Deploy the Gemma-3-27B-it using Inferless

Introduction

Gemma-3-27B-it is a state‑of‑the‑art, 27B vision‑language model from the Gemma team. Built for instruction‑tuned applications, it seamlessly integrates robust language understanding with cutting‑edge visual analysis. Whether it’s detecting intricate visual patterns, parsing complex documents, or analyzing extended video content by highlighting key moments, this model is engineered to act as a versatile visual agent. With capabilities that include generating structured outputs (such as bounding boxes and JSON) and supporting multilingual text within images, it paves the way for innovative interactive chatbots, advanced multimedia content analysis, and beyond.

Defining Dependencies

We are using the transformers to serve the model on a single A100.

Our Observations

We have deployed the model on an A100 GPU(80GB). Here are our observations:

Library	Inference Time	Cold Start Time	Tokens/Sec	Output Tokens Length
transformers	11.73 sec	21.64 sec	10.59	128

Note: The inference time and cold start time are average values.

Defining Dependencies

We are using the transformers to serve the model on a single A100 (80GB).

Constructing the GitHub/GitLab Template

Now quickly construct the GitHub/GitLab template, this process is mandatory and make sure you don’t add any file named model.py.

gemma-3-27b-it/
├── app.py
├── inferless-runtime-config.yaml
└── inferless.yaml

You can also add other files to this directory.

Create the Input Schema with Pydantic

Using the inferless Python client and Pydantic, you can define structured schemas directly in your code for input and output, eliminating the need for external file.

Input Schema

When defining an input schema with Pydantic, you need to annotate your class attributes with the appropriate types, such as str, float, int, etc. These type annotations specifys what type of data each field should contain. The default value serves as the example input for testing with the infer function.

@inferless.request
class RequestObjects(BaseModel):
    prompt: str = Field(default="What is Deep Learning?")
    image_url: Optional[str] = None
    system_prompt: Optional[str] = "You are a helpful assistant."
    temperature: Optional[float] = 0.7
    top_p: Optional[float] = 0.1
    repetition_penalty: Optional[float] = 1.18
    top_k: Optional[int] = 40
    max_tokens: Optional[int] = 100
    do_sample: Optional[bool] = False

Output Schema

The @inferless.response decorator helps you define structured output schemas.

@inferless.response
class ResponseObjects(BaseModel):
    generated_text: str = Field(default="Test output")

Usage in the `infer` Function

Once you have annotated the objects you can expect the infer function to receive RequestObjects as input, and returns a ResponseObjects instance as output, ensuring the results adhere to a defined structure.

class InferlessPythonModel:
    def infer(self, request: RequestObjects) -> ResponseObjects:
        
        return ResponseObjects(generated_result = generated_text)

Create the class for inference

In the app.py we will define the class and import all the required functions

def initialize: In this function, you will initialize your model and define any variable that you want to use during inference.
def infer: This function gets called for every request that you send. Here you can define all the steps that are required for the inference.
def finalize: This function cleans up all the allocated memory.

import os
os.environ["HF_HUB_ENABLE_HF_TRANSFER"] = "1"
from huggingface_hub import snapshot_download
from pydantic import BaseModel, Field
from typing import Optional
import inferless
import torch
from transformers import AutoProcessor, Gemma3ForConditionalGeneration

@inferless.request
class RequestObjects(BaseModel):
    prompt: str = Field(default="What is Deep Learning?")
    image_url: Optional[str] = None
    system_prompt: Optional[str] = "You are a helpful assistant."
    temperature: Optional[float] = 0.7
    top_p: Optional[float] = 0.1
    repetition_penalty: Optional[float] = 1.18
    top_k: Optional[int] = 40
    max_tokens: Optional[int] = 100
    do_sample: Optional[bool] = False

@inferless.response
class ResponseObjects(BaseModel):
    generated_text: str = Field(default="Test output")

class InferlessPythonModel:
    def initialize(self):
        model_id = "google/gemma-3-27b-it"
        snapshot_download(repo_id=model_id, allow_patterns=["*.safetensors"])
        self.model = Gemma3ForConditionalGeneration.from_pretrained(
            model_id, device_map="cuda"
        ).eval()
        self.processor = AutoProcessor.from_pretrained(model_id)

    def infer(self, request: RequestObjects) -> ResponseObjects:
        messages = [
            {
                "role": "system",
                "content": [{"type": "text", "text": request.system_prompt}]
            }
        ]
        # Build the user message based on provided inputs.
        user_content = []
        if request.image_url is not None:
            user_content.append({"type": "image", "image": request.image_url})
        
