Weyaxi/Stable-Platypus2-13B-QLoRA-0.80-epoch

TEXT GENERATIONConcurrency Cost:1Model Size:13BQuant:FP8Ctx Length:4kPublished:Aug 27, 2023License:llama2Architecture:Transformer Open Weights Cold

Weyaxi/Stable-Platypus2-13B-QLoRA-0.80-epoch is a 13 billion parameter language model, created by Weyaxi, resulting from a merge of StableBeluga-13B and Platypus2-13B-QLoRA-0.80-epoch. This model is designed for general language understanding and generation tasks, leveraging the strengths of its merged components. It offers a 4096-token context length and demonstrates balanced performance across various benchmarks, making it suitable for diverse applications requiring robust language capabilities.

Loading preview...

Model Overview

Weyaxi/Stable-Platypus2-13B-QLoRA-0.80-epoch is a 13 billion parameter language model developed by Weyaxi. This model is a strategic merge of two distinct base models: stabilityai/StableBeluga-13B and Platypus2-13B-QLoRA-0.80-epoch. The merging approach aims to combine the strengths of both foundational models to achieve enhanced performance across a spectrum of natural language processing tasks.

Performance Benchmarks

The model's capabilities have been evaluated on the Open LLM Leaderboard, showcasing its general proficiency. Key benchmark results include:

  • Avg. Score: 54.53
  • ARC (25-shot): 62.29
  • HellaSwag (10-shot): 82.46
  • MMLU (5-shot): 57.09
  • TruthfulQA (0-shot): 51.41
  • Winogrande (5-shot): 76.56

While demonstrating solid performance in reasoning, common sense, and factual recall, the model shows a lower score in mathematical reasoning (GSM8K: 3.56). Its 4096-token context length supports processing moderately long inputs.

Ideal Use Cases

This model is well-suited for applications requiring:

  • General-purpose text generation and understanding.
  • Tasks benefiting from a balanced performance across various language benchmarks.
  • Scenarios where a 13B parameter model with a 4096-token context window is appropriate for resource constraints and performance needs.