文件列表:

assets/
lib/
pitorch/
report.md

# Pitorch The first AI framewrok designed by myself, which is the final project of class "Programming in Artificial Intelligence" ## TODO :pencil2: - [ ] Data preprocessing - [ ] Accelerate training ## Code structure The code is orgnized as follows: ```bash ├── data │ └── MNIST ├── lib # C++ and Cuda code of pitorch │ ├── cpu_func.h # functions working on cpu │ ├── cpu_func.cc │ ├── gpu_func.h # functions working on gpu │ ├── gpu_func.cu │ ├── Tensor.h # Tensor class │ ├── Tensor.cu │ ├── tensor_func.h # wrapped cpu/gpu functions for Tensor │ ├── tensor_func.cu │ ├── utils.h # utils for error detection │ ├── pybind.cpp │ └── setup.py # install the lib │ ├── pitorch # Python code of pitorch │ ├── __init__.py │ ├── basic_operator.py # basic elements for Computinng Graph │ ├── Pisor.py # difination of Pisor and Operators │ ├── autodiff.py # gradient back propagation │ ├── optimizer.py # optimizer zone │ ├── Unittest.py # validate C++/Cuda operators │ ├── test_forward.py │ ├── test_backward.py │ ├── train_fc_net.py # train a fully connected network │ ├── train_conv_net.py # train a convolution network │ └── utils.py # some util function (not finished yet) └── README.md ``` ## Usage ### Installation ```bash conda create -n pitorch python=3.9 conda activate pitorch # install pytorch according to your cuda version. # (We just use torch.utils.cpp_extension to conveniently bind cuda code to python, and not utilize torch to conduct any operation) cd lib python setup.py develop ``` ### Running demo You can start training on MNIST as follows. ```bash cd pitorch # train convolution network python train_conv_net.py # train fully connected network python train_fc_net.py ``` On the bottom of training script, you can tweak hyper-parameters, select optimizers as well as change device. Here is an example of training log:

近期下载者：

相关文件：

评论：[我要评论] [举报此文件]

收藏者：