Comments (5)
这个问题在给出的问题讨论链接里已经解决。
https://devtalk.nvidia.com/default/topic/1005119/cudnn-v6-int8-convolution-failing-with-cudnn_status_not_supported/
cudnn_conv_int8.cc文件main函数修改代码如下即可:
#define FIX
// input dimensions
size_t width = 960;
size_t height = 600;
#ifdef FIX
size_t channels = 4;
#else
size_t channels = 3;
#endif
int batch_size = 1;
// Create layer architecture
#ifdef FIX
int out_channels = 4;
#else
int out_channels = 1;
#endif
int kernel_size = 3;
原因是因为cudnn INT8函数的输入和输出必须是4的倍数,文档原文为"Input and output features maps must be multiple of 4",详细见文档
from tensorrt_tutorial.
@LitLeo 修改后还是出现同样的问题:
Begin forwrad pass
CUDNN failure: CUDNN_STATUS_ARCH_MISMATCH
cudnn_conv_int8.cpp:236
Aborting...
我在win7 系统 cudnn_v6.0 cuda8.0 vs2013运行.
这和平台有关吗,为什么还是无法运行
from tensorrt_tutorial.
你是什么显卡?INT8需要计算能力>=6.1的显卡。比如GTX 1080,P4,P40。
from tensorrt_tutorial.
噢噢 我的是GTX750的 那是显卡的问题了 谢谢哈
from tensorrt_tutorial.
@LitLeo
测试demo跑通了,两个数据类,float32占用显存1000M左右,时间为0.4ms ,而int8占用显存通用为1000M,但是时间却达到2.2ms,我的显卡GTX1070 cudnn_v6.0
这是为何??
from tensorrt_tutorial.
Related Issues (11)
- fatal error: cuda.h: 没有那个文件或目录
- 在网络中插入自己定义的layer时,如何获取前一层的输出作为自定义层的输入?
- tensorrt 7.2.2.3版本
- 是否有关于支持custom layer QDQ INT8 explicitly quantization 相关的教程或示例? HOT 1
- 如何处理网络模型中的BatchNorm层? HOT 4
- 怎么在centos 下面安装tensorrt HOT 1
- TensorRT 可以部署到Windows 平台吗? HOT 1
- How use mtcnn to call tensorRT?
- 您好,您的网站是不是挂掉了?
- CUDA_R_32I was not declared
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from tensorrt_tutorial.