qwen3-4b-sft-merged-v2v5ver1/README.md

base_model, datasets, language, license, library_name, pipeline_tag, tags

base_model	datasets	language	license	library_name	pipeline_tag	tags
Qwen/Qwen3-4B-Instruct-2507	takami2022/structured_data_merged_v2v5_0222	en	apache-2.0	transformers	text-generation	merged sft structured-output

qwen3-4b-sft-merged-v2v5ver1

This repository provides a merged 16-bit model produced by fine-tuning Qwen/Qwen3-4B-Instruct-2507 with QLoRA (4-bit, Unsloth) and then merging the LoRA adapter into the base model weights.

This model is fully self-contained — no adapter loading required. Intended as the base model for subsequent DPO training.

Training Configuration

• Base model: Qwen/Qwen3-4B-Instruct-2507
• Dataset: takami2022/structured_data_merged_v2v5_0222
• Method: QLoRA (4-bit) → merged to 16-bit
• Max sequence length: 1024
• Epochs: 3
• Learning rate: 1e-06
• LoRA: r=64, alpha=128
• CoT mask: enabled

Usage

from transformers import AutoModelForCausalLM, AutoTokenizer
import torch

model_id = "takami2022/qwen3-4b-sft-merged-v2v5ver1"

tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(
    model_id,
    torch_dtype=torch.bfloat16,
    device_map="auto",
)

Notes

• This merged model is the output of SFT and serves as the starting point for DPO.