Downloads · 30 days
23
33% of all-time downloads
JingxuanWang/Qwen2.5-VL-7B_LoRA_image2poem_French
Qwen2.5-VL-7B_LoRA_image2poem_French is a machine learning model from JingxuanWang. Use it for the machine learning task on the model card, and read the license before you ship it in a product. The card lists the license as mit.
[!IMPORTANT] 1. We build an automated poetry data annotation pipeline driven by Prompt Engineering and powered by the Gemini-2.5-Pro API. A total of 2,369 images were processed: 2,345 successfully annotated, and 24 fi…
Downloads · 30 days
23
33% of all-time downloads
All-time downloads
70
Public
Parameters
8.3B
16.6 GB on disk
Likes
1
Public
Click a slice to open those files.
.safetensors16.6 GB · 100%
From the Hugging Face model README
[!IMPORTANT]
We build an automated poetry data annotation pipeline driven by Prompt Engineering and powered by the Gemini-2.5-Pro API. A total of 2,369 images were processed: 2,345 successfully annotated, and 24 filtered out as unsuitable for poem generation.
Using the 2,345 successfully annotated image–poem pairs, we perform Supervised Fine-Tuning (SFT) with the LoRA method via the open-source project LLaMA-Factory (hiyouga/LLaMA-Factory: Unified Efficient Fine-Tuning of 100+ LLMs & VLMs (ACL 2024)), applying it to Qwen2.5-VL-7B and obtaining a high-quality fine-tuned model.
We provide a Web UI dedicated to generating poetry from images. It supports three generation modes: invoking the Gemini-2.5-Pro API, using the original Qwen2.5-VL-7B model, and loading the LoRA fine-tuned Qwen2.5-VL-7B model. Users can upload an image and supply a custom prompt to steer style, theme, tone, or form, enabling personalized poetic output.
git clone https://github.com/hiyouga/LLaMA-Factory.git
conda create -n llama_factory python=3.10
conda activate llama_factory
cd LLaMA-Factory
pip install -e '.[torch,metrics]'
pip install -q -U google-genai
pip install qwen_vl_utils
git clone https://github.com/lawrencewjx58/Vison_NLP_Project.git
cd Vison_NLP_Project
python generate.py
python delete.py
python merge.py
python count.py
python transform.py
huggingface-cli download --resume-download Qwen/Qwen2.5-VL-7B-Instruct --local-dir {your_path} --local-dir-use-symlinks False --token {your_token}
cd LLaMA-Factory
llamafactory-cli webui
lawrencewjx58/Qwen2.5-VL-7B_LoRA_image2poem_Chinese
lawrencewjx58/Qwen2.5-VL-7B_LoRA_image2poem_French
cd Vison_NLP_Project
python UI_Web.py