PyTorch framework A simple and complete framework for PyTorch, providing a variety of data loading and simple task solutions that are easy to extend and migrate

Last update: Dec 19, 2021

Related tags

Pytorch Utilities PyTorch-framework

Overview

PyTorch framework

一个简单且完整的PyTorch的框架，提供了各种数据的加载以及简单任务的解决方案，易于扩展和迁移。

1.该框架提供了各种数据类型的加载(.wav .mat .jpg .csv .npy)方案。

2.该框架提供了简单分类任务和回归任务的解决方案，以及几个基础模型：CNN、RNN、Attention （ResNet、LSTM、Transformer-encoder)

3.该框架是一个简单且完整的框架，只保留了必要的部分并有详细的注释，方便阅读和理解。

并且解耦了各个模块，易于扩展和迁移。迁移到其他任务上只需要更改dataloader和model部分 (还有损失函数)。

用法：

训练和验证

python main.py --dataset_path ./data/audio/wav2vec/ --model_path  wav2vec --feature wav2vec --feature_dim 768 --task regression --model lstm

python main.py --dataset_path ./data/vision/AU/ --model_path  AU --feature AU --feature_dim 34 --task regression --model lstm

python main.py --dataset_path ./data/vision/vggface/ --model_path  vggface --feature vggface --feature_dim 128 --task regression --model lstm

python main.py --dataset_path ./data/vision/image/ --model_path  image --feature image  --task classification --model resnet

测试

python test.py --dataset_path ./data/audio/wav2vec/ --model_path  ./model/wav2vec_regression_1.pth --feature wav2vec --feature_dim 768 --task regression --model lstm

多卡训练

CUDA_VISIBLE_DEVICES=0,1 python main.py --dataset_path ./data/vision/image/ --model_path  image --feature image  --task classification --model resnet --parallel

CUDA_VISIBLE_DEVICE 和 parallel 搭配使用，单用 parallel 会默认使用所有卡。

如果有任何问题，欢迎联系我（[email protected])

PyTorch framework A simple and complete framework for PyTorch, providing a variety of data loading and simple task solutions that are easy to extend and migrate

Related tags

Overview

PyTorch framework

Owner

Cong Cai

Distiller is an open-source Python package for neural network compression research.

TorchSSL: A PyTorch-based Toolbox for Semi-Supervised Learning

Training RNNs as Fast as CNNs (https://arxiv.org/abs/1709.02755)

Tacotron 2 - PyTorch implementation with faster-than-realtime inference

higher is a pytorch library allowing users to obtain higher order gradients over losses spanning training loops rather than individual training steps.

An optimizer that trains as fast as Adam and as good as SGD.

A PyTorch repo for data loading and utilities to be shared by the PyTorch domain libraries.

Tez is a super-simple and lightweight Trainer for PyTorch. It also comes with many utils that you can use to tackle over 90% of deep learning projects in PyTorch.

PyTorch to TensorFlow Lite converter

A tiny scalar-valued autograd engine and a neural net library on top of it with PyTorch-like API

Code for paper "Energy-Constrained Compression for Deep Neural Networks via Weighted Sparse Projection and Layer Input Masking"

PyGCL: Graph Contrastive Learning Library for PyTorch

PyTorch Lightning Optical Flow models, scripts, and pretrained weights.

A PyTorch implementation of L-BFGS.

Pytorch bindings for Fortran

Training PyTorch models with differential privacy

A Pytorch Implementation for Compact Bilinear Pooling.

Pytorch implementation of Distributed Proximal Policy Optimization

High-level batteries-included neural network training library for Pytorch

You like pytorch? You like micrograd? You love tinygrad! ❤️