base_model: Matukaze/test105
datasets:
- u-10bei/sft_alfworld_trajectory_dataset_v5
language:
- en
license: apache-2.0
library_name: peft
pipeline_tag: text-generation
tags:
- lora
- agent
- tool-use
- alfworld
- dbbench
qwen2.5-7b-Instruct-trajectory-lora
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
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.