APMIC/caigun-lora-model-34B-v3

TEXT GENERATIONConcurrency Cost:2Model Size:34BQuant:FP8Ctx Length:32kPublished:Dec 21, 2023License:cc-by-nc-nd-4.0Architecture:Transformer Open Weights Cold

APMIC/caigun-lora-model-34B-v3 is a 34 billion parameter large language model based on the LLaMA architecture, fine-tuned on an Orca-style dataset. This model is designed for various general-purpose applications, leveraging its LLaMA foundation and specific fine-tuning. It incorporates a new optimizer in its latest version, aiming for broad utility across different tasks.

Loading preview...

Model Overview

APMIC/caigun-lora-model-34B-v3 is a 34 billion parameter large language model built upon the LLaMA architecture. It has been fine-tuned using an Orca-style dataset, which typically involves instruction-following and complex reasoning tasks, aiming to enhance its general-purpose capabilities. The model's development has seen several iterations, with version 3.0 utilizing a new optimizer to improve performance.

Key Characteristics

  • Architecture: Based on the robust LLaMA model family.
  • Parameter Count: Features 34 billion parameters, providing significant capacity for understanding and generating complex language.
  • Training Data: Fine-tuned on an Orca-style dataset, suggesting a focus on instruction-following and diverse task performance.
  • Version History: Evolved from earlier versions, including one fine-tuned for fake news detection (v1.0) and subsequent updates focusing on Orca-style datasets and optimizer improvements (v2.0, v3.0).

Intended Use

This model is designed for a variety of general-purpose applications, making it suitable for diverse use cases where a capable large language model is required. Its Orca-style fine-tuning suggests proficiency in following instructions and handling complex queries.

Finetunes
2 models