Downloads ยท 30 days
0
ncnn
camenduru/ncnn
ncnn is a machine learning model from camenduru. Use it for the machine learning task on the model card, and read the license before you ship it in a product.
Access
Public
Updated Sep 13, 2023
Repo size
279 MB
Likes
1
Public
Repo makeup
Click a slice to open those files.
Other288 MB ยท 66%
At a glance
- Access
- Public
- Created
- Sep 13, 2023
- Updated
- Sep 13, 2023
- SHA
- be903e21
Model card
From the Hugging Face model README
ncnn
ncnn is a high-performance neural network inference computing framework optimized for mobile platforms. ncnn is deeply considerate about deployment and uses on mobile phones from the beginning of design. ncnn does not have third party dependencies. It is cross-platform, and runs faster than all known open source frameworks on mobile phone cpu. Developers can easily deploy deep learning algorithm models to the mobile platform by using efficient ncnn implementation, create intelligent APPs, and bring the artificial intelligence to your fingertips. ncnn is currently being used in many Tencent applications, such as QQ, Qzone, WeChat, Pitu and so on.
ncnn ๆฏไธไธชไธบๆๆบ็ซฏๆ่ดไผๅ็้ซๆง่ฝ็ฅ็ป็ฝ็ปๅๅ่ฎก็ฎๆกๆถใ ncnn ไป่ฎพ่ฎกไนๅๆทฑๅป่่ๆๆบ็ซฏ็้จ็ฝฒๅไฝฟ็จใ ๆ ็ฌฌไธๆนไพ่ต๏ผ่ทจๅนณๅฐ๏ผๆๆบ็ซฏ cpu ็้ๅบฆๅฟซไบ็ฎๅๆๆๅทฒ็ฅ็ๅผๆบๆกๆถใ ๅบไบ ncnn๏ผๅผๅ่ ่ฝๅคๅฐๆทฑๅบฆๅญฆไน ็ฎๆณ่ฝปๆพ็งปๆคๅฐๆๆบ็ซฏ้ซๆๆง่ก๏ผ ๅผๅๅบไบบๅทฅๆบ่ฝ APP๏ผๅฐ AI ๅธฆๅฐไฝ ็ๆๅฐใ ncnn ็ฎๅๅทฒๅจ่ พ่ฎฏๅคๆฌพๅบ็จไธญไฝฟ็จ๏ผๅฆ๏ผQQ๏ผQzone๏ผๅพฎไฟก๏ผๅคฉๅคฉ P ๅพ็ญใ
ๆๆฏไบคๆต QQ ็พค๏ผ637093648 (่ถ ๅคๅคงไฝฌ) ็ญๆก๏ผๅทๅทๅทๅทๅท ๏ผๅทฒๆปก๏ผ
Pocky QQ ็พค๏ผMLIR YES!๏ผ: 677104663(่ถ ๅคๅคงไฝฌ) ็ญๆก๏ผmulti-level intermediate representation
Telegram Group https://t.me/ncnnyes
Discord Channel https://discord.gg/YRsxgmF
Current building status matrix
| System | CPU (32bit) | CPU (64bit) | GPU (32bit) | GPU (64bit) |
|---|---|---|---|---|
| Linux (GCC) |  |  | โ |  |
| Linux (Clang) |  |  | โ |  |
| Linux (ARM) |  |  | โ | โ |
| Linux (MIPS) |  |  | โ | โ |
| Linux (RISC-V) | โ |  | โ | โ |
| Linux (LoongArch) | โ |  | โ | โ |
| Windows |  |  | โ |  |
| Windows (ARM) |  |  | โ | โ |
| macOS | โ |  | โ |  |
| macOS (ARM) | โ |  | โ |  |
| Android |  |  |  |  |
| Android-x86 |  |  |  |  |
| iOS |  | | โ |  |
| iOS Simulator |  | | โ |  |
| WebAssembly |  | โ | โ | โ |
| RISC-V GCC/Newlib |  |  | โ | โ |
Support most commonly used CNN network
ๆฏๆๅคง้จๅๅธธ็จ็ CNN ็ฝ็ป
- Classical CNN: VGG AlexNet GoogleNet Inception ...
