| <span align="center"> | |
| <a href="https://huggingface.co/SajjadAyoubi/"><img src="https://img.shields.io/static/v1?label=%F0%9F%A4%97%20Hugging%20Face&message=SajjadAyoubi&color=yellow"></a> | |
| <a href="https://colab.research.google.com/github/sajjjadayobi/PersianQA/blob/main/notebooks/Demo.ipynb"><img src="https://img.shields.io/static/v1?label=Colab&message=Fine-tuning Example&logo=Google%20Colab&color=f9ab00"></a> | |
| </span> | |
| # ParsBigBird: Persian Bert For **Long-Range** Sequences | |
| The [Bert](https://arxiv.org/abs/1810.04805) and [ParsBert](https://arxiv.org/abs/2005.12515) algorithms can handle texts with token lengths of up to 512, however, many tasks such as summarizing and answering questions require longer texts. In our work, we have trained the [BigBird](https://arxiv.org/abs/2007.14062) model for the Persian language to process texts up to 4096 in the Farsi (Persian) language using sparse attention. | |
| ## Evaluation: 🌡️ | |
| We have evaluated the model on three tasks with different sequence lengths | |
| | Name | Params | SnappFood (F1) | Digikala Magazine(F1) | PersianQA (F1) | | |
| | :--------------------------------------------------------------: | :----: | :-----------------: | :---------------: | :--------------: | | |
| | [distil-bigbird-fa-zwnj](https://github.com/sajjjadayobi/ParsBigBird) | 78M | 85.43% | **94.05%** | **73.34%** | | |
| | [bert-base-fa](https://github.com/hooshvare/parsbert) | 118M | **87.98%** | 93.65% | 70.06% | | |
| - Despite being as big as distill-bert, the model performs equally well as ParsBert and is much better on PersianQA which requires much more context | |
| - This evaluation was based on `max_lentgh=2048` (It can be changed up to 4096) | |
| ## How to use❓ | |
| ### As Contextualized Word Embedding | |
| ```python | |
| from transformers import BigBirdModel, AutoTokenizer | |
| MODEL_NAME = "SajjadAyoubi/distil-bigbird-fa-zwnj" | |
| # by default its in `block_sparse` block_size=32 | |
| model = BigBirdModel.from_pretrained(MODEL_NAME, block_size=32) | |
| # you can use full attention like the following: use this when input isn't longer than 512 | |
| model = BigBirdModel.from_pretrained(MODEL_NAME, attention_type="original_full") | |
| text = "😃 امیدوارم مدل بدردبخوری باشه چون خیلی طول کشید تا ترین بشه" | |
| tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME) | |
| tokens = tokenizer(text, return_tensors='pt') | |
| output = model(**tokens) # contextualized embedding | |
| ``` | |
| ### As Fill Blank | |
| ```python | |
| from transformers import pipeline | |
| MODEL_NAME = 'SajjadAyoubi/distil-bigbird-fa-zwnj' | |
| fill = pipeline('fill-mask', model=MODEL_NAME, tokenizer=MODEL_NAME) | |
| results = fill('تهران پایتخت [MASK] است.') | |
| print(results[0]['token_str']) | |
| >>> 'ایران' | |
| ``` | |
| ## Pretraining details: 🔭 | |
| This model was pretrained using a masked language model (MLM) objective on the Persian section of the Oscar dataset. Following the original BERT training, 15% of tokens were masked. This was first described in this [paper](https://arxiv.org/abs/2007.14062) and released in this [repository](https://github.com/google-research/bigbird). Documents longer than 4096 were split into multiple documents, while documents much smaller than 4096 were merged using the [SEP] token. Model is warm started from `distilbert-fa`’s [checkpoint](https://huggingface.co/HooshvareLab/distilbert-fa-zwnj-base). | |
| - For more details, you can take a look at config.json at the model card in 🤗 Model Hub | |
| ## Fine Tuning Recommendations: 🐤 | |
| Due to the model's memory requirements, `gradient_checkpointing` and `gradient_accumulation` should be used to maintain a reasonable batch size. Considering this model isn't really big, it's a good idea to first fine-tune it on your dataset using Masked LM objective (also called intermediate fine-tuning) before implementing the main task. In block_sparse mode, it doesn't matter how many tokens are input. It just attends to 256 tokens. Furthermore, original_full should be used up to 512 sequence lengths (instead of block sparse). | |
| ### Fine Tuning Examples 👷♂️👷♀️ | |
| | Dataset | Fine Tuning Example | | |
| | ------------------------------------- | ------------------------------------------------------------ | | |
| | Digikala Magazine Text Classification | <a href="https://colab.research.google.com/github/sajjjadayobi/PersianQA/blob/main/notebooks/Demo.ipynb"><img src="https://img.shields.io/static/v1?label=Colab&message=Fine-tuning Example&logo=Google%20Colab&color=f9ab00"></a> | | |
| ## Contact us: 🤝 | |
| If you have a technical question regarding the model, pretraining, code or publication, please create an issue in the repository. This is the fastest way to reach us. | |
| ## Citation: ↩️ | |
| we didn't publish any papers on the work. However, if you did, please cite us properly with an entry like one below. | |
| ```bibtex | |
| @misc{ParsBigBird, | |
| author = {Ayoubi, Sajjad}, | |
| title = {ParsBigBird: Persian Bert For Long-Range Sequences}, | |
| year = 2021, | |
| publisher = {GitHub}, | |
| journal = {GitHub repository}, | |
| howpublished = {\url{https://github.com/SajjjadAyobi/ParsBigBird}}, | |
| } | |
| ``` | |