 clock cycles respectively, but the clock ratio of their outputs is still 3:1, which means that the DSP utilization can still be maintained at a very high level. A proper zero-setting is still necessary for convolution core to obtain correct and error-free accumulation results.请帮我把上边这段话润色一下,让他读起来更像是论文里边的句子
时间: 2023-02-21 18:17:17 浏览: 41
当核大小为7×7时,与卷积核大小为3×3的情况相同,MB的两个输出并不完全管道化。这两个输出需要分别累积6和2个时钟周期,但它们输出的时钟比仍为3:1,这意味着DSP利用率仍可保持在很高的水平。为了获得正确且错误率低的累积结果,仍需对卷积核进行适当的零设置。
相关问题
请帮我润色一下下边双引号里边的这段话,注意不是翻译,让他读起来更像是论文里边的句子。“When the kernel size is 7×7, as with convolution where the kernel size is 3×3, the two outputs of MB are not fully pipelined. These two outputs need to accumulate 6 and 2 clock cycles respectively, but the clock ratio of their outputs is still 3:1, which means that the DSP utilization can still be maintained at a very high level. A proper zero-setting is still necessary for convolution core to obtain correct and error-free accumulation results.”
当核大小为7×7时,与卷积(其中核大小为3×3)相同,MB 的两个输出不能完全流水线处理。这两个输出分别需要累积6和2个时钟周期,但是它们输出的时钟比例仍然为3:1,这意味着 DSP 的利用率仍然可以保持在非常高的水平。对卷积核进行适当的清零仍然是必要的,以获得正确无误的累积结果。
employed 9×9 convolution kernels to replace part of the 3×3 convolution kernels
in their neural network architecture.
This change was made in order to increase the receptive field of the network, allowing it to better capture spatial relationships between pixels in an image. By using larger kernels, the network is able to consider a wider area of the image when making predictions, leading to improved accuracy.
However, using larger kernels also increases the computational cost of the network, as more parameters need to be learned. Therefore, the researchers used a combination of 9×9 and 3×3 kernels, with the larger kernels being used in the deeper layers of the network where the receptive field is larger, and the smaller kernels being used in the shallower layers where the receptive field is smaller.
Overall, this approach resulted in a significant improvement in performance on benchmark datasets, demonstrating the effectiveness of using larger kernels in certain parts of a convolutional neural network.
相关推荐
![zip](https://img-home.csdnimg.cn/images/20210720083736.png)
![zip](https://img-home.csdnimg.cn/images/20210720083736.png)
![rar](https://img-home.csdnimg.cn/images/20210720083606.png)
![](https://csdnimg.cn/download_wenku/file_type_ask_c1.png)
![](https://csdnimg.cn/download_wenku/file_type_ask_c1.png)
![](https://csdnimg.cn/download_wenku/file_type_ask_c1.png)
![](https://csdnimg.cn/download_wenku/file_type_ask_c1.png)
![](https://csdnimg.cn/download_wenku/file_type_ask_c1.png)
![](https://csdnimg.cn/download_wenku/file_type_ask_c1.png)
![](https://csdnimg.cn/download_wenku/file_type_ask_c1.png)
![](https://csdnimg.cn/download_wenku/file_type_ask_c1.png)
![](https://csdnimg.cn/download_wenku/file_type_ask_c1.png)
![](https://csdnimg.cn/download_wenku/file_type_ask_c1.png)
![](https://csdnimg.cn/download_wenku/file_type_ask_c1.png)
![](https://csdnimg.cn/download_wenku/file_type_ask_c1.png)
![](https://csdnimg.cn/download_wenku/file_type_ask_c1.png)