kyujinpy/PlatYi-34B-Llama-Q-v2

TEXT GENERATIONConcurrency Cost:2Model Size:34BQuant:FP8Ctx Length:32kPublished:Dec 10, 2023License:cc-by-nc-sa-4.0Architecture:Transformer Open Weights Cold

PlatYi-34B-Llama-Q-v2 by Kyujin Han (kyujinpy) is a 34 billion parameter auto-regressive language model built on the Yi-34B transformer architecture, fine-tuned using Q-LoRA on the Open-Platypus dataset. This model is designed for general text generation tasks, offering a balance of performance and efficiency. It achieves an average score of 67.88 on the Open LLM Leaderboard benchmarks, including 76.59 on MMLU and 85.09 on HellaSwag, making it suitable for a range of natural language processing applications.

Loading preview...

Model Overview

PlatYi-34B-Llama-Q-v2 is a 34 billion parameter auto-regressive language model developed by Kyujin Han (kyujinpy). It is based on the Yi-34B transformer architecture and was fine-tuned using Q-LoRA with a lora_r value of 64. The training utilized the garage-bAInd/Open-Platypus dataset, and the developer notes modifications to templates and warmup steps to address prior model issues.

Key Capabilities & Performance

This model is designed for general text generation. Its performance is evaluated on the Open LLM Leaderboard, where it achieved an average score of 67.88. Notable benchmark results include:

  • MMLU (5-Shot): 76.59
  • HellaSwag (10-Shot): 85.09
  • ARC (25-Shot): 61.09
  • TruthfulQA (0-shot): 52.65
  • Winogrande (5-shot): 82.79
  • GSM8k (5-shot): 49.05

Use Cases

Given its benchmark performance, PlatYi-34B-Llama-Q-v2 is suitable for a variety of natural language processing tasks requiring robust text generation and understanding. Its fine-tuning approach suggests potential for efficient deployment, making it a candidate for applications where a 34B parameter model can be effectively utilized.

Finetunes
2 models