heegyu/LIMA2-13b-hf

TEXT GENERATIONConcurrency Cost:1Model Size:13BQuant:FP8Ctx Length:4kPublished:Aug 7, 2023Architecture:Transformer Cold

heegyu/LIMA2-13b-hf is a 13 billion parameter Llama 2 model, fine-tuned by heegyu, optimized for dialogue use cases. This model leverages the Llama 2 architecture developed by Meta, featuring a 4096-token context length. It is specifically fine-tuned using the 64bits/lima_vicuna_format data over 10 epochs, enhancing its performance in assistant-like chat scenarios.

Loading preview...

Model Overview

heegyu/LIMA2-13b-hf is a fine-tuned variant of Meta's Llama 2 13 billion parameter model. The base Llama 2 architecture is an auto-regressive language model utilizing an optimized transformer architecture, trained on 2 trillion tokens of publicly available data with a cutoff of September 2022. This specific model has been further fine-tuned for 10 epochs using the 64bits/lima_vicuna_format dataset, making it suitable for interactive conversational applications.

Key Capabilities

  • Dialogue Optimization: Fine-tuned for assistant-like chat use cases, building upon the Llama-2-Chat's alignment to human preferences for helpfulness and safety.
  • Llama 2 Foundation: Benefits from the robust Llama 2 pretraining, which includes a new mix of publicly available online data and a 4096-token context length.
  • Text Generation: Capable of generating text based on input prompts, leveraging its 13 billion parameters for nuanced responses.

Good For

  • Chatbots and Virtual Assistants: Its fine-tuning for dialogue makes it well-suited for creating conversational agents.
  • English Language Applications: Intended for commercial and research use primarily in English.
  • Instruction Following: Designed to follow specific formatting for chat versions, including INST and <<SYS>> tags, for optimal performance.