Spaces:
Running
on
Zero
Running
on
Zero
Update app.py
Browse files
app.py
CHANGED
|
@@ -1,85 +1,106 @@
|
|
| 1 |
import gradio as gr
|
| 2 |
import numpy as np
|
| 3 |
import random
|
| 4 |
-
from diffusers import AuraFlowPipeline
|
| 5 |
import torch
|
| 6 |
-
import
|
|
|
|
| 7 |
|
| 8 |
-
|
| 9 |
-
|
| 10 |
-
|
| 11 |
-
|
| 12 |
-
|
| 13 |
-
|
| 14 |
-
#torch._inductor.config.epilogue_fusion = False
|
| 15 |
-
#torch._inductor.config.coordinate_descent_check_all_directions = True
|
| 16 |
-
|
| 17 |
-
pipe = AuraFlowPipeline.from_pretrained(
|
| 18 |
-
"fal/AuraFlow",
|
| 19 |
-
torch_dtype=torch.float16
|
| 20 |
-
).to("cuda")
|
| 21 |
|
| 22 |
-
|
| 23 |
-
#pipe.transformer = torch.compile(pipe.transformer, mode="reduce-overhead", fullgraph=True)
|
| 24 |
-
#pipe.transformer.to(memory_format=torch.channels_last)
|
| 25 |
-
#pipe.vae.to(memory_format=torch.channels_last)
|
| 26 |
|
| 27 |
-
|
| 28 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 29 |
|
| 30 |
MAX_SEED = np.iinfo(np.int32).max
|
| 31 |
MAX_IMAGE_SIZE = 1024
|
| 32 |
|
| 33 |
-
|
| 34 |
-
def infer(prompt, negative_prompt="", seed=42, randomize_seed=False, width=1024, height=1024, guidance_scale=5.0, num_inference_steps=28, progress=gr.Progress(track_tqdm=True)):
|
| 35 |
-
|
| 36 |
if randomize_seed:
|
| 37 |
seed = random.randint(0, MAX_SEED)
|
| 38 |
-
|
| 39 |
-
generator = torch.Generator().manual_seed(seed)
|
|
|
|
|
|
|
| 40 |
|
| 41 |
image = pipe(
|
| 42 |
-
prompt
|
| 43 |
-
|
| 44 |
-
|
| 45 |
height=height,
|
| 46 |
-
|
| 47 |
-
num_inference_steps
|
| 48 |
-
|
| 49 |
-
|
|
|
|
|
|
|
| 50 |
|
| 51 |
return image, seed
|
| 52 |
|
| 53 |
examples = [
|
| 54 |
-
"A photo of a lavender cat",
|
| 55 |
-
"Astronaut in a jungle, cold color palette, muted colors, detailed, 8k",
|
| 56 |
-
"An astronaut riding a green horse",
|
| 57 |
-
"A delicious ceviche cheesecake slice",
|
| 58 |
]
|
| 59 |
|
| 60 |
css="""
|
| 61 |
#col-container {
|
| 62 |
margin: 0 auto;
|
| 63 |
-
max-width:
|
| 64 |
}
|
| 65 |
"""
|
| 66 |
-
|
| 67 |
-
if torch.cuda.is_available():
|
| 68 |
-
power_device = "GPU"
|
| 69 |
-
else:
|
| 70 |
-
power_device = "CPU"
|
| 71 |
-
|
| 72 |
with gr.Blocks(css=css) as demo:
|
| 73 |
-
|
| 74 |
with gr.Column(elem_id="col-container"):
|
| 75 |
gr.Markdown(f"""
|
| 76 |
-
#
|
| 77 |
-
Demo of the
|
| 78 |
-
[[blog](https://blog.fal.ai/auraflow/)] [[model](https://huggingface.co/fal/AuraFlow)] [[fal](https://fal.ai/models/fal-ai/aura-flow)]
|
| 79 |
""")
|
| 80 |
|
| 81 |
with gr.Row():
|
| 82 |
-
|
| 83 |
prompt = gr.Text(
|
| 84 |
label="Prompt",
|
| 85 |
show_label=False,
|
|
@@ -87,19 +108,18 @@ with gr.Blocks(css=css) as demo:
|
|
| 87 |
placeholder="Enter your prompt",
|
| 88 |
container=False,
|
| 89 |
)
|
| 90 |
-
|
| 91 |
run_button = gr.Button("Run", scale=0)
|
| 92 |
|
| 93 |
-
|
|
|
|
|
|
|
| 94 |
|
| 95 |
with gr.Accordion("Advanced Settings", open=False):
|
| 96 |
-
|
| 97 |
negative_prompt = gr.Text(
|
| 98 |
label="Negative prompt",
