Kernel size can‘t be greater than actual input size的解决方案

最新推荐文章于 2023-03-07 14:22:22 发布
最新推荐文章于 2023-03-07 14:22:22 发布
阅读量5.1k 收藏 4
点赞数 3
文章标签： 深度学习 神经网络 人工智能 python
版权声明：本文为博主原创文章，遵循 CC 4.0 BY-SA 版权协议，转载请附上原文出处链接和本声明。
本文链接：https://blog.csdn.net/weixin_61745097/article/details/128642752
版权

报错信息：RuntimeError: Calculated padded input size per channel: (49). Kernel size: (64). Kernel size can't be greater than actual input size

解决方案：不用修改卷积核大小，修改stride和padding即可

修改前：

self.conv1 = nn.Conv1d(in_channels=3, out_channels=3, kernel_size=64, stride=16, padding=24)

修改后：

self.conv1 = nn.Conv1d(in_channels=3, out_channels=3, kernel_size=64, stride=2, padding=60)

小白小王
关注
  • 3
    点赞
  • 踩
  • 4
    收藏
    觉得还不错? 一键收藏
  • 0
    评论
RuntimeError: Calculated padded input size per channel: (1 x 1). Kernel size: (5 x 5). Kernel size c
good18Levin的博客
06-25 7404
torchvision.model.inception_v3报错
Calculated padded input size per channel: (99 x 4). Kernel size: (129 x 1). Kernel size can‘t be gre
qq_46079584的博客
06-17 5184
报错信息。
AtarodiAbbasali_SUT_AdvDSP_TH1.2.zip_Greater
07-15
It solved the contradiction of the convergence rate, tracking capabilities and the steady-state error, but when the error close to zero, step size &#61549 changes too much, that lead to greater step in steady state
基于Arduino：球轮机器人（完整程序+3D建模文件）.zip
12-14
主程序： #include "gyro.h" #include "pid.h" #include "motor.h" #define dt 10 // in ms /* Object definitions and settings */ Gyro myGyro(dt); PIDControl pid(1.0,0,0.005,-255,255,dt); OmniMotorDriver omd; /* Deadzone definition. With a rotation of less than this value, * the robot will stand still */ double deadzone = 1.5; /* Buffers for timing, gyro data and PID outputs */ long lastMillis; double xAngle, yAngle; double xOut, yOut; /* Main setup call * Initializes Serial, the IMU and prepares the motor driver, * calbriates the IMU and enables the PID controller */ void setup() { pinMode(13, OUTPUT); Serial.begin(115200); if(!myGyro.init()) { while(1) { digitalWrite(13, HIGH); delay(1000); digitalWrite(13, LOW); delay(2000); } } myGyro.calibrate(); omd.init(); pid.enable(); } /* Main loop * The main loop connects all the libraries and classes together, * in that it describes the order of function calls that is necessary * for the function of the robot */ void loop() { /* Save the current point in time in lastMillis */ lastMillis = millis(); /* Get Data from Gyro */ myGyro.update(&xAngle, &yAngle); /* Compute PID Output with the data */ pid.compute(xAngle, yAngle, &xOut, &yOut); /* Plot Angle on Serial Plotter Serial.print(xAngle); Serial.print(" "); Serial.print(yAngle); Serial.println(" "); */ /* Plot Angle on Serial Plotter Serial.print(xOut); Serial.print(" "); Serial.println(yOut); */ /* If any calculated angle is larger than the deadzone */ if(abs(xAngle) > deadzone || abs(yAngle) > deadzone) { /* Actually drive the motors */ omd.drive(xOut, yOut); /* Otherwise, stop them */ } else omd.stop(); /* Wait until the next dt cycle * if processing took less than dt, wait for the remaining time, * otherwise wait for the next full cycle to start * This is needed for gyro and PID accuracy */ int pTime = millis() - lastMillis; int resttime = (dt > pTime
pytorch模型提示超出内存RuntimeError: CUDA out of memory.
01-20
跑模型时出现RuntimeError: CUDA out of memory.错误 查阅了许多相关内容，原因是：GPU显存内存不够 简单总结一下解决方法： 将batch_size改小。 取torch变量标量值时使用item()属性。 可以在测试阶段添加如下代码： with torch.no_grad(): # 停止自动反向计算梯度 参考： https://wisdomai.xyz/tool/pytorch/archives/2301 https://ptorch.com/news/160.html 作者：菜叶儿掉啦
Inception-v3报错：RuntimeError: Calculated padded input size per channel: (3 x 3). Kernel size: (5 x 5)
热门推荐
xjtu_yyl的博客
09-28 2万+
在PyTorch1.7.1环境下运行PyTorch官方提供的ImageNet预训练模型Inception-v3报错如下： RuntimeError: Calculated padded input size per channel: (3 x 3). Kernel size: (5 x 5). Kernel size can't be greater than actual input size 报错原因是我混用了其他模型的预处理transform方法，导致图片大小与网络结构不匹配。解决办法是采用PyTo
