Build multimodal AI applications with cloud-native technologies

Jina lets you build multimodal AI services and pipelines that communicate via gRPC, HTTP and WebSockets, then scale them up and deploy to production. You can focus on your logic and algorithms, without worrying about the infrastructure complexity.

Jina provides a smooth Pythonic experience for serving ML models transitioning from local deployment to advanced orchestration frameworks like Docker-Compose, Kubernetes, or Jina AI Cloud. Jina makes advanced solution engineering and cloud-native technologies accessible to every developer.

• Build and serve models for any data type and any mainstream deep learning framework.
• Design high-performance services, with easy scaling, duplex client-server streaming, batching, dynamic batching, async/non-blocking data processing and any protocol.
• Serve LLM models while streaming their output.
• Docker container integration via Executor Hub, OpenTelemetry/Prometheus observability.
• Streamlined CPU/GPU hosting via Jina AI Cloud.
• Deploy to your own cloud or system with our Kubernetes and Docker Compose integration.

Wait, how is Jina different from FastAPI?

Jina's value proposition may seem quite similar to that of FastAPI. However, there are several fundamental differences:

Data structure and communication protocols

• FastAPI communication relies on Pydantic and Jina relies on DocArray allowing Jina to support multiple protocols to expose its services. The support for gRPC protocol is specially useful for data intensive applications as for embedding services where the embeddings and tensors can be more efficiently serialized.

Advanced orchestration and scaling capabilities

• Jina allows you to easily containerize and orchestrate your services and models, providing concurrency and scalability.
• Jina lets you deploy applications formed from multiple microservices that can be containerized and scaled independently.

Journey to the cloud

• Jina provides a smooth transition from local development (using DocArray) to local serving using Deployment and Flow to having production-ready services by using Kubernetes capacity to orchestrate the lifetime of containers.
• By using Jina AI Cloud you have access to scalable and serverless deployments of your applications in one command.

Documentation

Install

pip install jina

Find more install options on Apple Silicon/Windows.

Get Started

Basic Concepts

Jina has three fundamental layers:

• Data layer: BaseDoc and DocList (from DocArray) are the input/output formats in Jina.
• Serving layer: An Executor is a Python class that transforms and processes Documents. By simply wrapping your models into an Executor, you allow them to be served and scaled by Jina. Gateway is the service making sure connecting all Executors inside a Flow.
• Orchestration layer: Deployment serves a single Executor, while a Flow serves Executors chained into a pipeline.

The full glossary is explained here.

Serve AI models

Let's build a fast, reliable and scalable gRPC-based AI service. In Jina we call this an Executor. Our simple Executor will wrap the StableLM LLM from Stability AI. We'll then use a Deployment to serve it.

Note A Deployment serves just one Executor. To combine multiple Executors into a pipeline and serve that, use a Flow.

Let's implement the service's logic:

executor.py

from jina import Executor, requests from docarray import DocList, BaseDoc from transformers import pipeline class Prompt(BaseDoc): text: str class Generation(BaseDoc): prompt: str text: str class StableLM(Executor): def __init__(self, **kwargs): super().__init__(**kwargs) self.generator = pipeline( 'text-generation', model='stabilityai/stablelm-base-alpha-3b' ) @requests def generate(self, docs: DocList[Prompt], **kwargs) -> DocList[Generation]: generations = DocList[Generation]() prompts = docs.text llm_outputs = self.generator(prompts) for prompt, output in zip(prompts, llm_outputs): generations.append(Generation(prompt=prompt, text=output)) return generations

Then we deploy it with either the Python API or YAML:

Python API: deployment.py	YAML: deployment.yml
from jina import Deployment from executor import StableLM dep = Deployment(uses=StableLM, timeout_ready=-1, port=12345) with dep: dep.block()	jtype: Deployment with: uses: StableLM py_modules: - executor.py timeout_ready: -1 port: 12345 And run the YAML Deployment with the CLI: jina deployment --uses deployment.yml

Use Jina Client to make requests to the service:

from jina import Client
from docarray import DocList, BaseDoc


class Prompt(BaseDoc):
    text: str


class Generation(BaseDoc):
    prompt: str
    text: str


prompt = Prompt(
    text='suggest an interesting image generation prompt for a mona lisa variant'
)

client = Client(port=12345)  # use port from output above
response = client.post(on='/', inputs=[prompt], return_type=DocList[Generation])

print(response[0].text)

a steampunk version of the Mona Lisa, incorporating mechanical gears, brass elements, and Victorian era clothing details

Note In a notebook, you can't use deployment.block() and then make requests to the client. Please refer to the Colab link above for reproducible Jupyter Notebook code snippets.

