# cpp-pytorch

**Repository Path**: lebranium/cpp-pytorch

## Basic Information

- **Project Name**: cpp-pytorch
- **Description**: No description available
- **Primary Language**: Unknown
- **License**: Not specified
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2020-08-08
- **Last Updated**: 2020-12-19

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

## cpp-pytorch
PyTroch 1.0 preview - LOADING A PYTORCH MODEL IN C++

Details refer to: https://shiftlab.github.io/pytorch_tutorials/advanced/cpp_export.html

Main requires: PyTroch 1.0, opencv, cmake

### STEP 1
CONVERTING YOUR PYTORCH MODEL TO TORCH SCRIPT and SERIALIZING YOUR SCRIPT MODULE TO A FILE

- run `python tracing.py` and get `model.pt`

### STEP 2
LOADING YOUR SCRIPT MODULE IN C++ and EXECUTING

1. Download LibTorch [here](https://pytorch.org/) and unzip

2. Cmake
```
mkdir build
cd build
cmake -DCMAKE_PREFIX_PATH=/Users/hankai/code/cpp-pytorch/libtorch ..
make
```

3. Run demo
```
./example-app ../model.pt ../dog.png ../synset_words.txt
```

Input image and predicted label: `n02108422 bull mastiff`, bingo!

![](dog.png)