使用pytorch+transformers复现了SimCSE论文中的有监督训练和无监督训练方法

Related tags

Text Data & NLPSimCSE
Overview

SimCSE复现

项目描述

SimCSE是一种简单但是很巧妙的NLP对比学习方法,创新性地引入Dropout的方式,对样本添加噪声,从而达到对正样本增强的目的。 该框架的训练目的为:对于batch中的每个样本,拉近其与正样本之间的距离,拉远其与负样本之间的距离,使得模型能够在大规模无监督语料(也可以使用有监督的语料)中学习到文本相似关系。 详见论文:Simple Contrastive Learning of Sentence EmbeddingsSimCSE官方代码仓库

本项目使用pytorch+transformers复现了SimCSE论文中的有监督训练和无监督训练方法,并且在STS-B数据集上进行消融实验,评价指标为Spearman相关系数,预训练模型为Bert-base-uncased, 验证了SimCSE的有效性。在STS-B数据集上,有监督训练和无监督训练的复现效果如下表。

在无监督训练中,dropout=0.1时,复现效果比原文略差,但也比较接近。当dropout=0.2时,复现效果比原文略高。 ** 但在有监督训练中,不知是否由于batch size过小(原论文使用512),复现效果与论文的效果相差较远,后续会进行排查。 **

训练方法 learning rate batch size dropout Spearman’s correlation
原论文 无监督 3e-5 64 0.1 0.763
复现 无监督 3e-5 64 0.2 0.771
复现 无监督 3e-5 64 0.1 0.748
原论文 有监督 5e-5 512 0.1 0.816
复现 有监督 5e-5 64 0.1 0.764

运行环境

python==3.6、transformers==3.1.0、torch==1.6.0

项目结构

  • data:存放训练数据
    • stsbenchmark:STS-B数据集
      • sts-dev.csv:STS-B验证集
      • sts-test.csv:STS-B验测试集
    • nli_for_simcse.csv:数量275601为的NLI数据集
    • wiki1m_for_simcse.txt:维基百科上获取的100w的文本
  • output:输出目录
  • pretrain_model:预训练模型存放位置
  • script:脚本存放位置。
  • dataset.py
  • model.py:模型代码,包含有监督和无监督损失函数的计算方式
  • train.py:训练代码

使用方法

Quick Start

下载训练数据:

bash script/download_nli.sh
bash script/download_wiki.sh

无监督训练,运行脚本

bash script/run_unsup_train.sh

有监督训练,运行脚本

bash script/run_sup_train.sh

实验

无监督训练

从前四条实验数据中可以看到,较大的batch size在一定程度上可以增加模型的泛化性。

dropout为0.2的时候,训练效果比0.1与0.3更好,有可能dropout=0.1加入的噪声过小,而dropout=0.3加入的噪声过大,增强得到的样本与原始样本差异较大。

learning rate batch size dropout 在哪一步得到best checkpoint 验证集上的得分 测试集上的得分
3e-5 256 0.1 6000 0.800 0.761
3e-5 128 0.1 4200 0.799 0.747
3e-5 64 0.1 10900 0.803 0.748
3e-5 32 0.1 21300 0.787 0.714
3e-5 64 0.2 11200 0.811 0.771
3e-5 64 0.3 6300 0.781 0.745
1e-5 64 0.1 16400 0.798 0.751

有监督训练

有监督实验的复现结果未达到预期,超参数相同时,在验证集上的得分略高于无监督,但是在测试集上,得分基本没有差异。增大有监督训练的学习率,有监督的训练的得分略高于无监督训练, 但还是与论文声称的0.816相差较远,原论文使用512的batch size, 不知是否由于batch size的设置有关,后续会对有监督的训练代码进一步排查。

不过从训练曲线可以看到,有监督训练的收敛速度明显快于无监督训练,这也符合我们的认知。

训练方法 learning rate batch size dropout 在哪一步得到best checkpoint 验证集上的得分 测试集上的得分
无监督 3e-5 64 0.1 10900 0.803 0.748
有监督 3e-5 64 0.1 200 0.810 0.748
有监督 5e-5 64 0.1 2300 0.809 0.764
有监督 3e-5 32 0.1 200 0.808 0.743
有监督 5e-5 32 0.1 200 0.806 0.746

无监督训练过程中,验证集得分的变化曲线: avatar

有监督训练过程中,验证集得分的变化曲线: avatar

REFERENCE

TODO

  • 排查有监督学习的效果不符合预期的原因
PatrickStar enables Larger, Faster, Greener Pretrained Models for NLP. Democratize AI for everyone.

PatrickStar enables Larger, Faster, Greener Pretrained Models for NLP. Democratize AI for everyone.

