You can deploy this model and start earning money today!

Ualikhan888/
qwen-kazsandra-homework

Name: qwen-kazsandra-homework
Availability: InStock

safetensors

Views

No views yet

Qwen KazSANDRA LoRA Adapter

This repository contains a LoRA adapter fine-tuned on the KazSANDRA dataset for sentiment classification.

Model

Base model: Qwen2.5-1.5B-Instruct
Method: LoRA fine-tuning
Task: Binary sentiment classification (negative / positive)

Training

Two configurations were tested:

Config A: r=8, alpha=16
Config B: r=32, alpha=64 (best)

The best configuration achieved higher accuracy on validation data.

Results

Best configuration:

r = 32
alpha = 64
accuracy ≈ 0.69

Compared to:

r = 8, alpha = 16 → accuracy ≈ 0.62

The larger LoRA rank showed better performance.

Usage

python

from peft import PeftModel
from transformers import AutoModelForCausalLM

base = AutoModelForCausalLM.from_pretrained("Qwen/Qwen2.5-1.5B-Instruct")
model = PeftModel.from_pretrained(base, "Ualikhan888/qwen-kazsandra-homework")