        user_content.append({"type": "text", "text": request.prompt})
        messages.append({
            "role": "user",
            "content": user_content
        })

        inputs = self.processor.apply_chat_template(
            messages,
            add_generation_prompt=True,
            tokenize=True,
            return_dict=True,
            return_tensors="pt"
        ).to(self.model.device, dtype=torch.bfloat16)

        input_len = inputs["input_ids"].shape[-1]

        with torch.inference_mode():
            generation = self.model.generate(
                **inputs,
                max_new_tokens=request.max_tokens,
                temperature=request.temperature,
                top_p=request.top_p,
                top_k=request.top_k,
                do_sample=request.do_sample,
                repetition_penalty=request.repetition_penalty,
            )
            generation = generation[0][input_len:]

        decoded = self.processor.decode(generation, skip_special_tokens=True)
        return ResponseObjects(generated_text=decoded)

    def finalize(self):
        self.model = None

Creating the Custom Runtime

This is a mandatory step where we allow the users to upload their custom runtime through inferless-runtime-config.yaml.

build:
  cuda_version: "12.1.1"
  python_packages:
    - "git+https://github.com/huggingface/transformers@v4.49.0-Gemma-3"
    - "hf-transfer==0.1.9"
    - "inferless==0.2.13"
    - "pydantic==2.10.2"
    - "accelerate==1.5.2"
    - "pillow==11.1.0"

Test your model with Remote Run

You can use the inferless remote-run(installation guide here) command to test your model or any custom Python script in a remote GPU environment directly from your local machine. Make sure that you use Python3.10 for seamless experience.

Step 1: Add the Decorators and local entry point

To enable Remote Run, simply do the following:

Import the inferless library and initialize Cls(gpu="A100"). The available GPU options are T4, A10 and A100.
Decorated the initialize and infer functions with @app.load and @app.infer respectively.
Create the Local Entry Point by decorating a function (for example, my_local_entry) with @inferless.local_entry_point. Within this function, instantiate your model class, convert any incoming parameters into a RequestObjects object, and invoke the model’s infer method.

import os
os.environ["HF_HUB_ENABLE_HF_TRANSFER"] = "1"
from huggingface_hub import snapshot_download
from pydantic import BaseModel, Field
from typing import Optional
import inferless
import torch
from transformers import AutoProcessor, Gemma3ForConditionalGeneration

app = inferless.Cls(gpu="A100")

@inferless.request
class RequestObjects(BaseModel):
    prompt: str = Field(default="What is Deep Learning?")
    image_url: Optional[str] = None
    system_prompt: Optional[str] = "You are a helpful assistant."
    temperature: Optional[float] = 0.7
    top_p: Optional[float] = 0.1
    repetition_penalty: Optional[float] = 1.18
    top_k: Optional[int] = 40
    max_tokens: Optional[int] = 100
    do_sample: Optional[bool] = False

@inferless.response
class ResponseObjects(BaseModel):
    generated_text: str = Field(default="Test output")

class InferlessPythonModel:
    @app.load
    def initialize(self):
        model_id = "google/gemma-3-27b-it"
        snapshot_download(repo_id=model_id, allow_patterns=["*.safetensors"])
        self.model = Gemma3ForConditionalGeneration.from_pretrained(
            model_id, device_map="cuda"
        ).eval()
        self.processor = AutoProcessor.from_pretrained(model_id)
    
    @app.infer
    def infer(self, request: RequestObjects) -> ResponseObjects:
        messages = [
            {
                "role": "system",
                "content": [{"type": "text", "text": request.system_prompt}]
            }
        ]
        # Build the user message based on provided inputs.
        user_content = []
        if request.image_url is not None:
            user_content.append({"type": "image", "image": request.image_url})
        
        user_content.append({"type": "text", "text": request.prompt})
        messages.append({
            "role": "user",
            "content": user_content
        })

        inputs = self.processor.apply_chat_template(
            messages,
            add_generation_prompt=True,
            tokenize=True,
            return_dict=True,
            return_tensors="pt"
        ).to(self.model.device, dtype=torch.bfloat16)

        input_len = inputs["input_ids"].shape[-1]

        with torch.inference_mode():
            generation = self.model.generate(
                **inputs,
                max_new_tokens=request.max_tokens,
                temperature=request.temperature,
                top_p=request.top_p,
                top_k=request.top_k,
                do_sample=request.do_sample,
                repetition_penalty=request.repetition_penalty,
            )
            generation = generation[0][input_len:]

        decoded = self.processor.decode(generation, skip_special_tokens=True)
        return ResponseObjects(generated_text=decoded)

    def finalize(self):
        self.model = None

@inferless.local_entry_point
def my_local_entry(dynamic_params):
    request_objects = RequestObjects(**dynamic_params)
    model_instance = InferlessPythonModel()

    return model_instance.infer(request_objects)