Build a pipeline

Sometimes you want to chain microservices together into a pipeline. That's where a Flow comes in.

A Flow is a DAG pipeline, composed of a set of steps, It orchestrates a set of Executors and a Gateway to offer an end-to-end service.

Note If you just want to serve a single Executor, you can use a Deployment.

For instance, let's combine our StableLM language model with a Stable Diffusion image generation model. Chaining these services together into a Flow will give us a service that will generate images based on a prompt generated by the LLM.

text_to_image.py

import numpy as np from jina import Executor, requests from docarray import BaseDoc, DocList from docarray.documents import ImageDoc class Generation(BaseDoc): prompt: str text: str class TextToImage(Executor): def __init__(self, **kwargs): super().__init__(**kwargs) from diffusers import StableDiffusionPipeline import torch self.pipe = StableDiffusionPipeline.from_pretrained( "CompVis/stable-diffusion-v1-4", torch_dtype=torch.float16 ).to("cuda") @requests def generate_image(self, docs: DocList[Generation], **kwargs) -> DocList[ImageDoc]: result = DocList[ImageDoc]() images = self.pipe( docs.text ).images # image here is in [PIL format](https://pillow.readthedocs.io/en/stable/) result.tensor = np.array(images) return result

Build the Flow with either Python or YAML:

Python API: flow.py	YAML: flow.yml
from jina import Flow from executor import StableLM from text_to_image import TextToImage flow = ( Flow(port=12345) .add(uses=StableLM, timeout_ready=-1) .add(uses=TextToImage, timeout_ready=-1) ) with flow: flow.block()	jtype: Flow with: port: 12345 executors: - uses: StableLM timeout_ready: -1 py_modules: - executor.py - uses: TextToImage timeout_ready: -1 py_modules: - text_to_image.py Then run the YAML Flow with the CLI: jina flow --uses flow.yml

Then, use Jina Client to make requests to the Flow:

from jina import Client
from docarray import DocList, BaseDoc
from docarray.documents import ImageDoc


class Prompt(BaseDoc):
    text: str


prompt = Prompt(
    text='suggest an interesting image generation prompt for a mona lisa variant'
)

client = Client(port=12345)  # use port from output above
response = client.post(on='/', inputs=[prompt], return_type=DocList[ImageDoc])

response[0].display()

Easy scalability and concurrency

Why not just use standard Python to build that service and pipeline? Jina accelerates time to market of your application by making it more scalable and cloud-native. Jina also handles the infrastructure complexity in production and other Day-2 operations so that you can focus on the data application itself.

Increase your application's throughput with scalability features out of the box, like replicas, shards and dynamic batching.

Let's scale a Stable Diffusion Executor deployment with replicas and dynamic batching:

• Create two replicas, with a GPU assigned for each.
• Enable dynamic batching to process incoming parallel requests together with the same model inference.

Normal Deployment	Scaled Deployment
jtype: Deployment with: uses: TextToImage timeout_ready: -1 py_modules: - text_to_image.py	jtype: Deployment with: uses: TextToImage timeout_ready: -1 py_modules: - text_to_image.py env: CUDA_VISIBLE_DEVICES: RR replicas: 2 uses_dynamic_batching: # configure dynamic batching /default: preferred_batch_size: 10 timeout: 200

Assuming your machine has two GPUs, using the scaled deployment YAML will give better throughput compared to the normal deployment.

These features apply to both Deployment YAML and Flow YAML. Thanks to the YAML syntax, you can inject deployment configurations regardless of Executor code.

Deploy to the cloud

Containerize your Executor

In order to deploy your solutions to the cloud, you need to containerize your services. Jina provides the Executor Hub, the perfect tool to streamline this process taking a lot of the troubles with you. It also lets you share these Executors publicly or privately.

You just need to structure your Executor in a folder:

TextToImage/
├── executor.py
├── config.yml
├── requirements.txt

config.yml	requirements.txt
jtype: TextToImage py_modules: - executor.py metas: name: TextToImage description: Text to Image generation Executor based on StableDiffusion url: keywords: []	diffusers accelerate transformers

Then push the Executor to the Hub by doing: jina hub push TextToImage.

