datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
MBQ
MBQ: Modality-Balanced Quantization for Large Vision-Language Models [paper]
Installation
Clone this repo
git clone --recurse-submodules git@github.com:thu-nics/MBQ.git
Create a conda env
conda create -n qmllm python=3.10
Install packages and 3rdparty repos.
# Install LLaVA-NeXT
cd ./3rdparty/LLaVA-NeXT
pip install -e .
# Install lmms-eval
cd ./3rdparty/lmms-eval
pip install -e .
# Install qmllm
pip install -r requirements.txt
pip install -e .
Apply… See the full description on the dataset page: https://huggingface.co/datasets/kihgfft/MBQ.mbqq
MBQ: Modality-Balanced Quantization for Large Vision-Language Models [paper]
Installation
Clone this repo
git clone --recurse-submodules git@github.com:thu-nics/MBQ.git
Create a conda env
conda create -n qmllm python=3.10
Install packages and 3rdparty repos.
# Install LLaVA-NeXT
cd ./3rdparty/LLaVA-NeXT
pip install -e .
# Install lmms-eval
cd ./3rdparty/lmms-eval
pip install -e .
# Install qmllm
pip install -r requirements.txt
pip install -e .
Apply… See the full description on the dataset page: https://huggingface.co/datasets/kihgfft/mbqq.eZ83e-MbqcKOBmnv64