关于GCU、沐曦GPGPU、MLU、0卡V100资源4月7日恢复上架的公告>>> 关于共建具身智能开源数据集的倡议>>> 关于云脑任务中统一路径访问方式的公告>>> 关于将启智集群GPU资源迁移至智算集群的公告>>>
JDOpenISCT
/
dabnn
mirror of https://github.com/JDAI-CV/dabnn
Not watched
8
19
Fork 9
Code
Releases 16 Wiki Activity
Issues 0 Datasets Model Cloudbrain
You can not select more than 25 topics Topics must start with a chinese character,a letter or number, can include dashes ('-') and can be up to 35 characters long.
328 Commits
1 Branch
3.0 MiB
117 Download
 
 
 
 
 
 
Branch: master
Branches Tags
${ item.name }
Create branch ${ searchTerm }
from 'master'
${ noResults }
dabnn/common
daquexian e801568031 Add support for prelu 4 years ago
..
Shaper.cpp init 5 years ago
Shaper.h init 5 years ago
StrKeyMap.h Make onnx2bnn c++11 compatible, add onnx2bnn build ci 5 years ago
argh.h onnx2bnn: Update the method to discover binary convs, add '--strict' arg (not test) 5 years ago
baseline.h Update baseline bconv, it works if gemm is compatible only if weight_mat->h * weight_mat->w * weight_mat->c % 2 == 0 4 years ago
common_bitpack.h stash 4 years ago
dab.fbs support prelu 4 years ago
dab_generated.h support prelu 4 years ago
flatbuffers_helper.h Add support for prelu 4 years ago
helper.h Improve assertions, remove unused funcs 4 years ago
log_helper.h init 5 years ago
macros.h Upgrade model version to 2, remove the binarize layer 4 years ago
Delete task

Are you sure you want to delete this task? Once this task is deleted, it cannot be recovered.

cancel
confirm

我们推出了 dabnn，一个用 ARM 汇编重写了卷积，高度优化的二值网络 inference 框架。实验显示 dabnn 相对于现有的二值网络和浮点网络 inference 框架均能带来极大的加速。在运行单个二值卷积时，我们的速度是 BMXNet （除 dabnn 外唯一一个二值网络 inference 框架） 的 800%~2400%，在运行二值化的 ResNet-18 时，我们的速度是 BMXNet 的约 700%。

C++ C Markdown Text Shell other

Please read the following content carefully:

Dear OpenI User

Thank you for your continuous support to the Openl Qizhi Community AI Collaboration Platform. In order to protect your usage rights and ensure network security, we updated the Openl Qizhi Community AI Collaboration Platform Usage Agreement in January 2024. The updated agreement specifies that users are prohibited from using intranet penetration tools. After you click "Agree and continue", you can continue to use our services. Thank you for your cooperation and understanding.

For more agreement content, please refer to the《Openl Qizhi Community AI Collaboration Platform Usage Agreement》

Agree and continue
Disagree, exit