NTA2/qwen3-4b-structured-lora

qwen3-4b-structured-output-lora (T4 Optimized)

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.

Training Hyperparameters

Parameter	Value
Base Model	Qwen/Qwen3-4B-Instruct-2507
Dataset	u-10bei/structured_data_with_cot_dataset_512_v5
Method	QLoRA (4-bit, Unsloth)
LoRA Rank (r)	128
LoRA Alpha	256
LoRA Dropout	0
Target Modules	q_proj, k_proj, v_proj, o_proj, gate_proj, up_proj, down_proj
Max Sequence Length	1024
Epochs	2
Batch Size	1 (per device)
Gradient Accumulation	16
Total Batch Size	16
Learning Rate	2e-4
Scheduler	cosine
Warmup Ratio	0.1
Weight Decay	0.05
Seed	3407
Optimizer	AdamW (8-bit)

Usage

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

base = "Qwen/Qwen3-4B-Instruct-2507"
adapter = "NTA2/qwen3-4b-structured-lora"

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_512_v5

Dataset License: MIT License. This dataset is used and distributed under the terms of the MIT License. Compliance: Users must comply with the MIT license (including copyright notice) and the base model's original terms of use.