p-tuning_NLU

Overview

这个小项目是受乐于分享的苏剑林大佬这篇p-tuning 文章启发，也实现了个使用P-tuning进行NLU分类的任务，思路是一样的，prompt实现方式有不同，这里是将[unused*]的embeddings参数抽取出用于初始化prompt_embed后，再接一个lstm和mlp用于关联各prompt，与最初p-tuning提出《GPT Understands, Too》的实现一样，结果显示在few-shot上p-tuning非常接近finetune效果。

Dataset

数据是情感分类，下载地址百度网盘提取码：osja

Evaluation

1. finetune

python few_shot_finetune.py

测试集效果：

epoch: 0 - acc: 0.897679 - best_test_acc: 0.8976788252013264
epoch: 1 - acc: 0.876362 - best_test_acc: 0.8976788252013264
epoch: 2 - acc: 0.884889 - best_test_acc: 0.8976788252013264
epoch: 3 - acc: 0.884415 - best_test_acc: 0.8976788252013264
epoch: 4 - acc: 0.884415 - best_test_acc: 0.8976788252013264

全量参数对小样本进行finetune，仅1个epoch就收敛了

2. p-tuning

python few_shot_ptuning.py

测试集效果：

epoch: 0 - acc: 0.546660 - best_test_acc: 0.5466603505447655
epoch: 1 - acc: 0.687826 - best_test_acc: 0.6878256750355282
epoch: 2 - acc: 0.737091 - best_test_acc: 0.7370914258645191
epoch: 3 - acc: 0.722406 - best_test_acc: 0.7370914258645191
epoch: 4 - acc: 0.776883 - best_test_acc: 0.7768829938417812
epoch: 5 - acc: 0.805306 - best_test_acc: 0.8053055423969683
epoch: 6 - acc: 0.833254 - best_test_acc: 0.8332543818095689
epoch: 7 - acc: 0.837991 - best_test_acc: 0.8379914732354334
epoch: 8 - acc: 0.854571 - best_test_acc: 0.8545712932259593
epoch: 9 - acc: 0.858361 - best_test_acc: 0.8583609663666508
epoch: 10 - acc: 0.856466 - best_test_acc: 0.8583609663666508
epoch: 11 - acc: 0.853150 - best_test_acc: 0.8583609663666508
epoch: 12 - acc: 0.868783 - best_test_acc: 0.8687825675035529
epoch: 13 - acc: 0.877309 - best_test_acc: 0.877309332070109
epoch: 14 - acc: 0.873993 - best_test_acc: 0.877309332070109
epoch: 15 - acc: 0.877783 - best_test_acc: 0.8777830412126955
epoch: 16 - acc: 0.882994 - best_test_acc: 0.8829938417811464
epoch: 17 - acc: 0.881573 - best_test_acc: 0.8829938417811464
epoch: 18 - acc: 0.889626 - best_test_acc: 0.8896257697773567
epoch: 19 - acc: 0.877783 - best_test_acc: 0.8896257697773567

仅prompt_embed和lstm及mlp去做p-tuning，20个epoch后接近收敛，acc=0.8896，略小于finetun的acc 0.8977

附上苏神结果对比：

p-tuning for few-shot NLU task

Related tags

Overview

p-tuning_NLU

Overview

Dataset

Evaluation

Owner

基于百度的语音识别，用python实现，pyaudio+pyqt

ByT5: Towards a token-free future with pre-trained byte-to-byte models

keras implement of transformers for humans

REST API for sentence tokenization and embedding using Multilingual Universal Sentence Encoder.

A natural language processing model for sequential sentence classification in medical abstracts.

LewusBot - Twitch ChatBot built in python with twitchio library

Flexible interface for high-performance research using SOTA Transformers leveraging Pytorch Lightning, Transformers, and Hydra.

A Word Level Transformer layer based on PyTorch and 🤗 Transformers.

Multi-Task Pre-Training for Plug-and-Play Task-Oriented Dialogue System

The RWKV Language Model

Simple Annotated implementation of GPT-NeoX in PyTorch

Pytorch-version BERT-flow: One can apply BERT-flow to any PLM within Pytorch framework.

Spam filtering made easy for you

Statistics and Mathematics for Machine Learning, Deep Learning , Deep NLP

Python library for processing Chinese text

PyTorch code for EMNLP 2019 paper "LXMERT: Learning Cross-Modality Encoder Representations from Transformers".

The NewSHead dataset is a multi-doc headline dataset used in NHNet for training a headline summarization model.

ChatterBot is a machine learning, conversational dialog engine for creating chat bots

Code for ACL 2020 paper "Rigid Formats Controlled Text Generation"

ELECTRA: Pre-training Text Encoders as Discriminators Rather Than Generators