license: mit
SEA-LION
SEA-LION is a collection of LLMs which has been pretrained and instruct-tuned for the Southeast Asia (SEA) region. The models range from 3 billion to 7 billion parameters. This is the card for the SEA-LION 3B model.
SEA-LION stands for Southeast Asia Languages In One Network.
Model Details
Model Description
The SEA-LION model is a significant leap forward in the field of natural language processing, specifically trained to understand Southeast Asia (SEA) regional context.
SEA-LION is built on the robust MPT architecture and utilize a vocabulary size of 256K.
The model employs our custom SEABPETokenizer for tokenization. Our SEABPETokenizer is specially tailored for SEA languages, ensuring optimal model performance.
The training data for SEA-LION encompasses 980B tokens.
- Developed by: Products Pillar, AI Singapore
- Funded by: Singapore NRF
- Model type: Decoder
- Languages: English, Chinese, Indonesian, Malay, Thai, Vietnamese, Filipino, Tamil, Burmese, Khmer, Lao
- License: MIT License
Training Details
Data
SEA-LION was trained on 980B tokens of the following data:
| Data Source | Tokens | Percentage |
|---|---|---|
| RefinedWeb - English | 571.3B | 58.20% |
| mC4 - Chinese | 91.2B | 9.29% |
| mC4 - Indonesian | 14.7B | 1.50% |
| mC4 - Malay | 2.9B | 0.29% |
| mC4 - Filipino | 5.3B | 0.54% |
| mC4 - Burmese | 4.9B | 0.49% |
| mC4 - Vietnamese | 63.4B | 6.46% |
| mC4 - Thai | 21.6B | 2.20% |
| mC4 - Lao | 1.1B | 0.12% |
| mC4 - Khmer | 3.9B | 0.40% |
| mC4 - Tamil | 10.2B | 1.04% |
| the Stack - Python | 41.8B | 4.26% |
| the Stack - Javascript | 55.6B | 5.66% |
| the Stack - Shell | 2.5B | 0.26% |
| the Stack - SQL | 12.8B | 1.31% |
| the Stack - Markdown | 26.6B | 2.71% |
| RedPajama - StackExchange | 21.2B | 2.16% |
| RedPajama - ArXiv | 30.6B | 3.12% |
Infrastructure
SEA-LION was trained using MosaicML Composer on the following hardware:
| Training Details | SEA-LION 3B |
|---|---|
| AWS EC2 p4d.24xlarge | 30 instances |
| Nvidia A100 40GB GPU | 240 |
| Training Duration | 14 days |
Configuration
| HyperParameter | SEA-LION 3B |
|---|---|
| Precision | bfloat16 |
| Optimizer | decoupled_adamw |
| Scheduler | cosine_with_warmup |
| Learning Rate | 1.6e-4 |
| Global Batch Size | 1200 |
| Micro Batch Size | 5 |
Technical Specifications
Model Architecture and Objective
SEA-LION is a decoder model using the MPT architecture.
| Parameter | SEA-LION 3B |
|---|---|
| Layers | 32 |
| d_model | 2560 |
| head_dim | 20 |
| Vocabulary | 256000 |
| Sequence Length | 2048 |
Tokenizer Details
We sample 20M lines from the training data to train the tokenizer.
The framework for training is SentencePiece.
The tokenizer type is Byte-Pair Encoding (BPE).
The Team
Lam Zhiwen Clarence
Leong Weiqi
Li Yier
Liu Darius
Lovenia Holy
Montalan Jann Railey
Ng Raymond
Ngui Jian Gang
Nguyen Ngan Thanh
Ong Tat-Wee David
Rengarajan Hamsawardhini
Susanto Yosephine
Tai Ngee Chia
Tan Choon Meng
Teo Jin Howe
Teo Leslie
Teo Wei Yi
Tjhi William
Yeo Yeow Tong
Yong Xianbin
Acknowledgements
AI Singapore is a national programme supported by the National Research Foundation, Singapore and hosted by the National University of Singapore. Any opinions, findings and conclusions or recommendations expressed in this material are those of the author(s) and do not reflect the views of National Research Foundation, Singapore.
Contact
For more info, please contact us at sealion@aisingapore.org