qwen3-4b-structured-output-lora-v7

qwen3-4b-structured-output-lora

This repository provides a LoRA adapter fine-tuned from Qwen/Qwen3-4B-Instruct-2507 using QLoRA (4-bit, Unsloth).

This repository contains LoRA adapter weights only. The base model must be loaded separately.

Training Objective

This adapter is trained to improve structured output accuracy (JSON / YAML / XML / TOML / CSV).

Loss is applied only to the final assistant output, while intermediate reasoning (Chain-of-Thought) is masked.

Data preprocessing: Markdown fences (```json, ```yaml, etc.) and text preambles ("Here's the converted ...", etc.) are automatically stripped from assistant responses before training, ensuring the model learns to produce clean structured output without formatting artifacts.

Training Configuration

• Base model: Qwen/Qwen3-4B-Instruct-2507
• Method: QLoRA (4-bit)
• Primary dataset: u-10bei/structured_data_with_cot_dataset_v2
• Secondary dataset: daichira/structured-5k-mix-sft
• Dataset mixing: both datasets concatenated after column harmonization
• Max sequence length: 1024
• Epochs: 2
• Learning rate: 2e-05
• LoRA: r=64, alpha=128
• CoT masking: enabled
• Data preprocessing: markdown fence stripping (enabled)

Usage

from transformers import AutoModelForCausalLM, AutoTokenizer
from peft import PeftModel
import torch

base = "Qwen/Qwen3-4B-Instruct-2507"
adapter = "a-kuratani/qwen3-4b-structured-output-lora-v7"

tokenizer = AutoTokenizer.from_pretrained(base)
model = AutoModelForCausalLM.from_pretrained(
    base,
    torch_dtype=torch.float16,
    device_map="auto",
)
model = PeftModel.from_pretrained(model, adapter)

Sources & Terms (IMPORTANT)

Training data:

• u-10bei/structured_data_with_cot_dataset_v2
• daichira/structured-5k-mix-sft

Dataset Licenses:

• u-10bei/structured_data_with_cot_dataset_v2: License not explicitly specified on HuggingFace. Please verify with the dataset author.
• daichira/structured-5k-mix-sft: CC-BY-4.0 (Creative Commons Attribution 4.0)

Compliance: Users must comply with each dataset's license terms (including attribution requirements for CC-BY-4.0) and the base model's original terms of use (Apache 2.0)."""