ha0yuan/Qwen3-ASR-LoRa-ChineseAviation-Tiny

Qwen3-ASR LoRA Chinese Fine-tuned Model

 

Model Description

This model is a LoRA fine-tuned version of Qwen/Qwen3-ASR-1.7B for Chinese speech recognition. The fine-tuning was performed using the PEFT library with LoRA method.

• —Base model: Qwen/Qwen3-ASR-1.7B
• —Fine-tuning method: LoRA
• —Language: Chinese
• —Task: Automatic Speech Recognition

Fine-tuning Details

• —LoRA rank: 16
• —LoRA alpha: 32
• —LoRA dropout: 0.05
• —Target modules: all-linear
• —Batch size: 4
• —Learning rate: 2e-4
• —Epochs: 20
• —Warmup ratio: 0.05

Usage

Installation

pip install transformers peft torch audio