Step 2: Run with Remote GPU

From your local terminal, navigate to the folder containing your app.py and your inferless-runtime-config.yaml and run:

inferless remote-run app.py -c inferless-runtime-config.yaml --prompt "What does this diagram illustrate?" --image_url "https://qianwen-res.oss-cn-beijing.aliyuncs.com/Qwen-VL/assets/demo.jpeg"

You can pass the other input parameters in the same way (e.g., --temperature, --system_prompt, etc.) as long as your code expects them in the inputs dictionary. If you want to exclude certain files or directories from being uploaded, use the --exclude or -e flag.

Method A: Deploying the model on Inferless Platform

Inferless supports multiple ways of importing your model. For this tutorial, we will use GitHub. Navigate to your desired workspace in Inferless and Click on Add a custom model button that you see on the top right. An import wizard will open up.

Step 2: Follow the UI to complete the model Import

Select the GitHub/GitLab Integration option to connect your source code repository with the deployment environment.
Navigate to the specific GitHub repository that contains your model’s code. Here, you will need to identify and enter the name of the model you wish to import.
Choose the appropriate type of machine that suits your model’s requirements. Additionally, specify the minimum and maximum number of replicas to define the scalability range for deploying your model.
Optionally, you have the option to enable automatic build and deployment. This feature triggers a new deployment automatically whenever there is a new code push to your repository.
If your model requires additional software packages, configure the Custom Runtime settings by including necessary pip or apt packages. Also, set up environment variables such as Inference Timeout, Container Concurrency, and Scale Down Timeout to tailor the runtime environment according to your needs.
Wait for the validation process to complete, ensuring that all settings are correct and functional. Once validation is successful, click on the “Import” button to finalize the import of your model.

Step 3: Wait for the model build to complete usually takes ~5-10 minutes

Step 4: Use the APIs to call the model

Once the model is in ‘Active’ status you can click on the ‘API’ page to call the model

Here is the Demo:

Method B: Deploying the model on Inferless CLI

Inferless allows you to deploy your model using Inferless-CLI. Follow the steps to deploy using Inferless CLI.

Clone the repository of the model

Let’s begin by cloning the model repository:

git clone https://github.com/inferless/gemma-3-27b-it.git

Deploy the Model

To deploy the model using Inferless CLI, execute the following command:

inferless deploy --gpu A100 --runtime inferless-runtime-config.yaml

Explanation of the Command:

--gpu A100: Specifies the GPU type for deployment. Available options include A10, A100, and T4.
--runtime inferless-runtime-config.yaml: Defines the runtime configuration file. If not specified, the default Inferless runtime is used.

How-to Guides

​Introduction

​Defining Dependencies

​Our Observations

​Defining Dependencies

​Constructing the GitHub/GitLab Template

​Create the Input Schema with Pydantic

​Input Schema

​Output Schema

​Usage in the infer Function

​Create the class for inference

​Creating the Custom Runtime

​Test your model with Remote Run

​Step 1: Add the Decorators and local entry point

​Step 2: Run with Remote GPU

​Method A: Deploying the model on Inferless Platform

​Step 1: Login to the inferless dashboard can click on Import model button

​Step 2: Follow the UI to complete the model Import

​Step 3: Wait for the model build to complete usually takes ~5-10 minutes

​Step 4: Use the APIs to call the model

​Here is the Demo:

​Method B: Deploying the model on Inferless CLI

​Clone the repository of the model

​Deploy the Model

Introduction

Defining Dependencies

Our Observations

Defining Dependencies

Constructing the GitHub/GitLab Template

Create the Input Schema with Pydantic

Input Schema

Output Schema

Usage in the `infer` Function

Create the class for inference

Creating the Custom Runtime

Test your model with Remote Run

Step 1: Add the Decorators and local entry point

Step 2: Run with Remote GPU

Method A: Deploying the model on Inferless Platform

Step 1: Login to the inferless dashboard can click on Import model button

Step 2: Follow the UI to complete the model Import

Step 3: Wait for the model build to complete usually takes ~5-10 minutes

Step 4: Use the APIs to call the model

Here is the Demo:

Method B: Deploying the model on Inferless CLI

Clone the repository of the model

Deploy the Model