|
| 99 |
max_lines=1,
|
| 100 |
placeholder="Enter a negative prompt",
|
| 101 |
)
|
| 102 |
-
|
| 103 |
seed = gr.Slider(
|
| 104 |
label="Seed",
|
| 105 |
minimum=0,
|
|
@@ -107,11 +127,8 @@ with gr.Blocks(css=css) as demo:
|
|
| 107 |
step=1,
|
| 108 |
value=0,
|
| 109 |
)
|
| 110 |
-
|
| 111 |
randomize_seed = gr.Checkbox(label="Randomize seed", value=True)
|
| 112 |
-
|
| 113 |
with gr.Row():
|
| 114 |
-
|
| 115 |
width = gr.Slider(
|
| 116 |
label="Width",
|
| 117 |
minimum=256,
|
|
@@ -119,7 +136,6 @@ with gr.Blocks(css=css) as demo:
|
|
| 119 |
step=32,
|
| 120 |
value=1024,
|
| 121 |
)
|
| 122 |
-
|
| 123 |
height = gr.Slider(
|
| 124 |
label="Height",
|
| 125 |
minimum=256,
|
|
@@ -127,9 +143,7 @@ with gr.Blocks(css=css) as demo:
|
|
| 127 |
step=32,
|
| 128 |
value=1024,
|
| 129 |
)
|
| 130 |
-
|
| 131 |
with gr.Row():
|
| 132 |
-
|
| 133 |
guidance_scale = gr.Slider(
|
| 134 |
label="Guidance scale",
|
| 135 |
minimum=0.0,
|
|
@@ -137,28 +151,34 @@ with gr.Blocks(css=css) as demo:
|
|
| 137 |
step=0.1,
|
| 138 |
value=5.0,
|
| 139 |
)
|
| 140 |
-
|
| 141 |
num_inference_steps = gr.Slider(
|
| 142 |
label="Number of inference steps",
|
| 143 |
minimum=1,
|
| 144 |
-
maximum=
|
| 145 |
step=1,
|
| 146 |
-
value=
|
| 147 |
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 148 |
|
| 149 |
gr.Examples(
|
| 150 |
-
examples
|
| 151 |
-
fn
|
| 152 |
-
inputs
|
| 153 |
-
outputs
|
| 154 |
cache_examples="lazy"
|
| 155 |
)
|
| 156 |
|
| 157 |
gr.on(
|
| 158 |
-
triggers=[run_button.click, prompt.submit
|
| 159 |
-
fn
|
| 160 |
-
inputs
|
| 161 |
-
outputs
|
| 162 |
)
|
| 163 |
|
| 164 |
demo.queue().launch()
|
|
|
|
| 1 |
import gradio as gr
|
| 2 |
import numpy as np
|
| 3 |
import random
|
|
|
|
| 4 |
import torch
|
| 5 |
+
from PIL import Image
|
| 6 |
+
import os
|
| 7 |
|
| 8 |
+
from transformers import CLIPVisionModelWithProjection, CLIPImageProcessor
|
| 9 |
+
from kolors.pipelines.pipeline_stable_diffusion_xl_chatglm_256_ipadapter import StableDiffusionXLPipeline
|
| 10 |
+
from kolors.models.modeling_chatglm import ChatGLMModel
|
| 11 |
+
from kolors.models.tokenization_chatglm import ChatGLMTokenizer
|
| 12 |
+
from kolors.models.unet_2d_condition import UNet2DConditionModel
|
| 13 |
+
from diffusers import AutoencoderKL, EulerDiscreteScheduler
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 14 |
|
| 15 |
+
from huggingface_hub import snapshot_download
|
|
|
|
|
|
|
|
|
|
| 16 |
|
| 17 |
+
device = "cuda" if torch.cuda.is_available() else "cpu"
|
| 18 |
+
root_dir = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
|
| 19 |
+
ckpt_dir = f'{root_dir}/weights/Kolors'
|
| 20 |
+
|
| 21 |
+
snapshot_download(repo_id="Kwai-Kolors/Kolors", local_dir=ckpt_dir)
|
| 22 |
+
snapshot_download(repo_id="Kwai-Kolors/Kolors-IP-Adapter-Plus", local_dir=f"{root_dir}/weights/Kolors-IP-Adapter-Plus")
|
| 23 |
+
|
| 24 |
+
# Load models
|
| 25 |
+
text_encoder = ChatGLMModel.from_pretrained(f'{ckpt_dir}/text_encoder', torch_dtype=torch.float16).half().to(device)
