Beta

Explore
Marketplace
Neural Labs
Playground
Wallet
Docs

Beta

Explore Marketplace Neural Labs Wallet Docs

Build

Manage

My assets Manage wallet Organization

Community

Discord X (Twitter)Telegram

Docs Privacy Policy Terms and Conditions

Proud member of

Proud member of

Docs Cookie Policy

Privacy Policy Terms and Conditions

© 2026 AlphaNeural AI. All rights reserved.

qwen3-4b-structured-output-lora-002 – AI Model by yuuuki11 | AlphaNeural AI | AlphaNeural AI

You can deploy this model and start earning money today!

yuuuki11/
qwen3-4b-structured-output-lora-002

like

0

structured-output

text-generation

u-10bei/structured_data_with_cot_dataset_512_v5

Qwen/Qwen3-4B-Instruct-2507

Views

No views yet

qwen3-4b-structured-output-lora-002

This repository provides a LoRA adapter fine-tuned from Qwen/Qwen3-4B-Instruct-2507 using QLoRA (4-bit with 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) for the course competition setting.

Loss is applied to assistant output tokens (with output-marker-aware masking in the training pipeline).

Training Configuration

Case ID: R2_v5_len768_seed42_s120
Base model: Qwen/Qwen3-4B-Instruct-2507
Dataset: u-10bei/structured_data_with_cot_dataset_512_v5
Method: QLoRA (4-bit, Unsloth)
Max sequence length: 768
Epochs: 2
Learning rate: 1e-5
Warmup ratio: 0.03
Max steps: 120
Seed: 42

Usage

python

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

base = "Qwen/Qwen3-4B-Instruct-2507"
adapter = "yuuuki11/qwen3-4b-structured-output-lora-002"

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

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.