[Deploy Code] Align-Anything support MiniCPM-o-2_6 inference
❕ We are very pleased to announce that align-anything now supports fine-tuning for MiniCPM-O-2.6. Compared to the community's implementation, we believe our solution is more user-friendly. You just need to run the following script after installation to start training without modifying any parameters.
- Installation:
# We tested on the H800 computing cluster, and this version of CUDA works well.
# You can adjust this version according to the actual situation of the computing cluster.
conda install nvidia/label/cuda-12.2.0::cuda
export CUDA_HOME=$CONDA_PREFIX
pip install -e .[train]
pip install -e .[minicpmo]
- train
cd scripts
bash minicpmo_sft.sh
🎉 We are very pleased to announce that align-anything now supports online inference for MiniCPM-O-2.6. This update brings significant enhancements to its multimodal inference capabilities across video, images, audio, and text. We have deployed a temporary interactive link for your use; welcome to try it out: https://618cded0a3363b9686.gradio.live. For more details, please refer to our repo: https://github.com/PKU-Alignment/align-anything
🏃 We will be fine-tuning it with our all-modality dataset shortly and will release the training code for SFT, DPO, and PPO. Stay tuned!