app.py

import gradio as gr
import torch
from torch import autocast
from kandinsky2 import get_kandinsky2

device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
model = get_kandinsky2('cuda', task_type='text2img', cache_dir='/tmp/kandinsky2', model_version='2.1', use_flash_attention=False)

def generate_text(prompt, quality="High (Default)"):
    length_dict = {"Low": 50, "High (Default)": 100, "Ultra": 150}
    length = length_dict[quality]
    return model.generate_text2img('''red cat, 4k photo''', num_steps=length,
                          batch_size=1, guidance_scale=4,
                           h=768, w=768
                           ,sampler='p_sampler', prior_cf_scale=4,
                            prior_steps="5",)[0]

iface = gr.Interface(
  fn=generate_text, 
  inputs=["textbox", gr.inputs.Dropdown(["Low", "Medium (Default)", "High"], label="Quality")],
  outputs=gr.outputs.Image(label="Generated image:")
)

if device.type == 'cpu':
    model.load_state_dict(torch.load('path/to/model.pth', map_location=device))
else:
    model.load_state_dict(torch.load('path/to/model.pth'))

iface.launch()