GitHub - facebookresearch/quip: Official repository for the paper "Question Answering Infused Pre-training of General-Purpose Contextualized Representations" by Robin Jia, Mike Lewis, and Luke Zettlemoyer. (original) (raw)

QuIP: Question Answering Infused Pre-training

This is the official GitHub repository for the following paper:

Question Answering Infused Pre-training of General-Purpose Contextualized Representations.
Robin Jia, Mike Lewis, and Luke Zettlemoyer.
arXiv, 2021.

Here, you can download our pre-trained QuIP model and reproduce our paper's main results on question answering, paraphrase detection, named entity recognition, and sentiment analysis.

Setup

This code has been tested with python 3.7.9. Key installed packages were:

• fairseq 1.0.0a0+9825786 (most recent version as of August 9, 2020)
• scikit-learn 0.24.0
• torch 1.7.1
• transformers 4.4.2

Note that you should install fairseq by following the instructions in the fairseq GitHub repository, not directly through pip.

A full list of installed packages is provided in requirements.txt (sufficient but likely not necessary for the purposes of this repository).

Getting the data

We have a single command that will download all relevant data for you:

This does (in order) the following:

• GLUE (for QQP, MRPC, SST-2): Gets the standard GLUE dataset files and puts them at data/glue_data.
• LM-BFF (for QQP, MRPC, MR, CR): Clones the LM-BFF repository to data/lm-bff and follows their instructions to download the data.
• SQuAD: Downloads the train-v1.1.json, dev-v1.1.json, and evaluate-v2.0.py files from the SQuAD website repo and puts them in the directory data/squad.
• MRQA: Clones the MRQA shared task repository to data/mrqa, follows their instructions to download the development data, and gunzips the files.
• PAWS: Clones the PAWS repository to data/paws and follows their instructions to download the data (Note: This requires python2.7 to exist and have nltk installed). Then, we re-generate our k-shot PAWS splits using the list of IDs in paws_kshot_splits and write them to data/lm-bff, where the other k-shot splits are stored.

Note that we used the few-shot NER splits from [Few-Shot Named Entity Recognition: A Comprehensive Study]https://arxiv.org/abs/2012.14978) by Huang et al. These are not yet publicly available but will be released by Huang et al. at a later date. That data should be placed in data/ner/few_shot.

Getting the model

We have a single command that will download the main model for you:

This downloads both a fairseq-compatible and Huggingface-compatible version of the model, and places them at models/quip and models/quip-hf, respectively.

The main QuIP model in fairseq-compatible format can be downloaded at https://dl.fbaipublicfiles.com/quip/quip.tar.gz. Un-tar the file and you will see 4 files:

• model.pt: The main Transformer model. This has the same architecture as RoBERTa-large, and can be loaded by fairseq exactly like a RoBERTa checkpoint.
• dict.txt: Same dict.txt file used by RoBERTa.
• model_qa_head_start.pt: The parameters of the head used by the bi-encoder QA model to predict the start of the answer. This can be loaded by the MockClassificationHead class found in src/mock_classification_head.py.
• model_qa_head_end.pt Same as above, except for predicting the end of the answer.

The same model in HuggingFace-compatible format can be downloaded at https://dl.fbaipublicfiles.com/quip/quip-hf.tar.gz. Un-tar the file and you will get a directory called quip-hf that contains a checkpoint that can be directly loaded with HuggingFace's code for RoBERTa large models. Note that this checkpoint does not include the QA start and end heads, but you can load those directly from the files in models/quip.

We have also made available the following other models:

• quip-noTeacher and quip-noTeacher-hf: QuIP trained without the teacher model, for comparison.
• quip-concat and quip-concat-hf: QuIP trained by concatenating passage representations in the same batch together, simulating negative examples. We hypothesize that this may be a useful starting point for applications that require handling of negative examples, such as open-domain QA.

These can be downloaded by replacing the second "quip" in the original URL with the name of the model.

Running experiments

Question answering

python src/biencoder_predict_qa.py squad predictions_squad.json models/quip
python src/run_sentiment.py mrqa predictions_mrqa.json models/quip

The predictions will be written to predictions_squad.json and predictions_mrqa.json.

Zero-shot sentiment analysis

python src/run_sentiment.py [sst2|mr|cr] qa models/quip

Zero-shot paraphrase ranking

python src/run_sentiment.py [qqp|mrpc|paws-qqp|paws-wiki] models/quip

Few-shot paraphrase classification

With frozen model:

python src/run_sentiment.py [qqp|mrpc|paws-qqp|paws-wiki] models/quip --train-ft

With full fine-tuning:

python src/run_sentiment.py [qqp|mrpc|paws-qqp|paws-wiki] models/quip --train-skl

Note: We are able to exactly reproduce our paper's numbers when using a V100 GPU, but observe slight differences on a GP100.

Few-shot named entity recognition

To initialize the output head with question embeddings and train 5 times with the 5 different random splits, evaluating on the test set:

bash scripts/sweep_few_shot_ner.sh conll2003 out_dir models/quip-hf --learning_rate 2e-5 --num_train_epochs 200 --use_head_initialization

This saves model checkpoints and results to out_dir. The command run on conll2003 should print the following:

0.7544890082228561 0.72753114288103 0.7445901913275964 0.6633886828953732 0.8107488986784142

which are the test F1 scores for each of the 5 seeds.

To randomly initialize the output head, just omit the --use_head_initialization flag.

License

The majority of "QuIP: Question Answering Infused Pretraining" is licensed under CC-BY-NC, however portions of the project are available under separate license terms: HuggingFace is licensed under the Apache 2.0 license.