indischepartij/OpenMia-Indo-Mistral-7b-v3-refined

TEXT GENERATIONConcurrency Cost:1Model Size:7BQuant:FP8Ctx Length:4kPublished:Feb 4, 2024License:cc-by-nc-4.0Architecture:Transformer0.0K Open Weights Cold

OpenMia-Indo-Mistral-7b-v3-refined by indischepartij is a 7 billion parameter language model, fine-tuned from Mistral-7b, specifically designed for conversational capabilities in Bahasa Indonesia. This model, with a 4096-token context length, leverages an Alpaca dataset for its Indonesian language adaptation. It is primarily intended for applications requiring natural language processing and generation in Indonesian, offering a specialized solution for this linguistic context.

Loading preview...

OpenMia-Indo-Mistral-7b-v3-refined: Indonesian Language Fine-tuned Model

OpenMia-Indo-Mistral-7b-v3-refined is a 7 billion parameter language model developed by indischepartij, building upon the Mistral-7b architecture. This model is specifically fine-tuned to enhance its conversational abilities in Bahasa Indonesia, utilizing an Alpaca dataset for its linguistic adaptation. It represents an alpha-stage project focused on providing robust Indonesian language processing.

Key Capabilities

  • Bahasa Indonesia Conversation: Optimized for engaging in natural conversations in Indonesian.
  • Mistral-7b Base: Benefits from the strong foundational capabilities of the Mistral-7b model.
  • Alpaca Dataset Integration: Fine-tuned using an Alpaca dataset to improve Indonesian language understanding and generation.

Good for

  • Indonesian Language Applications: Ideal for chatbots, virtual assistants, and other NLP tasks requiring proficiency in Bahasa Indonesia.
  • Research and Development: Suitable for developers and researchers exploring Indonesian language models.
  • Early-stage Integration: Can be used in projects where a specialized Indonesian language model is beneficial, with awareness of its alpha status.

Performance Highlights

Evaluations on the Open LLM Leaderboard show an average score of 68.00. Notable scores include 84.22 on HellaSwag (10-Shot) and 81.53 on Winogrande (5-shot), indicating strong performance in common sense reasoning tasks. Detailed results are available on the Open LLM Leaderboard.

Finetunes
3 models