deepkick/qwen3-4b-structured-sft-lora-v05-merged

qwen3-4b-structured-sft-lora-v05-merged

This is a fully merged model (base + LoRA adapter) fine-tuned from Qwen/Qwen3-4B-Instruct-2507 using QLoRA (4-bit, Unsloth).

This model can be used directly without loading a separate adapter.

Training Configuration

• —Base model: Qwen/Qwen3-4B-Instruct-2507
• —Method: QLoRA (4-bit) → merged
• —Max sequence length: 1024
• —Epochs: 2
• —Learning rate: 2e-06
• —LoRA: r=64, alpha=128
• —Dataset: 9 datasets merged & cleaned (1500 samples)

Usage

from transformers import AutoModelForCausalLM, AutoTokenizer
import torch

model_id = "deepkick/qwen3-4b-structured-sft-lora-v05-merged"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(
    model_id,
    torch_dtype=torch.float16,
    device_map="auto",
)

Sources & Terms (IMPORTANT)

Training data: 9 datasets from u-10bei and daichira (see metadata above). Dataset License: MIT License. Compliance: Users must comply with the MIT license and the base model's original terms of use.