aisingapore/Llama-SEA-LION-v3-8B-IT

TEXT GENERATIONConcurrency Cost:1Model Size:8BQuant:FP8Ctx Length:32kTool Calling:SupportedPublished:Dec 11, 2024License:llama3.1Architecture:Transformer0.0K Cold

Llama-SEA-LION-v3-8B-IT is an 8 billion parameter instruction-tuned decoder-only language model developed by AI Singapore, based on the Llama 3.1 architecture. It is specifically optimized for Southeast Asian languages, supporting Burmese, Chinese, English, Filipino, Indonesian, Javanese, Khmer, Lao, Malay, Sundanese, Tamil, Thai, and Vietnamese. With a context length of 128k tokens, this model excels in general language capabilities and instruction-following tasks across the SEA region.

Loading preview...

Overview

Llama-SEA-LION-v3-8B-IT is an 8 billion parameter instruction-tuned model developed by AI Singapore, building upon the Llama 3.1 architecture. It is part of the SEA-LION (Southeast Asian Languages In One Network) collection, which focuses on pretraining and instruct-tuning for the Southeast Asia region. The model supports a wide array of languages including Burmese, Chinese, English, Filipino, Indonesian, Javanese, Khmer, Lao, Malay, Sundanese, Tamil, Thai, and Vietnamese.

Key Capabilities

  • Multilingual Instruction Following: Instruction-tuned in English and various SEA languages like Indonesian, Javanese, Sundanese, Tamil, Thai, and Vietnamese.
  • Extended Context Length: Features a context length of 128k tokens, utilizing the default Llama 3.1 8B Instruct tokenizer.
  • Comprehensive Evaluation: Evaluated using the SEA-HELM evaluation benchmark for general language capabilities (QA, Sentiment, Toxicity, Translation, Summarization, Reasoning, NLI, linguistic diagnostics) and instruction-following capabilities with SEA-IFEval and SEA-MTBench.

Good For

  • Applications requiring strong instruction-following in Southeast Asian languages.
  • Research and development focusing on multilingual NLP tasks within the SEA region.
  • Use cases benefiting from a large context window for processing longer texts in supported languages.

Limitations

Users should be aware that the model may exhibit hallucinations and occasional inconsistencies in reasoning. It has not been aligned for safety, and developers are advised to implement their own safety fine-tuning and security measures.

Finetunes
2 models