The Codebase for Causal Distillation for Language Models.

Overview

Python 3.7 License CC BY-NC

Causal Distillation for Language Models

Zhengxuan Wu*,Atticus Geiger*, Josh Rozner, Elisa Kreiss, Hanson Lu, Thomas Icard, Christopher Potts, Noah D. Goodman

The is an implementation of our preprint Causal Distillation for Language Models. The standard approach to distillation trains a student model against two objectives: a task-specific objective (e.g., language modeling) and an imitation objective that encourages the hidden states of the student model to be similar to those of the larger teacher model. In this paper, we show that it is beneficial to augment distillation with a third objective that encourages the student to imitate the causal computation process of the teacher through interchange intervention training (IIT).

We fork our main codebase from the Huggingface Distillation Interface.

Release Notes

12/02/2021 Our paper on Interchange Intervention Training (IIT) is released! Read this more formal definition of the method.
12/06/2021 Released the causal distillation codebase with the preprint.
12/06/2021 Released evaluation results on distilled tiny-BERT (3 layers) with the Wiki-Text 103M dataset.
⬜️ Released evaluation results on causal-distilled tiny-BERT (3 layers) with the Wiki-Text 103M + BookCorpus dataset.
⬜️ Released evaluation results on causal-distilled BERT (6 layers) with the Wiki-Text 103M + BookCorpus dataset.
⬜️ Released more ablation studies.
⬜️ Released causal-distilled tiny-BERT (3 layers) model files.
⬜️ Released causal-distilled BERT (6 layers) model files.

If you experience any issues or have suggestions, please contact me either thourgh the issues page or at [email protected].

Benchmark Results

Here are the results on the dev sets of GLUE:

Model Average-score CoLA MNLI MRPC QNLI QQP RTE SST-2 STS-B WNLI
DistilBERT (3 layers) 67.81 22.8 71.6 78.2 82.1 84.3 55.4 86.5 56.7 24.2
CausalBERT (3 layers) 69.71 25.0 72.9 78.6 83.1 84.9 55.4 86.9 66.5 21.5

1 Average-score computed without WNLI.

Main Contents

Citation

If you use this repository, please cite the following two papers: paper for interchange intervention training, and paper for the our distillation method.

  @article{geiger-etal-2021-iit,
        title={Inducing Causal Structure for Interpretable Neural Networks}, 
        author={Geiger, Atticus and Wu, Zhengxuan and Lu, Hanson and Rozner, Josh and Kreiss, Elisa and Icard, Thomas and Goodman, Noah D. and Potts, Christopher},
        year={2021},
        eprint={2112.00826},
        archivePrefix={arXiv},
        primaryClass={cs.LG}
  }

  @article{wu-etal-2021-distill,
        title={Causal Distillation for Language Models}, 
        author={Wu, Zhengxuan and Geiger, Atticus and Rozner, Josh and Kreiss, Elisa and Lu, Hanson and Icard, Thomas and Potts, Christopher and Goodman, Noah D.},
        year={2021},
        eprint={2112.02505},
        archivePrefix={arXiv},
        primaryClass={cs.CL}
  }

Requirements

  • Python 3.6 or 3.7 are supported.
  • Pytorch Version: 1.9.0
  • Transfermers Version: 4.11.3
  • Datasets Version: Version: 1.8.0
  • We have performed experiments on Titan V GPU. We assume 12GB of GPU memory (more memory can expedite training).
  • Since we build our codebase off the Huggingface Distillation Interface, please review their doc for requirements.

Dataset

Following the Huggingface Distillation Interface, we need to pre-process the datasets before we do distillation. You can refer to their repo for details. We adapt their pre-processing scripts, and update with a few improvements. For example, we can now binarize datasets from the Dataset Hub from huggingface directly.

