README.md

NOS Playground 🛝

Playground for all kinds of examples with NOS.

Installation

We assume the NOS server package and all its dependencies has been installed in prior. Check the quick start document from NOS docs first if you haven't setup the develop enviroment yet.

Quick Start

Simply run the following commands to serve the model you would like to deply:

cd examples/MODEL_ID
nos serve up -c serve.yaml

Then you can use the nos python client library to run the inference:

from PIL import Image
from nos.client import Client

client = Client("[::]:50051")

model_id = "YOUR-MODEL-ID"
models: List[str] = client.ListModels()
assert model_id in models 
# Check if the selected model has been served.

inputs = YOUR-MODEL-INPUT
response = model(inputs) # Get output as response.
#change to model.DEFAULT_METHOD_NAME if the default method is defined as  "__call__"

Available Examples

Chat completion

model_id: str = "meta-llama/Llama-2-7b-chat-hf"

Video transcription

model_id: str = "m-bain/whisperx-large-v2"

Text to image

model_id: List[str] = ["sd-xl-turbo", 
                        "playground-v2",
                        "latent-consistency-model"]

Text to video

model_id: str = "animate-diff"

Image to video

model_id: str = "stable-video-diffusion"

Text to 360-view images

model_id: str = "mv-dream"

Text to Speech

model_id: str = "bark"

Text to Music

model_id: str = "music-gen"

Reach US

• 💬 Send us an email at [email protected] or join our Discord for help.
• 📣 Follow us on Twitter, and LinkedIn to keep up-to-date on our products.