All LoRA adapter weights have been merged into the base model, so the resulting checkpoint is a self-contained model that can be loaded directly without needing a separate adapter.

Training Objective

The model was trained to improve multi-turn agent task performance on ALFWorld (household tasks) and DBBench (database operations).

Loss was 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 Configuration

Base model: Qwen/Qwen3-4B-Instruct-2507
Method: LoRA (full-precision base) - weights merged after training
Max sequence length: 2048
Epochs: 1
Learning rate: 2e-04
LoRA: r=16, alpha=32

Usage

python

from transformers import AutoModelForCausalLM, AutoTokenizer
import torch

model_name = "edomurasaki/qwen3-4b-agent-trajectory"   # the merged model repo

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

Sources & Terms (IMPORTANT)

Training data: u-10bei/sft_alfworld_trajectory_dataset_v5

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.