This is v5.2 of the SFT training, focusing on hyperparameter tuning to address overfitting in v5. v5 achieved 0.73981, which was lower than v2 (0.75074) due to Epoch=2 causing overfitting.

Changes from v5

Parameter	v5	v5.2	Rationale
Dataset	3,869 samples	3,869 samples	Same (XML errors removed)
MAX_SEQ_LEN	1024	1024	Same
Epochs	2	1	Reduced to prevent overfitting
Learning Rate	5e-6	1e-06	Lower for more stable training
Warmup Ratio	10%	10%	Increased to reduce early instability

v5.2 Key Improvements

Epoch=1: v5's Epoch=2 caused overfitting (train/loss increased at end)
Learning Rate 5e-6: More conservative learning to prevent overfitting
Warmup Ratio 10%: Longer warmup for training stability
Data unchanged: 3,869 samples with XML errors removed

Score History

Version	Data	Score	Notes
v2	3,933	0.75074	Best score baseline
v5	3,869	0.73981	Epoch=2 overfitting
v5.2	3,869	(pending)	Hyperparam tuning

Training Objective

This adapter is trained to improve structured output accuracy (JSON / YAML / XML / TOML / CSV) for the StructEval-T benchmark.

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

Training Configuration

Base model: Qwen/Qwen3-4B-Instruct-2507
Method: QLoRA (4-bit, Unsloth)
Max sequence length: 1024
Epochs: 1
Learning rate: 1e-06
Warmup ratio: 10%
Batch size: 2 (effective: 16)
Gradient accumulation: 8
LoRA: r=64, alpha=128
CoT masking: enabled (loss on final output only)

Usage

python

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

base = "Qwen/Qwen3-4B-Instruct-2507"
adapter = "your_id/qwen3-4b-structured-output-lora-v5.5A"

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_v2

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.

kmd2525/qwen3-4b-structured-output-lora-v5.5A

qwen3-4b-structured-output-lora-v5.5A

Version: v5.2 — Hyperparameter Tuning