AmalMa/qwen2.5-1.5b-github-actions-qlora

Qwen2.5-1.5B GitHub Actions QLoRA

A QLoRA fine-tuned version of Qwen2.5-1.5B-Instruct specialized in GitHub Actions troubleshooting.

The model was trained on approximately 1.7K Q&A examples covering caching, matrix builds, secrets, permissions, artifacts, runners, deployments, reusable workflows, concurrency, and expressions.

Training

• Base model: Qwen/Qwen2.5-1.5B-Instruct
• Method: QLoRA / PEFT
• Training data: ~1.7K GitHub Actions troubleshooting Q&A examples
• Mean token accuracy: 67.1%

Usage

This repository contains a LoRA adapter, so the Qwen2.5-1.5B-Instruct base model is required.

from transformers import AutoModelForCausalLM, AutoTokenizer
from peft import PeftModel

base_id = "Qwen/Qwen2.5-1.5B-Instruct"
adapter_id = "AmalMa/qwen2.5-1.5b-github-actions-qlora"

tokenizer = AutoTokenizer.from_pretrained(base_id)
base_model = AutoModelForCausalLM.from_pretrained(
    base_id,
    device_map="auto"
)

model = PeftModel.from_pretrained(base_model, adapter_id)

Limitations

This is an experimental portfolio model, not a production-ready GitHub Actions assistant. It may generate incorrect technical recommendations, so important answers should be verified against official documentation.