Bert Tiny Finetuned Stsb

🚀 BERT-tiny model finetuned with M-FAC

This model is a fine-tuned BERT-tiny model using the state-of-the-art second-order optimizer M-FAC on the STS-B dataset, offering enhanced performance for specific tasks.

🚀 Quick Start

This model is finetuned on the STS-B dataset with the state-of-the-art second-order optimizer M-FAC. For more details on M-FAC, check the NeurIPS 2021 paper: https://arxiv.org/pdf/2107.03356.pdf.

✨ Features

• Advanced Optimizer: Utilizes the M-FAC optimizer, a state-of-the-art second-order optimizer, for finetuning.
• Fair Comparison Setup: Ensures a fair comparison against the default Adam baseline by using the same framework.

📦 Installation

No specific installation steps are provided in the original README, so this section is skipped.

💻 Usage Examples

Basic Usage

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

The hyperparameters used by the M-FAC optimizer are as follows:

learning rate = 1e-4
number of gradients = 512
dampening = 1e-6

Advanced Usage

The results can be reproduced by adding the M-FAC optimizer code in https://github.com/huggingface/transformers/blob/master/examples/pytorch/text-classification/run_glue.py and running the following bash script:

CUDA_VISIBLE_DEVICES=0 python run_glue.py \
  --seed 7 \
  --model_name_or_path prajjwal1/bert-tiny \
  --task_name stsb \
  --do_train \
  --do_eval \
  --max_seq_length 128 \
  --per_device_train_batch_size 32 \
  --learning_rate 1e-4 \
  --num_train_epochs 5 \
  --output_dir out_dir/ \
  --optim MFAC \
  --optim_args '{"lr": 1e-4, "num_grads": 512, "damp": 1e-6}'

📚 Documentation

Results

We share the best model out of 5 runs with the following score on the STS-B validation set:

pearson = 80.66
spearman = 81.13

The mean and standard deviation for 5 runs on the STS-B validation set are as follows:

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, so this section is skipped.

📚 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}
}