File size: 8,475 Bytes
dc155d4
 
 
 
 
 
 
97ff6b8
dc155d4
 
 
 
 
 
 
048bf77
dc155d4
 
 
048bf77
dc155d4
a0e2bdf
 
dc155d4
 
048bf77
6687fea
048bf77
 
fe70d6a
d4fdc43
a40e3c4
 
 
 
 
 
 
 
 
 
fe70d6a
dc155d4
 
349c4aa
dc155d4
048bf77
 
 
 
 
736f1ae
dc155d4
 
 
 
 
 
 
 
165d45c
 
dc155d4
 
7d594f1
 
 
 
 
 
 
 
1278c72
 
7d594f1
 
5e477f6
dc155d4
5e477f6
a0e2bdf
dc155d4
 
a0e2bdf
 
 
dc155d4
 
1278c72
 
dc155d4
 
 
6687fea
dc155d4
6687fea
 
dc155d4
 
6687fea
dc155d4
 
 
 
048bf77
 
dc155d4
 
 
 
 
 
 
 
 
 
6687fea
dc155d4
 
 
 
 
 
 
6687fea
dc155d4
 
a0e2bdf
dc155d4
 
a0e2bdf
dc155d4
 
1278c72
 
a0e2bdf
dc155d4
048bf77
dc155d4
a0e2bdf
dc155d4
a0e2bdf
dc155d4
a0e2bdf
5fe028b
 
 
 
 
 
 
 
 
 
dffc412
 
 
 
 
 
 
 
 
 
 
2afa171
 
 
6ad1aae
dffc412
 
 
023b880
f4f5c62
ff5aca4
f4f5c62
 
 
 
 
 
 
ff5aca4
 
023b880
a0e2bdf
 
 
1278c72
 
dc155d4
a0e2bdf
dc155d4
ff5aca4
dc155d4
 
 
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
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
# PyTorch 2.8 (temporary hack)
import os
os.system('pip install --upgrade --pre --extra-index-url https://download.pytorch.org/whl/nightly/cu126 "torch<2.9" spaces')

# Actual demo code
import spaces
import torch
from diffusers import WanPipeline, AutoencoderKLWan, UniPCMultistepScheduler
from diffusers.models.transformers.transformer_wan import WanTransformer3DModel
from diffusers.utils.export_utils import export_to_video
import gradio as gr
import tempfile
import numpy as np
from PIL import Image
import random
import gc
from optimization import optimize_pipeline_


MODEL_ID = "Wan-AI/Wan2.2-T2V-A14B-Diffusers"

LANDSCAPE_WIDTH = 1024
LANDSCAPE_HEIGHT = 1024
MAX_SEED = np.iinfo(np.int32).max

FIXED_FPS = 16
MIN_FRAMES_MODEL = 1
MAX_FRAMES_MODEL = 81

vae = AutoencoderKLWan.from_pretrained("Wan-AI/Wan2.2-T2V-A14B-Diffusers", subfolder="vae", torch_dtype=torch.float32)
pipe = WanPipeline.from_pretrained(MODEL_ID,
    transformer=WanTransformer3DModel.from_pretrained('linoyts/Wan2.2-T2V-A14B-Diffusers-BF16',
        subfolder='transformer',
        torch_dtype=torch.bfloat16,
        device_map='cuda',
    ),
    transformer_2=WanTransformer3DModel.from_pretrained('linoyts/Wan2.2-T2V-A14B-Diffusers-BF16',
        subfolder='transformer_2',
        torch_dtype=torch.bfloat16,
        device_map='cuda',
    ),
    vae=vae,
    torch_dtype=torch.bfloat16,
).to('cuda')
pipe.scheduler = UniPCMultistepScheduler.from_config(pipe.scheduler.config, flow_shift=2.0)


for i in range(3): 
    gc.collect()
    torch.cuda.synchronize() 
    torch.cuda.empty_cache()

optimize_pipeline_(pipe,
    prompt='prompt',
    height=LANDSCAPE_HEIGHT,
    width=LANDSCAPE_WIDTH,
    num_frames=MAX_FRAMES_MODEL,
)



default_negative_prompt = ""


def get_duration(
    prompt,
    negative_prompt,
    guidance_scale,
    guidance_scale_2,
    steps,
    seed,
    randomize_seed,
    # width,
    # height,
    progress,
):
    return steps * 2

@spaces.GPU(duration = get_duration)
def generate_image(
    prompt,
    negative_prompt=default_negative_prompt,
    guidance_scale = 3.5,
    guidance_scale_2 = 4,
    steps = 27,
    seed = 42,
    randomize_seed = False,
    # width=1024, 
    # height=1024,
    progress=gr.Progress(track_tqdm=True),
):
    """
    Generate an image from a text prompt using the Wan 2.2 14B T2V model.
    
    This function takes an input prompt and generates an image based on the provided
    prompt and parameters. It uses an FP8 qunatized Wan 2.2 14B Text-to-Video model.
    
