This repository provides a LoRA adapter fine-tuned from Matukaze/test105 using LoRA + Unsloth. This model initialised from Qwen2.5-7B-Instruct.

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 Configuration

Base model: Matukaze/test105
Method: LoRA (full precision base)
Max sequence length: 1024
Epochs: 0.7
Learning rate: 5e-06
LoRA: r=32, alpha=128
This model was fine-tuned sequentially using LoRA
First stage:fine-tuned on DBBench SFT dataset and merged.
Second stage:furter fine-tuned on ALFWorld SFT datasset and merged.

-This model architecture remains unchanged.

Usage

python

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

base = "Matukaze/test105"
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/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.

Matukaze/test107

qwen2.5-7b-Instruct-trajectory-lora

Training Objective

Training Configuration

Usage

Sources & Terms (IMPORTANT)

Matukaze/
test107