- Practical CNN: ResNet DenseNet SENet FPN ...
- Light-weight CNN: SqueezeNet MobileNetV1 MobileNetV2/V3 ShuffleNetV1 ShuffleNetV2 MNasNet ...
- Face Detection: MTCNN RetinaFace scrfd ...
- Detection: VGG-SSD MobileNet-SSD SqueezeNet-SSD MobileNetV2-SSDLite MobileNetV3-SSDLite ...
- Detection: Faster-RCNN R-FCN ...
- Detection: YOLOv2 YOLOv3 MobileNet-YOLOv3 YOLOv4 YOLOv5 YOLOv7 YOLOX ...
- Detection: NanoDet
- Segmentation: FCN PSPNet UNet YOLACT ...
- Pose Estimation: SimplePose ...
HowTo
how to build ncnn library on Linux / Windows / macOS / Raspberry Pi3, Pi4 / POWER / Android / NVIDIA Jetson / iOS / WebAssembly / AllWinner D1 / Loongson 2K1000
- Build for Linux / NVIDIA Jetson / Raspberry Pi3, Pi4 / POWER
- Build for Windows x64 using VS2017
- Build for macOS
- Build for ARM Cortex-A family with cross-compiling
- Build for Hisilicon platform with cross-compiling
- Build for Android
- Build for iOS on macOS with xcode
- Build for WebAssembly
- Build for AllWinner D1
- Build for Loongson 2K1000
- Build for Termux on Android
- Build for QNX
download prebuild binary package for android and ios
use ncnn with alexnet with detailed steps, recommended for beginners :)
ncnn ็ปไปถไฝฟ็จๆๅ alexnet ้ๅธฆ่ฏฆ็ปๆญฅ้ชค๏ผๆฐไบบๅผบ็ๆจ่ :)
use netron for ncnn model visualization
out-of-the-box web model conversion
ncnn param and model file spec
ncnn operation param weight table
how to implement custom layer step by step
FAQ
Features
- Supports convolutional neural networks, supports multiple input and multi-branch structure, can calculate part of the branch
- No third-party library dependencies, does not rely on BLAS / NNPACK or any other computing framework
- Pure C++ implementation, cross-platform, supports Android, iOS and so on
- ARM NEON assembly level of careful optimization, calculation speed is extremely high
- Sophisticated memory management and data structure design, very low memory footprint
- Supports multi-core parallel computing acceleration, ARM big.LITTLE CPU scheduling optimization
- Supports GPU acceleration via the next-generation low-overhead Vulkan API
- Extensible model design, supports 8bit quantization and half-precision floating point storage, can import caffe/pytorch/mxnet/onnx/darknet/keras/tensorflow(mlir) models
- Support direct memory zero copy reference load network model
- Can be registered with custom layer implementation and extended
- Well, it is strong, not afraid of being stuffed with ๅท QvQ
ๅ่ฝๆฆ่ฟฐ
- ๆฏๆๅท็งฏ็ฅ็ป็ฝ็ป๏ผๆฏๆๅค่พๅ ฅๅๅคๅๆฏ็ปๆ๏ผๅฏ่ฎก็ฎ้จๅๅๆฏ
- ๆ ไปปไฝ็ฌฌไธๆนๅบไพ่ต๏ผไธไพ่ต BLAS/NNPACK ็ญ่ฎก็ฎๆกๆถ
