search

camenduru / MiniGPT-4-colab

The Unlicense
120 stars 13 forks source link

readme

šŸ£ Please follow me for new updates https://twitter.com/camenduru
šŸ”„ Please join our discord server https://discord.gg/k5BwmmvJJU
šŸ„³ Please join my patreon community https://patreon.com/camenduru

šŸš¦ WIP šŸš¦

šŸ¦’ Colab

Colab Info
Open In Colab 13B Pro Colab (A100) we need more than 20GB vram šŸ˜­
Open In Colab 7B Pro Colab (T4) we need more than 12GB vram šŸ˜€ but we need High-RAM šŸ˜

Main Repo

https://github.com/Vision-CAIR/MiniGPT-4

Paper

https://arxiv.org/abs/2304.10592

Tutorial

https://www.youtube.com/watch?v=WzYBMWc6Zqk

Models License

Model License
https://huggingface.co/lmsys/vicuna-13b-delta-v0 From https://vicuna.lmsys.org: The online demo is a research preview intended for non-commercial use only, subject to the model License of LLaMA, Terms of Use of the data generated by OpenAI, and Privacy Practices of ShareGPT. Please contact us If you find any potential violation. The code is released under the Apache License 2.0.
https://huggingface.co/decapoda-research/llama-13b-hf https://huggingface.co/decapoda-research/llama-13b-hf/blob/main/LICENSE

Output

Screenshot 2023-04-20 182643 Screenshot 2023-04-20 182019

For RunPod

https://runpod.io/gsc?template=9m2wl11ypy&ref=iqi9iy8y

Please copy-paste the code below into a new notebook. (We need more than 20GB vram)

!apt update
!apt-get install git-lfs
!git lfs install

%env HF_HOME=/workspace/cache/huggingface

!git clone -b dev https://github.com/camenduru/minigpt4
!wget https://huggingface.co/ckpt/minigpt4/resolve/main/minigpt4.pth -O /workspace/minigpt4/checkpoint.pth
!wget https://huggingface.co/ckpt/minigpt4/resolve/main/blip2_pretrained_flant5xxl.pth -O /workspace/minigpt4/blip2_pretrained_flant5xxl.pth

!pip install -q salesforce-lavis
!pip install -q bitsandbytes
!pip install -q accelerate
!pip install -q gradio==3.27.0
!pip install -q git+https://github.com/huggingface/transformers -U

%cd /workspace/minigpt4
!python app.py