    Args:
        prompt (str): Text prompt describing the desired image.
        negative_prompt (str, optional): Negative prompt to avoid unwanted elements. 
            Defaults to default_negative_prompt (contains unwanted visual artifacts).
        guidance_scale (float, optional): Controls adherence to the prompt. Higher values = more adherence.
            Defaults to 1.0. Range: 0.0-20.0.
        guidance_scale_2 (float, optional): Controls adherence to the prompt. Higher values = more adherence.
            Defaults to 1.0. Range: 0.0-20.0.
        steps (int, optional): Number of inference steps. More steps = higher quality but slower.
            Defaults to 4. Range: 1-30.
        seed (int, optional): Random seed for reproducible results. Defaults to 42.
            Range: 0 to MAX_SEED (2147483647).
        randomize_seed (bool, optional): Whether to use a random seed instead of the provided seed.
            Defaults to False.
        progress (gr.Progress, optional): Gradio progress tracker. Defaults to gr.Progress(track_tqdm=True).
    
    Returns:
        tuple: A tuple containing:
            - image_path (str): Path to the generated image 
            - current_seed (int): The seed used for generation (useful when randomize_seed=True)
    
    Raises:
        gr.Error: If input_image is None (no image uploaded).
    
    Note:
        - The function uses GPU acceleration via the @spaces.GPU decorator
        - Generation time varies based on steps
    """
    
   
    current_seed = random.randint(0, MAX_SEED) if randomize_seed else int(seed)

    out_img = pipe(
        prompt=prompt,
        negative_prompt=negative_prompt,
        height=1024,
        width=1024,
        num_frames=1,
        guidance_scale=float(guidance_scale),
        guidance_scale_2=float(guidance_scale_2),
        num_inference_steps=int(steps),
        output_type="pil",
        generator=torch.Generator(device="cuda").manual_seed(current_seed),
    ).frames[0][0]

    return out_img, current_seed
css="""
#col-container {
    margin: 0 auto;
    max-width: 620px;
}
"""
with gr.Blocks(css=css) as demo:
    with gr.Column(elem_id="col-container"):
       gr.Markdown("# Wan 2.2 (14B) Image")
       gr.Markdown("generate high quality images with Wan 2.2 14B")
       with gr.Row():
          prompt_input = gr.Textbox(show_label=False,
                max_lines=1,
                placeholder="Enter your prompt",
                container=False,)
          generate_button = gr.Button("Run", variant="primary", scale=0)
       img_output = gr.Image(label="Generated Image", interactive=False)
       with gr.Accordion("Advanced Settings", open=False):
            negative_prompt_input = gr.Textbox(label="Negative Prompt", value=default_negative_prompt, lines=3)
            seed_input = gr.Slider(label="Seed", minimum=0, maximum=MAX_SEED, step=1, value=42, interactive=True)
            randomize_seed_checkbox = gr.Checkbox(label="Randomize seed", value=True, interactive=True)
            with gr.Row():
                width = gr.Slider(label="Width", value=1024, minimum=64, maximum=1280, step=16)
                height = gr.Slider(label="Height", value=1024, minimum=64, maximum=1280, step=16)
            steps_slider = gr.Slider(minimum=1, maximum=40, step=1, value=27, label="Inference Steps") 
            guidance_scale_input = gr.Slider(minimum=0.0, maximum=10.0, step=0.5, value=3.5, label="Guidance Scale - high noise stage")
            guidance_scale_2_input = gr.Slider(minimum=0.0, maximum=10.0, step=0.5, value=4, label="Guidance Scale 2 - low noise stage")

       gr.Examples(
        examples=[
            [
                "Extreme close-up portrait of a theatrical villain character with chalk-white face paint, blood-red lips stretched into an unsettling wide grin, wild green hair disheveled and unkempt, piercing eyes with dark smudged makeup, dramatic high contrast lighting casting deep shadows across angular facial features, menacing expression with visible teeth, gritty urban atmosphere, cinematic quality, sharp focus on facial details"
            ], 
            [
                "Whimsical fiber art sculpture of a wise green alien sage character crafted entirely from soft wool yarn, sitting contentedly in a vintage clawfoot bathtub filled with fluffy white soap bubbles, the yarn figure's earth-tone knitted robes darkened and soggy from bathwater, intricate crochet textures visible on the wet yarn surface, iridescent soap bubbles floating through the air catching rainbow reflections, small rubber duck and bath toys scattered around the tub, warm bathroom lighting creating cozy shadows, steam gently rising from the warm water, the character's embroidered facial features showing peaceful relaxation, macro photography highlighting the contrast between wet yarn fibers and delicate soap foam, quirky handcrafted art piece with playful domestic setting"
            ],
            [
                "Macro photograph of a honey bee delicately perched on vibrant purple lavender blossoms during golden hour, warm amber sunlight filtering through translucent wings, soft bokeh background with gradient sky transitioning from orange to deep blue, dewdrops glistening on flower petals, shallow depth of field emphasizing the bee's fuzzy texture and intricate wing details, peaceful twilight atmosphere, cinematic lighting, ultra-sharp focus on the bee"
            ],
        ],
        inputs=[prompt_input], outputs=[img_output, seed_input], fn=generate_image, cache_examples="lazy")            
    ui_inputs = [ 
        prompt_input,
        negative_prompt_input,
        guidance_scale_input, guidance_scale_2_input, steps_slider, seed_input, randomize_seed_checkbox 
        # width, height
    ]
    generate_button.click(fn=generate_image, inputs=ui_inputs, outputs=[img_output, seed_input])

    

if __name__ == "__main__":
    demo.queue().launch(mcp_server=True)