Tencent 633 Dec 28, 2022
precise iris segmentation

PI-DECODER Introduction PI-DECODER, a decoder structure designed for Precise Iris Segmentation and Location. The decoder structure is shown below: Ple

8 Aug 08, 2022
Tool to check whether a GCP bucket is public or not.

Tool to check publicly accessible GCP bucket. Blog https://justm0rph3u5.medium.com/gcp-inspector-auditing-publicly-exposed-gcp-bucket-ac6cad55618c Wha

DIVYANSHU SHUKLA 7 Nov 24, 2022
Hierarchical unsupervised and semi-supervised topic models for sparse count data with CorEx

Anchored CorEx: Hierarchical Topic Modeling with Minimal Domain Knowledge Correlation Explanation (CorEx) is a topic model that yields rich topics tha

Greg Ver Steeg 592 Dec 18, 2022
Persian Bert For Long-Range Sequences

ParsBigBird: Persian Bert For Long-Range Sequences The Bert and ParsBert algorithms can handle texts with token lengths of up to 512, however, many ta

Sajjad Ayoubi 63 Dec 14, 2022
Rich Prosody Diversity Modelling with Phone-level Mixture Density Network

Phone Level Mixture Density Network for TTS This repo contains pytorch implementation of paper Rich Prosody Diversity Modelling with Phone-level Mixtu

Rishikesh (ऋषिकेश) 42 Dec 13, 2022
Unsupervised text tokenizer for Neural Network-based text generation.

SentencePiece SentencePiece is an unsupervised text tokenizer and detokenizer mainly for Neural Network-based text generation systems where the vocabu

Google 6.4k Jan 01, 2023
Synthetic data for the people.

zpy: Synthetic data in Blender. Website • Install • Docs • Examples • CLI • Contribute • Licence Abstract Collecting, labeling, and cleaning data for

Zumo Labs 253 Dec 21, 2022
Open-Source Toolkit for End-to-End Speech Recognition leveraging PyTorch-Lightning and Hydra.

🤗 Contributing to OpenSpeech 🤗 OpenSpeech provides reference implementations of various ASR modeling papers and three languages recipe to perform ta

Openspeech TEAM 513 Jan 03, 2023
Py65 65816 - Add support for the 65C816 to py65

Add support for the 65C816 to py65 Py65 (https://github.com/mnaberez/py65) is a

4 Jan 04, 2023
End-to-end MLOps pipeline of a BERT model for emotion classification.

image source EmoBERT-MLOps The goal of this repository is to build an end-to-end MLOps pipeline based on the MLOps course from Made with ML, but this

Dimitre Oliveira 4 Nov 06, 2022
Code examples for my Write Better Python Code series on YouTube.

Write Better Python Code This repository contains the code examples used in my Write Better Python Code series published on YouTube: https:/

858 Dec 29, 2022
Super easy library for BERT based NLP models

Fast-Bert New - Learning Rate Finder for Text Classification Training (borrowed with thanks from https://github.com/davidtvs/pytorch-lr-finder) Suppor

Utterworks 1.8k Dec 27, 2022
ChatterBot is a machine learning, conversational dialog engine for creating chat bots

ChatterBot ChatterBot is a machine-learning based conversational dialog engine build in Python which makes it possible to generate responses based on

Gunther Cox 12.8k Jan 03, 2023
Lightweight utility tools for the detection of multiple spellings, meanings, and language-specific terminology in British and American English

Breame ( British English and American English) Breame is a lightweight Python package with a number of utility tools to aid in the detection of words

Charles 8 Oct 10, 2022
CodeBERT: A Pre-Trained Model for Programming and Natural Languages.

CodeBERT This repo provides the code for reproducing the experiments in CodeBERT: A Pre-Trained Model for Programming and Natural Languages. CodeBERT

Microsoft 1k Jan 03, 2023
👄 The most accurate natural language detection library for Python, suitable for long and short text alike

1. What does this library do? Its task is simple: It tells you which language some provided textual data is written in. This is very useful as a prepr

Peter M. Stahl 334 Dec 30, 2022
Model parallel transformers in JAX and Haiku

Table of contents Mesh Transformer JAX Updates Pretrained Models GPT-J-6B Links Acknowledgments License Model Details Zero-Shot Evaluations Architectu

Ben Wang 4.9k Jan 04, 2023
CPC-big and k-means clustering for zero-resource speech processing

The CPC-big model and k-means checkpoints used in Analyzing Speaker Information in Self-Supervised Models to Improve Zero-Resource Speech Processing.

Benjamin van Niekerk 5 Nov 23, 2022
⚡ boost inference speed of T5 models by 5x & reduce the model size by 3x using fastT5.

Reduce T5 model size by 3X and increase the inference speed up to 5X. Install Usage Details Functionalities Benchmarks Onnx model Quantized onnx model

Kiran R 399 Jan 05, 2023