Stable Diffusion Chatbot

Generate FREE AI images on your laptop with this stable diffusion chatbot.

Requirements for running it locally on laptop

• Windows / Mac / Linux with Git installed
• Python 3.9+
• Conda Environment (Python)
• A FREE MessengerX.io Creator Account
• Ngrok for Tunneling
• Desktop / Laptop with a minimum of 12GB RAM
• GPU is required for faster inference (MPS on MAC M1 & CUDA is supported for Nvidia GPU)
• Note: Inference time can take anywhere between 60s to 180s depending on your device capabilities
• Some optimizations may affect the quality of image produced for complex prompts

Setup Conda Environment

A suitable conda environment named ldm can be created

For windows / linux:
conda env create -f environment.yml

For m1 mac: MAC OS 12.3+ (minimum requirements)
brew install cmake
conda env create -f environment-m1.yml

conda activate ldm

Setup Weights

Currently the following checkpoints are available:

• sd-v1-1.ckpt
• sd-v1-2.ckpt
• sd-v1-3.ckpt
• sd-v1-4.ckpt

Download any one of the above weights

Linking & Running the chatbot

After obtaining the stable-diffusion-v1-*-original weights For Linux / Mac OS

mkdir -p models/ldm/stable-diffusion-v1/
ln -s <path/to/model.ckpt> models/ldm/stable-diffusion-v1/model.ckpt 

For Windows

mkdir -p models\ldm\stable-diffusion-v1\
mklink /h models\ldm\stable-diffusion-v1\model.ckpt <path\to\model.ckpt>

Usage: Can take about 30s to 60s to intialize and load the model

python3 run.py

You should get an output in the console on successful start:

Loading model from models/ldm/stable-diffusion-v1/model.ckpt
Global Step: 470000
UNet: Running in eps-prediction mode
...

Start your ngrok.io tunnel

ngrok http 5000

Update your webhook

Update your bot Webhook URL at MessengerX.io Portal with the url provided as shown below to continue development Your webhook URL would be of the format as in the example below

https://<Your NGROK URL>/webhooks/machaao/incoming