Sparse Progressive Distillation: Resolving Overfitting under Pretrain-and-Finetune Paradigm

Last update: Dec 05, 2022

Overview

Sparse Progressive Distillation: Resolving Overfitting under Pretrain-and-Finetune Paradigm

This is the Pytorch implementation for sparse progressive distillation (SPD). For more details about the motivation, techniques and experimental results, refer to our paper here.

Running

Environment Preparation (using python3)
```
pip install -r requirements.txt
```
Dataset Preparation

The original GLUE dataset could be downloaded here.

BERT_base fine-tuning on GLUE

We use finetuned BERT_base as the teacher. For each task of GLUE benchmark, we obtain the finetuned model using the original huggingface transformers code with the following script.

python run_glue.py \
          --model_name_or_path $INT_DIR \
          --task_name $TASK_NAME \
          --do_train \
          --do_eval \
          --data_dir $GLUE_DIR/$TASK_NAME/ \
          --max_seq_length 128 \
          --per_gpu_train_batch_size 32 \
          --per_gpu_eval_batch_size 32 \
          --learning_rate 3e-5 \
          --num_train_epochs 4.0 \
          --output_dir $OUT_DIR \
          --evaluate_during_training \
          --overwrite_output_dir \
          --logging_steps 400 \
          --logging_dir $OUT_DIR \
          --save_steps 10000

Sparse Progressive Distillation

We use run_glue.py to run the sparse progressive distillation. --num_prune_epochs is the epochs for pruning. --num_train_epochs is the total number of epochs (pruning, progressive distillation, finetuning).

python run_glue.py \
  --model_name_or_path PATH_TO_FINETUNED_MODEL \
  --task_name $TASK_NAME \
  --do_train \
  --do_eval \
  --do_lower_case \
  --data_dir $GLUE_DIR/$TASK_NAME/ \
  --max_seq_length 128 \
  --per_gpu_train_batch_size 32 \
  --per_gpu_eval_batch_size 32 \
  --learning_rate 6.4e-4 \
  --save_steps 50 \
  --num_prune_epochs 30 \
  --num_train_epochs 60 \
  --sparsity 0.9 \
  --output_dir $OUT_DIR \
  --evaluate_during_training \
  --replacing_rate 0.8 \
  --overwrite_output_dir \
  --steps_for_replacing 0 \
  --scheduler_type linear

To Dos

Provide our teacher model for each task.
Provide best performed model checkpoint for each task.

Sparse Progressive Distillation: Resolving Overfitting under Pretrain-and-Finetune Paradigm

Related tags

Overview

Sparse Progressive Distillation: Resolving Overfitting under Pretrain-and-Finetune Paradigm

Running

BERT_base fine-tuning on GLUE

Sparse Progressive Distillation

To Dos

Owner

Making Structure-from-Motion (COLMAP) more robust to symmetries and duplicated structures

[NAACL & ACL 2021] SapBERT: Self-alignment pretraining for BERT.

Official PyTorch code for Mutual Affine Network for Spatially Variant Kernel Estimation in Blind Image Super-Resolution (MANet, ICCV2021)

Implementation supporting the ICCV 2017 paper "GANs for Biological Image Synthesis"

PyTorch implementation of Advantage Actor Critic (A2C), Proximal Policy Optimization (PPO), Scalable trust-region method for deep reinforcement learning using Kronecker-factored approximation (ACKTR) and Generative Adversarial Imitation Learning (GAIL).

Fbone (Flask bone) is a Flask (Python microframework) starter/template/bootstrap/boilerplate application.

RNN Predict Street Commercial Vitality

This is an example of a reproducible modelling project

QAHOI: Query-Based Anchors for Human-Object Interaction Detection (paper)

PyTorch/GPU re-implementation of the paper Masked Autoencoders Are Scalable Vision Learners

Personalized Federated Learning using Pytorch (pFedMe)

Gym Threat Defense

Code for the Active Speakers in Context Paper (CVPR2020)

SMD-Nets: Stereo Mixture Density Networks

Localization Distillation for Object Detection

Official implementation of the article "Unsupervised JPEG Domain Adaptation For Practical Digital Forensics"

EMNLP'2021: Simple Entity-centric Questions Challenge Dense Retrievers

Public repository containing materials used for Feed Forward (FF) Neural Networks article.

DeepVoxels is an object-specific, persistent 3D feature embedding.

Code for the paper Relation Prediction as an Auxiliary Training Objective for Improving Multi-Relational Graph Representations (AKBC 2021).