# preprocessing from disk
python script/binarized_data.py \
--file_path ../../bert-mid-tuning/data-files/wikitext-15M \
--split train \
--field_name text \
--max_parsing_example 1000 \
--tokenizer_type bert \
--tokenizer_name bert-base-uncased \
--dump_file ./data/binarized_text

# preprocessing from huggingface.
python scripts/binarized_data.py \
--dataset_name bookcorpus \
--split train \
--field_name text \
--tokenizer_type bert \
--tokenizer_name bert-base-uncased \
--dump_file bookcorpus-dataset/binarized_text \
--cache_dir ./distill_cache/

python scripts/binarized_data.py \
--dataset_name wikitext \
--split train \
--field_name text \
--tokenizer_type bert \
--tokenizer_name bert-base-uncased \
--dump_file wikitext-dataset/binarized_text \
--cache_dir ./distill_cache/

python scripts/binarized_data.py \
--dataset_name wikitext+bookcorpus \
--split train \
--field_name text \
--tokenizer_type bert \
--tokenizer_name bert-base-uncased \
--dump_file wikitext+bookcorpus-dataset/binarized_text \
--cache_dir ./distill_cache/

# helper scripts to combine two binarized data files
python scripts/data_combinator.py \
--file_path_left ./bookcorpus-dataset/binarized_text.train.bert-base-uncased.pickle \
--file_path_right ./wikitext-dataset/binarized_text.train.bert-base-uncased.pickle \
--split train \
--tokenizer_name bert-base-uncased \
--dump_file wikitext+bookcorpus-dataset/binarized_text

# multiprocessing preprocessor.
python scripts/binarized_data.py \
--dataset_name bookcorpus \
--split train \
--field_name text \
--tokenizer_type bert \
--tokenizer_name bert-base-uncased \
--dump_file bookcorpus-dataset/binarized_text \
--cache_dir ./distill_cache/ \
--fast_process \
--preprocessing_num_workers 48

After you get the datasets ready, you need to generate token counts as well.

python scripts/token_counts.py \
--data_file data/binarized_text.train.bert-base-uncased.pickle \
--token_counts_dump data/binarized_text.train.token_counts.bert-base-uncased.pickle \
--vocab_size 30522

Distillation

Before training, we recommand you to initialize your student model with weights extracted from the teacher model.

python scripts/extract_distilbert.py \
--model_type bert \
--model_name bert-base-uncased \
--dump_checkpoint ./distillation_checkpoints/bert-base-uncased_num_layer_3.pth \
--num_layers 3

Now, here is an example for you to distill with our causal distillation objective or without,

CUDA_VISIBLE_DEVICES=9,4 python causal_train.py \
--force \
--n_gpu 2 \
--is_wandb \
--log_interval 10 \
--student_type distilbert \
--student_config ./training_configs/distilbert-base-uncased-small.json \
--student_pretrained_weights ./distillation_checkpoints/bert-base-uncased_num_layer_3.pth \
--teacher_type bert \
--teacher_name bert-base-uncased \
--neuron_mapping ./training_configs/single_middle.nm \
--mlm --alpha_ce 0.25 --alpha_mlm 0.25 --alpha_cos 0.25 --alpha_clm 0.0 --alpha_causal 0.25 \
--freeze_pos_embs \
--dump_path ./results/ \
--data_file ./wikitext-15M/binarized_text.train.bert-base-uncased.pickle \
--token_counts ./wikitext-15M/binarized_text.train.token_counts.bert-base-uncased.pickle \
--seed 42 \
--gradient_accumulation_steps 50 \
--n_epoch 3 \
--batch_size 5

CUDA_VISIBLE_DEVICES=0,1,2,3 python causal_train.py \
--force \
--n_gpu 4 \
--is_wandb \
--log_interval 10 \
--student_type distilbert \
--student_config ./training_configs/distilbert-base-uncased-small.json \
--student_pretrained_weights ./distillation_checkpoints/bert-base-uncased_num_layer_3.pth \
--teacher_type bert \
--teacher_name bert-base-uncased \
--neuron_mapping ./training_configs/single_middle.nm \
--mlm --alpha_ce 0.33 --alpha_mlm 0.33 --alpha_cos 0.33 --alpha_clm 0.0 --alpha_causal 0.00 \
--freeze_pos_embs \
--dump_path ./results/ \
--data_file ./wikitext-15M/binarized_text.train.bert-base-uncased.pickle \
--token_counts ./wikitext-15M/binarized_text.train.token_counts.bert-base-uncased.pickle \
--seed 42 \
--gradient_accumulation_steps 124 \
--n_epoch 6 \
--batch_size 4

Note that you can simply turn our causal distillation objective on/off through setting the arguments.

