---
base_model: Qwen/Qwen3-4B-Instruct-2507
datasets:
- u-10bei/sft_alfworld_trajectory_dataset_v5
- u-10bei/dbbench_sft_dataset_react
language:
- en
license: mit
library_name: peft
pipeline_tag: text-generation
tags:
- lora
- agent
- tool-use
- alfworld
- dbbench
- merge
- dare-ties
---

# qwen3-4b-agent-trajectory-lora (Merged Model)

This repository provides a **LoRA adapter** fine-tuned from
**Qwen3-4B-Instruct-2507** using **LoRA + Unsloth**.

**[Merge Information]**
This model is a merged adapter created using **MergeKit (DARE-TIES)** to combine the strengths of the following two models:
1. `maru-miya/lora_agentbench_qwen3_4b_d20_t1` (General Agent focus)
2. `maru-miya/lora_agentbench_qwen3_4b_d21_t9_db` (DB specialized focus)

This repository contains **LoRA adapter weights only**.
The base model must be loaded separately.

## Training Objective

This adapter is trained to improve **multi-turn agent task performance**
on ALFWorld (household tasks) and DBBench (database operations).

Loss is applied to **all assistant turns** in the multi-turn trajectory,
enabling the model to learn environment observation, action selection,
tool use, and recovery from errors.

## Training & Merge Configuration

**[Merge Settings (MergeKit)]**
- Method: DARE-TIES
- Base model for merge: Qwen/Qwen3-4B-Instruct-2507
- Models & Parameters:
  - `maru-miya/lora_agentbench_qwen3_4b_d20_t1` (weight: 1.0, density: 0.7)
  - `maru-miya/lora_agentbench_qwen3_4b_d21_t9_db` (weight: 1.2, density: 0.7)

**[Original LoRA Adapter 1: d20_t1]**
- Base model: Qwen/Qwen3-4B-Instruct-2507
- Method: LoRA (full precision base)
- Max sequence length: 2048
- Epochs: 2
- Learning rate: 2e-06
- LoRA: r=16, alpha=32

**[Original LoRA Adapter 2: d21_t9_db]**
- Base model: Qwen/Qwen3-4B-Instruct-2507
- Method: LoRA (full precision base)
- Max sequence length: 2048
- Epochs: 2
- Learning rate: 2e-05
- LoRA: r=16, alpha=32

## Usage

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

base = "Qwen/Qwen3-4B-Instruct-2507"
adapter = "maru-miya/merged_agentbench_qwen3_4b_dare_ties_t1"

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/sft_alfworld_trajectory_dataset_v5

u-10bei/dbbench_sft_dataset_react

This repository does NOT redistribute the dataset.
Users must comply with the dataset license and base model terms.