- ็บฏ C++ ๅฎ็ฐ๏ผ่ทจๅนณๅฐ๏ผๆฏๆ Android / iOS ็ญ
- ARM Neon ๆฑ็ผ็บง่ฏๅฟไผๅ๏ผ่ฎก็ฎ้ๅบฆๆๅฟซ
- ็ฒพ็ป็ๅ ๅญ็ฎก็ๅๆฐๆฎ็ปๆ่ฎพ่ฎก๏ผๅ ๅญๅ ็จๆไฝ
- ๆฏๆๅคๆ ธๅนถ่ก่ฎก็ฎๅ ้๏ผARM big.LITTLE CPU ่ฐๅบฆไผๅ
- ๆฏๆๅบไบๅ จๆฐไฝๆถ่็ Vulkan API GPU ๅ ้
- ๅฏๆฉๅฑ็ๆจกๅ่ฎพ่ฎก๏ผๆฏๆ 8bit ้ๅ ๅๅ็ฒพๅบฆๆตฎ็นๅญๅจ๏ผๅฏๅฏผๅ ฅ caffe/pytorch/mxnet/onnx/darknet/keras/tensorflow(mlir) ๆจกๅ
- ๆฏๆ็ดๆฅๅ ๅญ้ถๆท่ดๅผ็จๅ ่ฝฝ็ฝ็ปๆจกๅ
- ๅฏๆณจๅ่ชๅฎไนๅฑๅฎ็ฐๅนถๆฉๅฑ
- ๆฉ๏ผๅพๅผบๅฐฑๆฏไบ๏ผไธๆ่ขซๅกๅท QvQ
supported platform matrix
- โ = known work and runs fast with good optimization
- โ๏ธ = known work, but speed may not be fast enough
- โ = shall work, not confirmed
- / = not applied
| Windows | Linux | Android | macOS | iOS | |
|---|---|---|---|---|---|
| intel-cpu | โ๏ธ | โ๏ธ | โ | โ๏ธ | / |
| intel-gpu | โ๏ธ | โ๏ธ | โ | โ | / |
| amd-cpu | โ๏ธ | โ๏ธ | โ | โ๏ธ | / |
| amd-gpu | โ๏ธ | โ๏ธ | โ | โ | / |
| nvidia-gpu | โ๏ธ | โ๏ธ | โ | โ | / |
| qcom-cpu | โ | โ๏ธ | โ | / | / |
| qcom-gpu | โ | โ๏ธ | โ๏ธ | / | / |
| arm-cpu | โ | โ | โ | / | / |
| arm-gpu | โ | โ | โ๏ธ | / | / |
| apple-cpu | / | / | / | โ๏ธ | โ |
| apple-gpu | / | / | / | โ๏ธ | โ๏ธ |
| ibm-cpu | / | โ๏ธ | / | / | / |
Project examples
- https://github.com/nihui/ncnn-android-squeezenet
- https://github.com/nihui/ncnn-android-styletransfer
- https://github.com/nihui/ncnn-android-mobilenetssd
- https://github.com/moli232777144/mtcnn_ncnn
- https://github.com/nihui/ncnn-android-yolov5
- https://github.com/xiang-wuu/ncnn-android-yolov7
- https://github.com/nihui/ncnn-android-scrfd ๐คฉ
- https://github.com/shaoshengsong/qt_android_ncnn_lib_encrypt_example
<img src="https://github.com/nihui/ncnn-assets/raw/master/20181217/ncnn-2.jpg" height ="230"/><img src="https://github.com/nihui/ncnn-assets/raw/master/20181217/4.jpg" height ="230"/><img src="https://github.com/nihui/ncnn-assets/raw/master/20181217/ncnn-33.jpg" height ="230"/><img src="https://github.com/nihui/ncnn-assets/raw/master/20181217/ncnn-m.png" height ="230"/><img src="https://github.com/nihui/ncnn-android-yolov5/raw/master/screenshot.jpg" height ="230"/><img src="https://github.com/nihui/ncnn-android-scrfd/raw/master/screenshot.jpg" height ="230"/><br>
- <img src="https://user-images.githubusercontent.com/13585785/189326958-f5a8d6f8-caef-49bf-88da-ae494371195d.jpg" width ="700"/>
-
https://github.com/mizu-bai/ncnn-fortran Call ncnn from Fortran
-
https://github.com/k2-fsa/sherpa Use ncnn for real-time speech recognition (i.e., speech-to-text); also support embedded devices and provide mobile Apps (e.g., Android App)
License
- Created
- Sep 13, 2023
- Updated
- Sep 13, 2023