DEV Community: AnyISalIn

Train Stylized Portraits on Novita.AI

AnyISalIn — Wed, 08 Nov 2023 12:08:14 +0000

Try this on the Hugging Face Spaces. https://huggingface.co/spaces/novita-ai/Face-Stylization-Playground

Obtain the Novita.AI API Key from the website https://novita.ai/.

The cost for LoRA training is approximately $0.2.

We recommend providing 6–10 high-quality self-portrait photos, with each photo requiring 200 steps and a maximum of 2000 steps. For instance, if you provide 7 images, it will take a total of 1400 steps.

Replace any object in a photograph with Stable Diffusion Python SDK.

AnyISalIn — Mon, 23 Oct 2023 09:22:40 +0000

You can install the Novita Python SDK using just one command

pip install novita-client

get NOVITA_API_KEY from https://novita.ai/get-started/

Finally, you can try the change cat to dog by following this code.

import os

from novita_client import NovitaClient, Txt2ImgRequest, UpscaleRequest
from novita_client.utils import base64_to_image, image_to_base64
from PIL import Image
from io import BytesIO

novita = NovitaClient(os.getenv('NOVITA_API_KEY'))

replace_object_res = novita.replace_object(image="https://cdn-images-1.medium.com/max/1600/1*E3jkiGoC8EfVlqqhE_6SWw.png", object_prompt="a cat", prompt="a dog", response_image_type="png")
replace_object_generated = base64_to_image(replace_object_res.image_file)
replace_object_generated.save("replace-object.png")

novita-client is the Python SDK of Novita.AI, which provides the following features.

cleanup - remove all your generated images
remove-background - remove background from image
remove-text - remove text from image
reimagine - reimagine image
doodle - doodle image
merge-face - merge face
mix-pose - mix pose
outpainting - outpainting
replace-object - remove object
replace-background - replace background
replace-sky - replace sky
create-tile - create tile

5 lines of Python code to change the posture of photography

AnyISalIn — Mon, 23 Oct 2023 04:01:38 +0000

You can install the Novita Python SDK using just one command

pip install novita-client

then you can get NOVITA_API_KEY from https://novita.ai/get-started/

Finally, you can try the mix pose demo by following this code.

import os

from novita_client import NovitaClient
from novita_client.utils import base64_to_image

client = NovitaClient(os.getenv('NOVITA_API_KEY'))
res = client.mixpose(
    image="https://image.uniqlo.com/UQ/ST3/my/imagesgoods/455359/item/mygoods_23_455359.jpg?width=494",
    pose_image="https://image.uniqlo.com/UQ/ST3/ca/imagesgoods/455359/item/cagoods_02_455359.jpg?width=494",
)

base64_to_image(res.image_file).save("./mixpose.png")

novita-client is the Python SDK of Novita.AI, which provides the following features.

cleanup - remove all your generated images
remove-background - remove background from image
remove-text - remove text from image
reimagine - reimagine image
doodle - doodle image
merge-face - merge face
mix-pose - mix pose
outpainting - outpainting
replace-object - remove object
replace-background - replace background
replace-sky - replace sky
create-tile - create tile