yolov5 yaml配置问题
set_jmp的博客
11-25 816
yolov5配置文件问题
调试代码错误记录 9.27.1
stujoy的博客
09-27 1552
RuntimeError: Calculated padded input size per channel: (4). Kernel size: (5). Kernel size can't be greater than actual input size 记录：修改basic_layers.py 62行kernel_size = 4 错误变了，下个错误见
Calculated padded input size per channel: (3 x 3). Kernel size: (5 x 5). Kernel size can‘t be greate
AI浩
06-10 8162
将图片的Resize改为（299，299）
.pth转.onnx出现问题解决。（记录，仅供参考）
sadjhaksdas的博客
03-07 1546
pth权重模型转换onnx
A greater than B 2_travel5gx_python_Greater_
10-02
Pycharm、Python程序编写以用来判断A大于B的正误、计算圆的周长、面积。
Greater than 2 kW all-passive fiber Raman amplifier with good beam quality
02-21
We report a 2 kW all-fiberized Raman fiber amplifier with efficient brightness enhancement based on the graded-index fiber. The maximum power output reaches up to 2.034 kW centered at 1130 nm, with a conversion efficiency of 79.35% with respect to the injected pump power. To the best of our knowledge, this is the highest conversion efficiency obtained for any Raman laser system using graded-index fiber. An optimized fiber combiner adopting graded-index fiber as the pigtail fiber was fabricated,
Laser-induced damage of high reflectors for Ti:sapphire laser system
02-12
A broadband (~176 nm, R>98%, 'lambda'0=800 nm) and high laser-induced damage threshold (LIDT=2.4J/cm2) TiO2/HfO2/SiO2 high reflector (HR) for Ti:sapphire chirped-pulse amplification (CPA) laser system is fabricated by the electron beam evaporation. The refractive index and extinction coefficient of TiO2 and HfO2 films are calculated from single-layer films' transmittance spectra. The properties of HR are mainly determined by the high refractive index material. The high refractive index leads
A greater than B 2_forthe2q_python_Greater_greaterthanpython_
09-29
python
pytorch dataloader 取batch_size时候出现bug的解决方式
12-20
1、 RuntimeError: invalid argument 0: Sizes of tensors must match except in dimension 0. Got 342 and 281 in dimension 3 at /pytorch/aten/src/TH/generic/THTensorMoreMath.cpp:1333 2、 RuntimeError: invalid argument 0: Sizes of tensors must match except in dimension 0. Got 3 and 1 in dimension 1 bug1
greater-than:验证值是否大于比较器值
06-23
var isGreaterThan = require ( 'validate.io-greater-than' ) ; isGreaterThan( 值，比较器) 验证value是否大于comparator值。 var a = 10 , b = 0 ; var bool = isGreaterThan ( a , b ) ; // returns true ...
源码 论文_PointNet论文笔记及源码解析
weixin_31358991的博客
01-04 946
PointNet论文笔记Tensor flow:https://github.com/charlesq34/pointnetpytorch复现：https://github.com/fxia22/pointnet.pytorchPointNet是来自斯坦福的Charles同学2017年在CVPR上的论文，这篇文章提出一种对点云运用深度学习的架构。该网络能够完成三维数据的分类，部分分割和语义分割任务...
python2代码转换python3遇到的问题
sinat_33487968的博客
07-10 994
1.SyntaxError: Non-ASCII character '\xe5' in file【现象】在编写Python时，当使用中文输出或注释时运行脚本，会提示错误信息：SyntaxError: Non-ASCII character '\xe5' in file *******【原因】python的默认编码文件是用的ASCII码，而你的python文件中使用了中文等非英语字符。【解决办法】...
Inception v3 输入图片尺寸大小
qq_34789262的博客
09-30 2万+
pytorch官网表示输入图片尺寸最少224，自己用Inception v3开始设置的size是224， 代码简单如下： import torchvision.models as models from torchsummary import summary import torch Incep = models.Inception3() Incep.fc = nn.Linear(2048,...
Calculated padded input size per channel: (3 x 3). Kernel size: (4 x 4). Kernel size can't be greater than actual input size
最新发布
04-23
哦，好的。让我来解答这个问题。在这种情况下，我们需要进行填充操作来使得卷积核可以和输入数据进行卷积操作。假设输入数据的尺寸为HxWxC（其中，H表示高度，W表示宽度，C表示通道数），卷积核的尺寸为KHxKW，同时，我们需要进行填充操作以确保卷积核可以覆盖到输入数据的边缘区域，那么，填充后的输入数据尺寸为：(H+(KH-1))x(W+(KW-1))x C。因此，在本问题中，填充后的尺寸为：(3+3-1) x (3+3-1) = 5 x 5。因为卷积核的尺寸为4x4，所以它不能大于填充后的输入数据尺寸，也就是说，不能大于5x5，否则无法进行卷积操作。

“相关推荐”对你有帮助么？

  • 非常没帮助
  • 没帮助
  • 一般
  • 有帮助
  • 非常有帮助
提交
评论
添加红包

请填写红包祝福语或标题

个

红包个数最小为10个

元

红包金额最低5元

当前余额3.43元 前往充值 >
需支付：10.00元
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付元
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明：

1.余额是钱包充值的虚拟货币，按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载，可以购买VIP、付费专栏及课程。

余额充值