This repository contains a PEFT LoRA adapter for Qwen/Qwen3.5-9B, fine-tuned on synthetic FormStruct training data for structured form understanding from document images.

The released adapter is the early SFT checkpoint selected on a held-out development split. The base model weights are not included.

Loading

python

import torch
from peft import PeftModel
from transformers import AutoModelForImageTextToText, AutoProcessor

base_id = "Qwen/Qwen3.5-9B"
adapter_id = "D2I-CUHK-Shenzhen/FormStruct-Bench-SFT-Qwen3.5-9B"

processor = AutoProcessor.from_pretrained(adapter_id, trust_remote_code=True)
base_model = AutoModelForImageTextToText.from_pretrained(
    base_id,
    dtype=torch.bfloat16,
    device_map="auto",
    trust_remote_code=True,
)
model = PeftModel.from_pretrained(base_model, adapter_id)
model.eval()

Collection

This adapter is part of the FormStruct-Bench-SFT model collection, together with the Qwen3.6-35B-A3B adapter.

License

Apache-2.0, following the base model license. See LICENSE.