Tsunami-th/Tsunami-1.0-7B-Instruct

TEXT GENERATIONPricing:Input $0.4 / Cached $0.08 / Output $0.8Concurrent Unit Cost:1Model Size:7.6BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Oct 28, 2024License:apache-2.0Architecture:Transformer0.0K Open Weights Featherless Exclusive Cold

Tsunami-1.0-7B-Instruct is a 7.6 billion parameter Thai Large Language Model developed by Pollakrit Lorprasertkul. It is fine-tuned from Qwen2.5-7B using a Thai dataset, making it specialized for understanding and generating content in the Thai language. This model utilizes a 32K context length and is designed for instruction-following tasks in Thai. Its primary strength lies in its optimized performance for Thai natural language processing applications.

Loading preview...

Tsunami-1.0-7B-Instruct Overview

Tsunami-1.0-7B-Instruct is a 7.6 billion parameter instruction-tuned language model developed by Pollakrit Lorprasertkul. It is specifically fine-tuned from the Qwen2.5-7B architecture using an extensive Thai dataset, making it a specialized resource for Thai language processing. The model supports a context length of 32,768 tokens.

Key Capabilities

  • Thai Language Specialization: Optimized for understanding and generating text in Thai due to its fine-tuning on a dedicated Thai dataset.
  • Instruction Following: Designed to respond effectively to instructions, leveraging its instruction-tuned nature.
  • ChatML Prompt Template: Utilizes the ChatML format for structured conversational inputs, ensuring consistent interaction.

Use Cases

This model is particularly well-suited for applications requiring robust Thai language capabilities, such as:

  • Building chatbots or virtual assistants for Thai-speaking users.
  • Content generation and summarization in Thai.
  • Language understanding tasks specific to the Thai context.

Technical Details

The model is built upon the Qwen2.5-7B base and can be easily integrated into projects using the Hugging Face transformers library, with support for torch_dtype="auto" and device_map="auto" for efficient deployment.