|
| 26 |
+
tokenizer = ChatGLMTokenizer.from_pretrained(f'{ckpt_dir}/text_encoder')
|
| 27 |
+
vae = AutoencoderKL.from_pretrained(f"{ckpt_dir}/vae", revision=None).half().to(device)
|
| 28 |
+
scheduler = EulerDiscreteScheduler.from_pretrained(f"{ckpt_dir}/scheduler")
|
| 29 |
+
unet = UNet2DConditionModel.from_pretrained(f"{ckpt_dir}/unet", revision=None).half().to(device)
|
| 30 |
+
|
| 31 |
+
image_encoder = CLIPVisionModelWithProjection.from_pretrained(
|
| 32 |
+
f'{root_dir}/weights/Kolors-IP-Adapter-Plus/image_encoder',
|
| 33 |
+
ignore_mismatched_sizes=True
|
| 34 |
+
).to(dtype=torch.float16, device=device)
|
| 35 |
+
|
| 36 |
+
ip_img_size = 336
|
| 37 |
+
clip_image_processor = CLIPImageProcessor(size=ip_img_size, crop_size=ip_img_size)
|
| 38 |
+
|
| 39 |
+
pipe = StableDiffusionXLPipeline(
|
| 40 |
+
vae=vae,
|
| 41 |
+
text_encoder=text_encoder,
|
| 42 |
+
tokenizer=tokenizer,
|
| 43 |
+
unet=unet,
|
| 44 |
+
scheduler=scheduler,
|
| 45 |
+
image_encoder=image_encoder,
|
| 46 |
+
feature_extractor=clip_image_processor,
|
| 47 |
+
force_zeros_for_empty_prompt=False
|
| 48 |
+
)
|
| 49 |
+
|
| 50 |
+
pipe = pipe.to(device)
|
| 51 |
+
#pipe.enable_model_cpu_offload()
|
| 52 |
+
|
| 53 |
+
if hasattr(pipe.unet, 'encoder_hid_proj'):
|
| 54 |
+
pipe.unet.text_encoder_hid_proj = pipe.unet.encoder_hid_proj
|
| 55 |
+
|
| 56 |
+
pipe.load_ip_adapter(f'{root_dir}/weights/Kolors-IP-Adapter-Plus', subfolder="", weight_name=["ip_adapter_plus_general.bin"])
|
| 57 |
|
| 58 |
MAX_SEED = np.iinfo(np.int32).max
|
| 59 |
MAX_IMAGE_SIZE = 1024
|
| 60 |
|
| 61 |
+
def infer(prompt, ip_adapter_image, negative_prompt="", seed=42, randomize_seed=False, width=1024, height=1024, guidance_scale=5.0, num_inference_steps=50, ip_adapter_scale=0.5, progress=gr.Progress(track_tqdm=True)):
|
|
|
|
|
|
|
| 62 |
if randomize_seed:
|
| 63 |
seed = random.randint(0, MAX_SEED)
|
| 64 |
+
|
| 65 |
+
generator = torch.Generator(device="cpu").manual_seed(seed)
|
| 66 |
+
|
| 67 |
+
pipe.set_ip_adapter_scale([ip_adapter_scale])
|
| 68 |
|
| 69 |
image = pipe(
|
| 70 |
+
prompt=prompt,
|
| 71 |
+
ip_adapter_image=[ip_adapter_image],
|
| 72 |
+
negative_prompt=negative_prompt,
|
| 73 |
height=height,
|
| 74 |
+
width=width,
|
| 75 |
+
num_inference_steps=num_inference_steps,
|
| 76 |
+
guidance_scale=guidance_scale,
|
| 77 |
+
num_images_per_prompt=1,
|
| 78 |
+
generator=generator,
|
| 79 |
+
).images[0]
|
| 80 |
|
| 81 |
return image, seed
|
| 82 |
|
| 83 |
examples = [
|
| 84 |
+
["A photo of a lavender cat", "https://upload.wikimedia.org/wikipedia/commons/thumb/4/4d/Cat_November_2010-1a.jpg/640px-Cat_November_2010-1a.jpg"],
|
| 85 |
+
["Astronaut in a jungle, cold color palette, muted colors, detailed, 8k", "https://upload.wikimedia.org/wikipedia/commons/thumb/b/b5/Astronaut_EVA.jpg/640px-Astronaut_EVA.jpg"],
|
| 86 |
+
["An astronaut riding a green horse", "https://upload.wikimedia.org/wikipedia/commons/thumb/f/f7/Haflinger_in-motion.jpg/640px-Haflinger_in-motion.jpg"],
