Create README.md

README.md

@@ -0,0 +1,57 @@
+---
+license: apache-2.0
+---
+
+# Model description
+
+This is an [mt5-base](https://huggingface.co/google/mt5-base) model, finetuned to generate questions using [TyDi QA](https://huggingface.co/datasets/tydiqa) dataset. It was trained to take the context and answer as input to generate questions.
+
+# Overview
+**Language model**: mT5-base
+
+**Language**: Arabic, Bengali, English, Finnish, Indonesian, Korean, Russian, Swahili, Telugu
+
+**Task**: Question Generation 
+
+**Data**: TyDi QA
+
+# Intented use and limitations
+One can use this model to generate questions. Biases associated with pre-training of mT5 and TyDiQA dataset may be present.
+
+## Usage
+One can use this model directly in the [PrimeQA](https://github.com/primeqa/primeqa) framework as this example [notebook](https://github.com/primeqa/primeqa/blob/tableqg/notebooks/qg/tableqg_inference.ipynb).
+
+Or 
+```python
+from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
+tokenizer = AutoTokenizer.from_pretrained("ibm/mt5-base-tydi-question-generator")
+model = AutoModelForSeq2SeqLM.from_pretrained("ibm/mt5-base-tydi-question-generator")
+
+def get_question(answer, context, max_length=64):
+  input_text = answer +" <<sep>> " + context
+  features = tokenizer([input_text], return_tensors='pt')
+  output = model.generate(input_ids=features['input_ids'], 
+               attention_mask=features['attention_mask'],
+               max_length=max_length)
+  return tokenizer.decode(output[0])
+context = "শচীন টেন্ডুলকারকে ক্রিকেট ইতিহাসের অন্যতম সেরা ব্যাটসম্যান হিসেবে গণ্য করা হয়।"
+answer = "শচীন টেন্ডুলকার"
+get_question(answer, context)
+# output: ক্রিকেট ইতিহাসের অন্যতম সেরা ব্যাটসম্যান কে?
+```
+
+## Citation
+```bibtex
+@inproceedings{xue2021mt5,
+  title={mT5: A Massively Multilingual Pre-trained Text-to-Text Transformer},
+  author={Xue, Linting and Constant, Noah and Roberts, Adam and
+          Kale, Mihir and Al-Rfou, Rami and Siddhant, Aditya and
+          Barua, Aditya and Raffel, Colin},
+  booktitle={Proceedings of the 2021 Conference of the North American
+             Chapter of the Association for Computational Linguistics:
+             Human Language Technologies},
+  pages={483--498},
+  year={2021}
+}
+```
+