Evaluation

After you get your distilled models, you need to fine-tune them and evaluate them with downstream tasks. We provide you all the scripts you need to run.

MLM Evaluation

CUDA_VISIBLE_DEVICES=5 python run_mlm.py \
--model_name_or_path ./results/s_distilbert_t_bert_data_wikitext-15M_seed_42_mlm_True_ce_0.25_mlm_0.25_cos_0.25_causal_0.25_nm_single_multilayer/ \
--dataset_dir ../../bert-mid-tuning/data-files/wikitext-15M/ \
--tokenizer_name bert-base-uncased \
--do_eval \
--output_dir /tmp/test-mlm \
--cache_dir ./distill_cache/

GLUE Evaluation

CUDA_VISIBLE_DEVICES=5,7,8,9 python run_glue.py \
--model_name_or_path ./results/s_distilbert_t_bert_data_wikitext-dataset_seed_42_mlm_True_ce_0.33_mlm_0.33_cos_0.33_causal_0.0_nm_single_middle/ \
--tokenizer_name bert-base-uncased \
--task_name sst2 \
--do_train \
--do_eval \
--max_seq_length 128 \
--per_device_train_batch_size 32 \
--learning_rate 2e-5 \
--num_train_epochs 3 \
--output_dir ./results/ \
--save_total_limit 1 \
--cache_dir ./distill_cache/

CoNLL Evaluation

CUDA_VISIBLE_DEVICES=2,3,7,8 python run_ner.py \
--model_name_or_path ./results/s_distilbert_t_bert_data_wikitext-dataset_seed_42_mlm_True_ce_0.33_mlm_0.33_cos_0.33_causal_0.0_nm_single_middle_crossway_False/ \
--tokenizer_name bert-base-uncased \
--dataset_name conll2003 \
--do_train \
--do_eval \
--output_dir ./ner_results/ \
--save_total_limit 1 \
--cache_dir ./distill_cache/

SQuAD Evaluation

CUDA_VISIBLE_DEVICES=2,3,7,8 python run_qa.py \
--model_name_or_path ./results/s_distilbert_t_bert_data_wikitext-dataset_seed_42_mlm_True_ce_0.33_mlm_0.33_cos_0.33_causal_0.0_nm_single_middle_crossway_False/ \
--tokenizer_name bert-base-uncased \
--dataset_name squad \
--do_train \
--do_eval \
--per_device_train_batch_size 12 \
--learning_rate 3e-5 \
--num_train_epochs 2 \
--max_seq_length 384 \
--doc_stride 128 \
--save_total_limit 1 \
--output_dir ./qa_results/
Distributing reference energies for SMIRNOFF implementations

Warning: This code is currently experimental and under active development. Is it not yet suitable for distribution or use as reference implementation.

Open Force Field Initiative 1 Dec 07, 2021
Optimal Camera Position for a Practical Application of Gaze Estimation on Edge Devices,

Optimal Camera Position for a Practical Application of Gaze Estimation on Edge Devices, Linh Van Ma, Tin Trung Tran, Moongu Jeon, ICAIIC 2022 (The 4th

Linh 11 Oct 10, 2022
Official pytorch implementation of the AAAI 2021 paper Semantic Grouping Network for Video Captioning

Semantic Grouping Network for Video Captioning Hobin Ryu, Sunghun Kang, Haeyong Kang, and Chang D. Yoo. AAAI 2021. [arxiv] Environment Ubuntu 16.04 CU

Hobin Ryu 43 Nov 25, 2022
Train emoji embeddings based on emoji descriptions.

emoji2vec This is my attempt to train, visualize and evaluate emoji embeddings as presented by Ben Eisner, Tim Rocktäschel, Isabelle Augenstein, Matko

Miruna Pislar 17 Sep 03, 2022
Video-Music Transformer

VMT Video-Music Transformer (VMT) is an attention-based multi-modal model, which generates piano music for a given video. Paper https://arxiv.org/abs/

Chin-Tung Lin 5 Jul 13, 2022
A sequence of Jupyter notebooks featuring the 12 Steps to Navier-Stokes