This will give you a URL that you can use in your Deployment and Flow to use the pushed Executors containers.

jtype: Flow
with:
    port: 12345
executors:
  - uses: jinai+docker://<user-id>/StableLM
  - uses: jinai+docker://<user-id>/TextToImage

Get on the fast lane to cloud-native

Using Kubernetes with Jina is easy:

jina export kubernetes flow.yml ./my-k8s
kubectl apply -R -f my-k8s

And so is Docker Compose:

jina export docker-compose flow.yml docker-compose.yml
docker-compose up

Note You can also export Deployment YAML to Kubernetes and Docker Compose.

That's not all. We also support OpenTelemetry, Prometheus, and Jaeger.

What cloud-native technology is still challenging to you? Tell us and we'll handle the complexity and make it easy for you.

Deploy to JCloud

You can also deploy a Flow to JCloud, where you can easily enjoy autoscaling, monitoring and more with a single command.

First, turn the flow.yml file into a JCloud-compatible YAML by specifying resource requirements and using containerized Hub Executors.

Then, use jina cloud deploy command to deploy to the cloud:

wget https://raw.githubusercontent.com/jina-ai/jina/master/.github/getting-started/jcloud-flow.yml
jina cloud deploy jcloud-flow.yml

Warning

Make sure to delete/clean up the Flow once you are done with this tutorial to save resources and credits.

Read more about deploying Flows to JCloud.

Streaming for LLMs

Large Language Models can power a wide range of applications from chatbots to assistants and intelligent systems. However, these models can be heavy and slow and your users want systems that are both intelligent and fast!

Large language models work by turning your questions into tokens and then generating new token one at a time until it decides that generation should stop. This means you want to stream the output tokens generated by a large language model to the client. In this tutorial, we will discuss how to achieve this with Streaming Endpoints in Jina.

Service Schemas

The first step is to define the streaming service schemas, as you would do in any other service framework. The input to the service is the prompt and the maximum number of tokens to generate, while the output is simply the token ID:

from docarray import BaseDoc


class PromptDocument(BaseDoc):
    prompt: str
    max_tokens: int


class ModelOutputDocument(BaseDoc):
    token_id: int
    generated_text: str

Service initialization

Our service depends on a large language model. As an example, we will use the gpt2 model. This is how you would load such a model in your executor

from jina import Executor, requests
from transformers import GPT2Tokenizer, GPT2LMHeadModel
import torch

tokenizer = GPT2Tokenizer.from_pretrained('gpt2')


class TokenStreamingExecutor(Executor):
    def __init__(self, **kwargs):
        super().__init__(**kwargs)
        self.model = GPT2LMHeadModel.from_pretrained('gpt2')

Implement the streaming endpoint

Our streaming endpoint accepts a PromptDocument as input and streams ModelOutputDocuments. To stream a document back to the client, use the yield keyword in the endpoint implementation. Therefore, we use the model to generate up to max_tokens tokens and yield them until the generation stops:

class TokenStreamingExecutor(Executor):
    ...

    @requests(on='/stream')
    async def task(self, doc: PromptDocument, **kwargs) -> ModelOutputDocument:
        input = tokenizer(doc.prompt, return_tensors='pt')
        input_len = input['input_ids'].shape[1]
        for _ in range(doc.max_tokens):
            output = self.model.generate(**input, max_new_tokens=1)
            if output[0][-1] == tokenizer.eos_token_id:
                break
            yield ModelOutputDocument(
                token_id=output[0][-1],
                generated_text=tokenizer.decode(
                    output[0][input_len:], skip_special_tokens=True
                ),
            )
            input = {
                'input_ids': output,
                'attention_mask': torch.ones(1, len(output[0])),
            }

Learn more about streaming endpoints from the Executor documentation.

Serve and send requests

The final step is to serve the Executor and send requests using the client. To serve the Executor using gRPC:

from jina import Deployment

with Deployment(uses=TokenStreamingExecutor, port=12345, protocol='grpc') as dep:
    dep.block()

To send requests from a client:

import asyncio
from jina import Client


async def main():
    client = Client(port=12345, protocol='grpc', asyncio=True)
    async for doc in client.stream_doc(
        on='/stream',
        inputs=PromptDocument(prompt='what is the capital of France ?', max_tokens=10),
        return_type=ModelOutputDocument,
    ):
        print(doc.generated_text)


asyncio.run(main())

The
The capital
The capital of
The capital of France
The capital of France is
The capital of France is Paris
The capital of France is Paris.

Support

• Join our Discord community and chat with other community members about ideas.
• Subscribe to the latest video tutorials on our YouTube channel

Join Us

Jina is backed by Jina AI and licensed under Apache-2.0.