Qwen2.5-VL-UniSVG-finetuned

<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->

scale_qwen_new

This model is a fine-tuned version of Qwen2.5-VL-7B-Instruct on UniSVG and MMSVG-Icon.

Model description

The model is a fine-tuned version of Qwen2.5-VL-7B-Instruct, using image-to-svg, text-to-svg and svg understanding data from UniSVG and MMSVG-Icon.

Intended uses & limitations

The model can generate SVG from image/text, answer the attributes of a given SVG. Though the model is an SVG expert, it's still long way to go to business usage.

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 0.0001
• train_batch_size: 1
• eval_batch_size: 8
• seed: 42
• distributed_type: multi-GPU
• num_devices: 32
• gradient_accumulation_steps: 8
• total_train_batch_size: 256
• total_eval_batch_size: 256
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.05
• num_epochs: 2.0

Training results

Framework versions

• Transformers 4.49.0
• Pytorch 2.6.0+cu124
• Datasets 3.3.2
• Tokenizers 0.21.1

Cite

@inproceedings{li2025unisvg,
  title={UniSVG: A Unified Dataset for Vector Graphic Understanding and Generation with Multimodal Large Language Models},
  author={Li, Jinke and Yu, Jiarui and Wei, Chenxing and Dong, Hande and Lin, Qiang and Yang, Liangjing and Wang, Zhicai and Hao, Yanbin},
  booktitle={Proceedings of the 33rd ACM international conference on multimedia},
  year={2025}
}