Qwen3-4B-Instruct-SFT-03

shotalab/Qwen3-4B-Instruct-SFT-03

This repository provides a fully merged model (Base model + LoRA adapter) fine-tuned from Qwen/Qwen3-4B-Instruct-2507.

This repository contains complete model weights. You can use it directly without any additional adapters.

Training Objective

This model is fine-tuned to improve structured output accuracy (JSON / YAML / XML / TOML / CSV). Loss was applied only to the final assistant output, while intermediate reasoning (Chain-of-Thought) was masked during the SFT process.

Training Configuration (Pre-merge)

• Base model: Qwen/Qwen3-4B-Instruct-2507
• Method: QLoRA (4-bit)
• Max sequence length: 512
• Epochs: 1
• Learning rate: 3e-05
• LoRA Config: r=64, alpha=128

Usage

from transformers import AutoModelForCausalLM, AutoTokenizer
import torch

model_id = "shotalab/Qwen3-4B-Instruct-SFT-03"

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

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.