curtainwang's repositories
BlockShuffleTest
BlockShuffle,就是在训练过程中使用分块打乱替代随机打乱的一种方法,即将原始数据按照数据长度进行排序,然后进行batch划分,在对batch训练进行打乱。这样操作,可以减少数据padding长度,缩短训练时长。
clip-training
Code to train CLIP model
ConSERT
Code for our ACL 2021 paper - ConSERT: A Contrastive Framework for Self-Supervised Sentence Representation Transfer
CPlusPlusThings
C++那些事
DeepLearningExamples
Deep Learning Examples
document-level-classification
超长文本分类(大于1000字);文档级/篇章级文本分类;主要是解决长距离依赖问题
EstateNameGenerator_Pytorch
从链家爬取全国楼盘信息,使用LSTM实现房地产楼盘名的AI生成
fprime
F' - A flight software and embedded systems framework
ImHex
A Hex Editor for Reverse Engineers, Programmers and people that value their eye sight when working at 3 AM.
MAE-pytorch
Unofficial PyTorch implementation of Masked Autoencoders Are Scalable Vision Learners
Megatron-DeepSpeed
Ongoing research training transformer language models at scale, including: BERT & GPT-2
minGPT
A minimal PyTorch re-implementation of the OpenAI GPT (Generative Pretrained Transformer) training
NLP-Data-Augmentation
NLP文本增强的两种方式:同义词替换(利用word2vec词表)和回译
NLP-Loss-Pytorch
Implementation of some unbalanced loss like focal_loss, dice_loss, DSC Loss, GHM Loss et.al
NLP-Series-sentence-embeddings
NLP句子编码、句子embedding、语义相似度:BERT_avg、BERT_whitening、SBERT、SmiCSE
NLP_pytorch_project_handwrite
Embedding, NMT, Text_Classification, Text_Generation, NER etc.
nlpcda
一键中文数据增强包 ; NLP数据增强、bert数据增强、EDA:pip install nlpcda
picoGPT
An unnecessarily tiny implementation of GPT-2 in NumPy.
ResnetGPT
用Resnet101+GPT搭建一个玩王者荣耀的AI
SimCSE
EMNLP'2021: SimCSE: Simple Contrastive Learning of Sentence Embeddings
SIMCSE_unsup
中文无监督SimCSE Pytorch实现
stanford_alpaca
Code and documentation to train Stanford's Alpaca models, and generate the data.
team-learning-nlp
主要存储Datawhale组队学习中“自然语言处理”方向的资料。
unilm-1
Large-scale Self-supervised Pre-training Across Tasks, Languages, and Modalities
vit-pytorch
Implementation of Vision Transformer, a simple way to achieve SOTA in vision classification with only a single transformer encoder, in Pytorch
word2vec_commented
Commented (but unaltered) version of original word2vec C implementation.