mesolitica/gemma-2B-16k-instructions

TEXT GENERATIONPricing:Input $0.32 / Cached $0.064 / Output $1.6Concurrent Unit Cost:1Model Size:2.5BQuant:BF16Context Size:8kPublished:Feb 26, 2024Architecture:Transformer0.0K Featherless Exclusive Cold

mesolitica/gemma-2B-16k-instructions is a 2.5 billion parameter instruction-tuned language model developed by Mesolitica, based on the Gemma 2B architecture. It features a 8192-token context length and is specifically fine-tuned on a Malaysian instructions dataset. This model is designed for conversational AI and instruction-following tasks, particularly within the Malaysian linguistic context, demonstrating the base model's adaptability.

Loading preview...

Overview

mesolitica/gemma-2B-16k-instructions is a 2.5 billion parameter instruction-tuned model built upon the Gemma 2B architecture, developed by Mesolitica. It has been fine-tuned using a comprehensive Malaysian instructions dataset, making it particularly suitable for applications requiring understanding and generation of Malaysian language content. The model utilizes the exact Gemma Instruct chat template and supports a context length of 8192 tokens.

Key Capabilities

  • Instruction Following: Optimized to understand and respond to instructions, especially in the Malaysian language.
  • Malaysian Language Proficiency: Fine-tuned on a dedicated Malaysian synthetic dataset, enhancing its performance for local linguistic nuances.
  • Conversational AI: Designed for chat-based interactions, leveraging the Gemma Instruct chat template.
  • Efficient Deployment: The 2.5B parameter size allows for more accessible deployment compared to larger models, with support for 4-bit quantization.

Good for

  • Malaysian Language Applications: Ideal for chatbots, virtual assistants, and content generation targeting Malaysian users.
  • Instruction-Based Tasks: Suitable for tasks requiring the model to follow specific commands or answer questions based on provided instructions.
  • Research and Development: Serves as a demonstration of fine-tuning Gemma 2B for specific regional languages and instruction sets.
  • Resource-Constrained Environments: Its smaller parameter count and quantization support make it viable for deployment in environments with limited computational resources.

Finetunes
4 models