File size: 1,908 Bytes
07d5247
058e9d8
 
6d70521
058e9d8
01807fb
07d5247
058e9d8
54a9b28
01807fb
 
058e9d8
07d5247
0c639c6
01807fb
 
0c639c6
e7f7ab4
01807fb
058e9d8
 
f2ae260
670d739
abc3bb5
934504b
26d38a8
 
058e9d8
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
import gradio as gr
import torch
import numpy as np
import modin.pandas as pd
from PIL import Image
from diffusers import DiffusionPipeline, StableDiffusionLatentUpscalePipeline

device = "cuda" if torch.cuda.is_available() else "cpu"
pipe = DiffusionPipeline.from_pretrained("dreamlike-art/dreamlike-photoreal-2.0", torch_dtype=torch.float16, safety_checker=None)
upscaler = StableDiffusionLatentUpscalePipeline.from_pretrained("stabilityai/sd-x2-latent-upscaler", torch_dtype=torch.float16)
upscaler = upscaler.to(device)
pipe = pipe.to(device)

def genie (Prompt, scale, steps, Seed):
    generator = torch.Generator(device=device).manual_seed(Seed)
     #images = pipe(prompt, num_inference_steps=steps, guidance_scale=scale, generator=generator).images[0]
    low_res_latents = pipe(Prompt, num_inference_steps=steps, guidance_scale=scale, generator=generator, output_type="latent").images
    upscaled_image = upscaler(prompt='', image=low_res_latents, num_inference_steps=5, guidance_scale=0, generator=generator).images[0]
    return upscaled_image
    
gr.Interface(fn=genie, inputs=[gr.Textbox(label='What you want the AI to generate. 77 Token Limit.'), 
                               gr.Slider(1, maximum=15, value=10, step=.25, label='Prompt Guidance Scale:', interactive=True), 
                               gr.Slider(1, maximum=100, value=50, step=1, label='Number of Iterations: 50 is typically fine.'), 
                               gr.Slider(minimum=1, step=10, maximum=999999999999999999, randomize=True, interactive=True)], 
             outputs=gr.Image(label='512x512 Generated Image'), 
             title="PhotoReal V2 with SD x2 Upscaler - GPU", 
             description="<br><br><b/>Warning: This Demo is capable of producing NSFW content.", 
             article = "Code Monkey: <a href=\"https://huggingface.co/Manjushri\">Manjushri</a>").launch(debug=True, max_threads=True)