|
| 87 |
+
["A delicious ceviche cheesecake slice", "https://upload.wikimedia.org/wikipedia/commons/thumb/9/9c/Ceviche_mixto.jpg/640px-Ceviche_mixto.jpg"],
|
| 88 |
]
|
| 89 |
|
| 90 |
css="""
|
| 91 |
#col-container {
|
| 92 |
margin: 0 auto;
|
| 93 |
+
max-width: 720px;
|
| 94 |
}
|
| 95 |
"""
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 96 |
with gr.Blocks(css=css) as demo:
|
|
|
|
| 97 |
with gr.Column(elem_id="col-container"):
|
| 98 |
gr.Markdown(f"""
|
| 99 |
+
# Kolors Demo
|
| 100 |
+
Demo of the Kolors model with IP-Adapter integration
|
|
|
|
| 101 |
""")
|
| 102 |
|
| 103 |
with gr.Row():
|
|
|
|
| 104 |
prompt = gr.Text(
|
| 105 |
label="Prompt",
|
| 106 |
show_label=False,
|
|
|
|
| 108 |
placeholder="Enter your prompt",
|
| 109 |
container=False,
|
| 110 |
)
|
|
|
|
| 111 |
run_button = gr.Button("Run", scale=0)
|
| 112 |
|
| 113 |
+
with gr.Row():
|
| 114 |
+
ip_adapter_image = gr.Image(label="IP-Adapter Image", type="pil")
|
| 115 |
+
result = gr.Image(label="Result", show_label=False)
|
| 116 |
|
| 117 |
with gr.Accordion("Advanced Settings", open=False):
|
|
|
|
| 118 |
negative_prompt = gr.Text(
|
| 119 |
label="Negative prompt",
|
| 120 |
max_lines=1,
|
| 121 |
placeholder="Enter a negative prompt",
|
| 122 |
)
|
|
|
|
| 123 |
seed = gr.Slider(
|
| 124 |
label="Seed",
|
| 125 |
minimum=0,
|
|
|
|
| 127 |
step=1,
|
| 128 |
value=0,
|
| 129 |
)
|
|
|
|
| 130 |
randomize_seed = gr.Checkbox(label="Randomize seed", value=True)
|
|
|
|
| 131 |
with gr.Row():
|
|
|
|
| 132 |
width = gr.Slider(
|
| 133 |
label="Width",
|
| 134 |
minimum=256,
|
|
|
|
| 136 |
step=32,
|
| 137 |
value=1024,
|
| 138 |
)
|
|
|
|
| 139 |
height = gr.Slider(
|
| 140 |
label="Height",
|
| 141 |
minimum=256,
|
|
|
|
| 143 |
step=32,
|
| 144 |
value=1024,
|
| 145 |
)
|
|
|
|
| 146 |
with gr.Row():
|
|
|
|
| 147 |
guidance_scale = gr.Slider(
|
| 148 |
label="Guidance scale",
|
| 149 |
minimum=0.0,
|
|
|
|
| 151 |
step=0.1,
|
| 152 |
value=5.0,
|
| 153 |
)
|
|
|
|
| 154 |
num_inference_steps = gr.Slider(
|
| 155 |
label="Number of inference steps",
|
| 156 |
minimum=1,
|
| 157 |
+
maximum=100,
|
| 158 |
step=1,
|
| 159 |
+
value=50,
|
| 160 |
)
|
| 161 |
+
ip_adapter_scale = gr.Slider(
|
| 162 |
+
label="IP-Adapter Scale",
|
| 163 |
+
minimum=0.0,
|
| 164 |
+
maximum=1.0,
|
| 165 |
+
step=0.01,
|
| 166 |
+
value=0.5,
|
| 167 |
+
)
|
| 168 |
|
| 169 |
gr.Examples(
|
| 170 |
+
examples=examples,
|
| 171 |
+
fn=infer,
|
| 172 |
+
inputs=[prompt, ip_adapter_image],
|
| 173 |
+
outputs=[result, seed],
|
| 174 |
cache_examples="lazy"
|
| 175 |
)
|
| 176 |
|
| 177 |
gr.on(
|
| 178 |
+
triggers=[run_button.click, prompt.submit],
|
| 179 |
+
fn=infer,
|
| 180 |
+
inputs=[prompt, ip_adapter_image, negative_prompt, seed, randomize_seed, width, height, guidance_scale, num_inference_steps, ip_adapter_scale],
|
| 181 |
+
outputs=[result, seed]
|
| 182 |
)
|
| 183 |
|
| 184 |
demo.queue().launch()
|