File size: 4,587 Bytes
443c139
 
cda261e
52ac0d1
 
 
 
 
 
 
 
 
cda261e
 
52ac0d1
 
 
 
 
 
 
 
 
 
 
 
 
5d9c9c6
c327aef
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
acc345c
c327aef
 
 
 
 
 
 
 
 
 
443c139
52ac0d1
 
 
 
443c139
 
 
 
52ac0d1
 
 
 
 
 
 
 
 
c327aef
52ac0d1
 
 
 
 
 
bce2b48
b6a8d3b
 
52ac0d1
 
 
 
 
 
 
b6a8d3b
52ac0d1
 
 
 
d8d70b6
 
c327aef
d8d70b6
49c6131
d8d70b6
49c6131
 
 
c327aef
17b9a96
49c6131
e3cd3bd
49c6131
 
1508448
c327aef
1508448
c327aef
49c6131
 
 
 
52ac0d1
 
 
 
443c139
52ac0d1
 
 
 
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
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
# credits : https://huggingface.co/spaces/black-forest-labs/FLUX.1-dev

import os
import gradio as gr
import numpy as np
import random
import spaces
import torch
from diffusers import  DiffusionPipeline, FlowMatchEulerDiscreteScheduler, AutoencoderTiny, AutoencoderKL
from transformers import CLIPTextModel, CLIPTokenizer,T5EncoderModel, T5TokenizerFast
from live_preview_helpers import calculate_shift, retrieve_timesteps, flux_pipe_call_that_returns_an_iterable_of_images

hf_token = os.getenv("HF_TOKEN")

dtype = torch.bfloat16
device = "cuda" if torch.cuda.is_available() else "cpu"

taef1 = AutoencoderTiny.from_pretrained("madebyollin/taef1", torch_dtype=dtype).to(device)
good_vae = AutoencoderKL.from_pretrained("black-forest-labs/FLUX.1-dev", subfolder="vae", torch_dtype=dtype).to(device)
pipe = DiffusionPipeline.from_pretrained("black-forest-labs/FLUX.1-dev", torch_dtype=dtype, vae=taef1).to(device)
torch.cuda.empty_cache()

MAX_SEED = np.iinfo(np.int32).max

pipe.flux_pipe_call_that_returns_an_iterable_of_images = flux_pipe_call_that_returns_an_iterable_of_images.__get__(pipe)

@spaces.GPU(duration=75)

def infer(name, pet, background, style, seed=42, randomize_seed=False, width=1024, height=1024, guidance_scale=3.5, num_inference_steps=28, progress=gr.Progress(track_tqdm=True)):
    if pet == "Kaatz":
        intro = "please generate an image of a cat sitting "
    elif pet == "Mupp":
        intro = "please generate an image of a dog sitting "
    elif pet == "Hues":
        intro = "please generate an image of a bunny sitting "
    else:
        intro = "please generate an image of an hamster sitting "

    if background == "Wunnzëmmer":
        place = intro + "in a living space "
    elif background == "Grafitti Mauer":  
        place = intro + "in front of a wall with graffiti "
    elif background == "Strooss":   
        place = intro + "in a street in the city "
    elif background == "Plage":
        place = intro + "at the beach "
    else:
        place = intro + " in the forest "

    if style == "Photo":
        prompt = place + "holding a signal that says " + name + "in a photorealistic style"
    elif style == "Cartoon":
        prompt = place + "holding a signal that says " + name + "in a cartoon style"
    elif style == "Woll":
        prompt = place + "holding a signal that says " + name + "in a knitted with wool style"
    elif style == "Aquarell":
        prompt = place + "holding a signal that says " + name + "in a watercolorl style"
    else:
        prompt = place + "holding a signal that says " + name + "in a 3D style"
        
    seed = random.randint(0, MAX_SEED)
    generator = torch.Generator().manual_seed(seed)
    
    for img in pipe.flux_pipe_call_that_returns_an_iterable_of_images(
            prompt=prompt,
            guidance_scale=3,5,
            num_inference_steps=28,
            width=1024,
            height=1024,
            generator=generator,
            output_type="pil",
            good_vae=good_vae,
        ):
            yield img, seed

css="""
#col-container {
    margin: 0 auto;
    max-width: 640px;
}
"""

with gr.Blocks(css=css) as demo:
    
    with gr.Column(elem_id="col-container"):
        gr.Markdown(f"""# Mäin éischt KI-Bild
        Mol mer e Bild mat méngem Hausdéier a méngem Numm op engem Schëld !
        """) 
        
        with gr.Row():
            
            prompt = gr.Text(
                label="Prompt",
                show_label=False,
                max_lines=1,
                placeholder="Schreiw däin Text mat dengem  Numm ",
                container=False,
            )
            
            run_button = gr.Button("Run", scale=0)

        with gr.Row():
            pet = gr.Radio(
                choices=["Kaatz", "Mupp", "Hues", "Hamster"],
                label="Hausdéier",
                value="Kaatz"
            ) 

        with gr.Row():
            background = gr.Radio(
                choices=["Wunnzëmmer", "Grafitti Mauer", "Strooss", "Plage", "Bësch"],
                label="Hannergronn",
                value="Strooss"
            ) 

        with gr.Row():
            style = gr.Radio(
                choices=["Photo", "Cartoon", "Woll", "Aquarell", "3D"],
                label="Style",
                value="Photo"
            ) 
           
        result = gr.Image(label="Result", show_label=False) 

    gr.on(
        triggers=[run_button.click, prompt.submit],
        fn = infer,
        inputs = [prompt, pet, background, style],
        outputs = [result, seed]
    )

demo.launch()