Examples

Nano-banana on CPU

from typing import List, Optional

from pydantic import BaseModel

from fma.toolkit import model as fma_model
from fma.toolkit.fields.image import Image


class Input(BaseModel):
    prompt: str
    image: Optional[Image]
    image1: Optional[Image]


class Output(BaseModel):
    sample: Image


class Model(fma_model.Model):
    requirements: List[str] = ["git+https://<token>@github.com/FlyMyAI/flymy_nano_banana.git"]

    def initialize(self):
        import os

        os.environ['GOOGLE_API_KEY'] = ''
        os.environ['MODEL'] = 'gemini-2.5-flash-image-preview'

        from flymy_nano_banana.image_generator import GoogleImageGenerator

        self.image_generator = GoogleImageGenerator()

    def predict(self, input: Input) -> Output:
        prompt = input["prompt"]
        image = input["image"]
        image1 = input["image1"]

        generated_image = self.image_generator.generate_image(prompt, image, image1)  # type: ignore
        if generated_image is None:
            raise ValueError("No image was generated")

        return {"sample": generated_image}

LoRA Training

This example demonstrates how to train a LoRA model using the FlyMyAI platform. The training process is asynchronous, so you'll need to poll for results.

import os

try:
    import flymyai
except ModuleNotFoundError:
    os.system('pip install flymyai')
    import flymyai

payload = {
    "lora_name": "random-lora",
    "trigger_word": "random",
    "dataset_url": "",  # dataset might be an zip archive with jpeg photos in it 
    "training_steps": 1000, 
}

fma_client = flymyai.client(apikey=os.getenv("FLYMYAI_API_KEY", ""))

prediction_task = fma_client.predict_async_task(
    model="flymyai/flux-lora-trainer-fast",
    payload=payload,
)
print(f'Got prediction task: {prediction_task}')

while True:
    try:
        # Get result
        result = fma_client.prediction_task_result(prediction_task)
        break
    except:
        continue

print(f"Prediction completed: {result.inference_responses}")
print(result.inference_responses[0].output_data['model'][0]['data'])
print(result.inference_responses[0].output_data['config'][0]['data'])

Parameters

• lora_name: Name for your LoRA model
• trigger_word: Word to activate the LoRA model in prompts
• dataset_url: Direct URL to your training dataset. The dataset should be a ZIP archive containing JPEG images
• training_steps: Number of training steps (typically 1000-2000)

Notes

• The training process is asynchronous and may take several minutes to complete
• Make sure your FLYMYAI_API_KEY environment variable is set
• The dataset should be a ZIP archive containing JPEG images for training