Bert Mini Finetuned Squadv2

🚀 BERT-mini model finetuned with M-FAC

This model is finetuned on the SQuAD version 2 dataset using the state-of-the-art second-order optimizer M-FAC, offering enhanced performance for question-answering tasks.

🚀 Quick Start

This model is finetuned on SQuAD version 2 dataset with state-of-the-art second-order optimizer M-FAC. Check NeurIPS 2021 paper for more details on M-FAC: https://arxiv.org/pdf/2107.03356.pdf.

✨ Features

• Finetuned on SQuAD version 2 dataset with M-FAC optimizer.
• Provides competitive results on SQuAD version 2 validation set.
• Code and tutorial for M-FAC integration are available.

📦 Installation

No specific installation steps are provided in the original README. However, you can follow the finetuning setup to use the model.

💻 Usage Examples

Basic Usage

To reproduce the results, add M-FAC optimizer code in https://github.com/huggingface/transformers/blob/master/examples/pytorch/question-answering/run_qa.py and run the following bash script:

CUDA_VISIBLE_DEVICES=0 python run_qa.py \
    --seed 8276 \
    --model_name_or_path prajjwal1/bert-mini \
    --dataset_name squad_v2 \
    --version_2_with_negative \
    --do_train \
    --do_eval \
    --per_device_train_batch_size 12 \
    --learning_rate 1e-4 \
    --num_train_epochs 2 \
    --max_seq_length 384 \
    --doc_stride 128 \
    --output_dir out_dir/ \
    --optim MFAC \
    --optim_args '{"lr": 1e-4, "num_grads": 1024, "damp": 1e-6}'

📚 Documentation

Finetuning setup

For fair comparison against default Adam baseline, we finetune the model in the same framework as described here https://github.com/huggingface/transformers/tree/master/examples/pytorch/question-answering and just swap Adam optimizer with M-FAC. Hyperparameters used by M-FAC optimizer:

learning rate = 1e-4
number of gradients = 1024
dampening = 1e-6

Results

We share the best model out of 5 runs with the following score on SQuAD version 2 validation set:

exact_match = 58.38
f1 = 61.65

Mean and standard deviation for 5 runs on SQuAD version 2 validation set:

We believe these results could be improved with modest tuning of hyperparameters: per_device_train_batch_size, learning_rate, num_train_epochs, num_grads and damp. For the sake of fair comparison and a robust default setup we use the same hyperparameters across all models (bert-tiny, bert-mini) and all datasets (SQuAD version 2 and GLUE).

Our code for M-FAC can be found here: https://github.com/IST-DASLab/M-FAC. A step-by-step tutorial on how to integrate and use M-FAC with any repository can be found here: https://github.com/IST-DASLab/M-FAC/tree/master/tutorials.

📄 License

No license information is provided in the original README.

🔧 Technical Details

The model is finetuned using the M-FAC optimizer, which is a state-of-the-art second-order optimizer. For more details on M-FAC, check the NeurIPS 2021 paper: https://arxiv.org/pdf/2107.03356.pdf.

📄 BibTeX entry and citation info

@article{frantar2021m,
  title={M-FAC: Efficient Matrix-Free Approximations of Second-Order Information},
  author={Frantar, Elias and Kurtic, Eldar and Alistarh, Dan},
  journal={Advances in Neural Information Processing Systems},
  volume={35},
  year={2021}
}