CFD Python Please cite as: Barba, Lorena A., and Forsyth, Gilbert F. (2018). CFD Python: the 12 steps to Navier-Stokes equations. Journal of Open Sour

Barba group 2.6k Dec 30, 2022
CLUES: Few-Shot Learning Evaluation in Natural Language Understanding

CLUES: Few-Shot Learning Evaluation in Natural Language Understanding This repo contains the data and source code for baseline models in the NeurIPS 2

Microsoft 29 Dec 29, 2022
Good Classification Measures and How to Find Them

Good Classification Measures and How to Find Them This repository contains supplementary materials for the paper "Good Classification Measures and How

Yandex Research 7 Nov 13, 2022
Systematic generalisation with group invariant predictions

Requirements are Python 3, TensorFlow v1.14, Numpy, Scipy, Scikit-Learn, Matplotlib, Pillow, Scikit-Image, h5py, tqdm. Experiments were run on V100 GPUs (16 and 32GB).

Faruk Ahmed 30 Dec 01, 2022
The DL Streamer Pipeline Zoo is a catalog of optimized media and media analytics pipelines.

The DL Streamer Pipeline Zoo is a catalog of optimized media and media analytics pipelines. It includes tools for downloading pipelines and their dependencies and tools for measuring their performace

8 Dec 04, 2022
Pytorch implementation of PTNet for high-resolution and longitudinal infant MRI synthesis

Pyramid Transformer Net (PTNet) Project | Paper Pytorch implementation of PTNet for high-resolution and longitudinal infant MRI synthesis. PTNet: A Hi

Xuzhe Johnny Zhang 6 Jun 08, 2022
[NeurIPS 2021] Code for Unsupervised Learning of Compositional Energy Concepts

Unsupervised Learning of Compositional Energy Concepts This is the pytorch code for the paper Unsupervised Learning of Compositional Energy Concepts.

45 Nov 30, 2022
Repository for "Space-Time Correspondence as a Contrastive Random Walk" (NeurIPS 2020)

Space-Time Correspondence as a Contrastive Random Walk This is the repository for Space-Time Correspondence as a Contrastive Random Walk, published at

A. Jabri 239 Dec 27, 2022
Codes of paper "Unseen Object Amodal Instance Segmentation via Hierarchical Occlusion Modeling"

Unseen Object Amodal Instance Segmentation (UOAIS) Seunghyeok Back, Joosoon Lee, Taewon Kim, Sangjun Noh, Raeyoung Kang, Seongho Bak, Kyoobin Lee This

GIST-AILAB 92 Dec 13, 2022
Image Captioning on google cloud platform based on iot

Image-Captioning-on-google-cloud-platform-based-on-iot - Image Captioning on google cloud platform based on iot

Shweta_kumawat 1 Jan 20, 2022
Official Implementation of Neural Splines

Neural Splines: Fitting 3D Surfaces with Inifinitely-Wide Neural Networks This repository contains the official implementation of the CVPR 2021 (Oral)

Francis Williams 56 Nov 29, 2022
Reference implementation of code generation projects from Facebook AI Research. General toolkit to apply machine learning to code, from dataset creation to model training and evaluation. Comes with pretrained models.

This repository is a toolkit to do machine learning for programming languages. It implements tokenization, dataset preprocessing, model training and m

Facebook Research 408 Jan 01, 2023
This repository collects 100 papers related to negative sampling methods.

Negative-Sampling-Paper This repository collects 100 papers related to negative sampling methods, covering multiple research fields such as Recommenda

RUCAIBox 119 Dec 29, 2022
Learn other languages ​​using artificial intelligence with python.

The main idea of ​​the project is to facilitate the learning of other languages. We created a simple AI that will interact with you. Just ask questions that if she knows, she will answer.

Pedro Rodrigues 2 Jun 07, 2022
Official implementation of "Learning to Discover Cross-Domain Relations with Generative Adversarial Networks"

DiscoGAN Official PyTorch implementation of Learning to Discover Cross-Domain Relations with Generative Adversarial Networks. Prerequisites Python 2.7

SK T-Brain 754 Dec 29, 2022