kabuizuchi-trading/qwen3-4b-instruct-2507-lora001

＜# Qwen3-4B-Instruct-2507-lora001＞

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 Configuration

• —Base model: Qwen/Qwen3-4B-Instruct-2507
• —Method: QLoRA (4-bit)
• —Max sequence length: 1536
• —Epochs: 2
• —Learning rate: 5e-05
• —LoRA: r=128, alpha=256

Data / Split

• —Dataset: u-10bei/structureddatawithcotdataset512v2
• —Validation split: 0.03
• —Seed: 3407

Batch / Steps

• —Train batch size (per device): 4
• —Eval batch size (per device): 4
• —Gradient accumulation steps: 6
• —Effective batch size: 24 × (numberofgpus) (effective = perdevicetrainbs × gradaccum × num_gpus)
• —Max steps: -1 (epoch-based if -1)
• —Logging steps: 20
• —Eval strategy: steps (eval_steps=200)
• —Save strategy: steps (savesteps=200, savetotal_limit=6)

Optimization

• —LR scheduler: cosine
• —Warmup ratio: 0.03
• —Weight decay: 0.01

CoT masking (Output-only supervision)

• —maskcot: enabled (SFTMASK_COT=1)
• —output_markers: Output:, OUTPUT:, Final:, Answer:, Result:, Response:
• —outputlearnmode: after_marker
• —upsampling: disabled (SFTUSEUPSAMPLING=0)

Usage

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

base = "Qwen/Qwen3-4B-Instruct-2507"
adapter = "your_id/your-repo"

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/structureddatawithcotdataset512v2

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.