Description
Jay is a 367 million parameters open source Generative pretrained Decoder Auto Regressive Language Model for Hindi.
Jay is a monolingual Hindi GenAI Decoder Language model.
This is a pretrained model from scratch at a context size of 1024.
This model is not either chat-tuned or fine-tuned.
We recommend to fine-tune/chat-tune this pretrained model on Hindi chat or Hindi instruction datasets. Please use PyTorch for fine-tuning/instruction-tuning.
We also recommend to perform continual pretraining on Hindi dataset if possible before fine-tuning.
This model is strictly prohibited to use for commercial purposes.
If you use our model, please cite our paper Niyogi and Bhattacharya, 2024
The architecture of Jay is different than the language models, mentioned in Niyogi and Bhattacharya, 2024
Model Architecture
Transformer Decoder Auto Regressive Model
Limitations
The model was trained on data that contains toxic language, unsafe content, and societal biases originally crawled from the internet. Therefore, the model may amplify those biases and return toxic responses especially when prompted with toxic prompts. The model may generate answers that may be inaccurate, omit key information, or include irrelevant or redundant text producing socially unacceptable or undesirable text, even if the prompt itself does not include anything explicitly offensive.
Gyan AI Research does own the output generated from the model.
Citations
@misc{niyogi2024paramanu,
title={Paramanu: A Family of Novel Efficient Indic Generative Foundation Language Models},
author={Mitodru Niyogi and Arnab Bhattacharya},
year={2024},
eprint={2401.18034},
archivePrefix={arXiv},
primaryClass={id='cs.CL' full_name='Computation and Language' is_active=True alt_name='cmp-lg' in_archive='cs' is_general=False description='Covers natural language processing. Roughly includes material in ACM Subject Class I.2.7. Note that work on artificial languages (programming languages, logics, formal systems) that does not explicitly address natural-language issues broadly construed (natural-language processing, computational linguistics, speech, text retrieval, etc.) is not appropriate for this area.'}
}