Spaces:
Runtime error
Runtime error
File size: 15,545 Bytes
714bf26 |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 |
from enum import Enum
import gc
import numpy as np
import torch
import decord
from diffusers import StableDiffusionInstructPix2PixPipeline, StableDiffusionControlNetPipeline, ControlNetModel, UNet2DConditionModel
from diffusers.schedulers import EulerAncestralDiscreteScheduler, DDIMScheduler
from text_to_video.text_to_video_pipeline import TextToVideoPipeline
import utils
import gradio_utils
decord.bridge.set_bridge('torch')
class ModelType(Enum):
Pix2Pix_Video = 1,
Text2Video = 2,
ControlNetCanny = 3,
ControlNetCannyDB = 4,
ControlNetPose = 5,
class Model:
def __init__(self, device, dtype, **kwargs):
self.device = device
self.dtype = dtype
self.generator = torch.Generator(device=device)
self.pipe_dict = {
ModelType.Pix2Pix_Video: StableDiffusionInstructPix2PixPipeline,
ModelType.Text2Video: TextToVideoPipeline,
ModelType.ControlNetCanny: StableDiffusionControlNetPipeline,
ModelType.ControlNetCannyDB: StableDiffusionControlNetPipeline,
ModelType.ControlNetPose: StableDiffusionControlNetPipeline,
}
self.controlnet_attn_proc = utils.CrossFrameAttnProcessor(unet_chunk_size=2)
self.pix2pix_attn_proc = utils.CrossFrameAttnProcessor(unet_chunk_size=3)
self.text2video_attn_proc = utils.CrossFrameAttnProcessor(unet_chunk_size=2)
self.pipe = None
self.model_type = None
self.states = {}
def set_model(self, model_type: ModelType, model_id: str, **kwargs):
if self.pipe is not None:
del self.pipe
torch.cuda.empty_cache()
gc.collect()
safety_checker = kwargs.pop('safety_checker', None)
self.pipe = self.pipe_dict[model_type].from_pretrained(model_id, safety_checker=safety_checker, **kwargs).to(self.device).to(self.dtype)
self.model_type = model_type
def inference_chunk(self, frame_ids, **kwargs):
if self.pipe is None:
return
image = kwargs.pop('image')
prompt = np.array(kwargs.pop('prompt'))
negative_prompt = np.array(kwargs.pop('negative_prompt', ''))
latents = None
if 'latents' in kwargs:
latents = kwargs.pop('latents')[frame_ids]
return self.pipe(image=image[frame_ids],
prompt=prompt[frame_ids].tolist(),
negative_prompt=negative_prompt[frame_ids].tolist(),
latents=latents,
generator=self.generator,
**kwargs)
def inference(self, split_to_chunks=False, chunk_size=8, **kwargs):
if self.pipe is None:
return
seed = kwargs.pop('seed', 0)
kwargs.pop('generator', '')
# self.generator.manual_seed(seed)
if split_to_chunks:
assert 'image' in kwargs
assert 'prompt' in kwargs
image = kwargs.pop('image')
prompt = kwargs.pop('prompt')
negative_prompt = kwargs.pop('negative_prompt', '')
f = image.shape[0]
chunk_ids = np.arange(0, f, chunk_size - 1)
result = []
for i in range(len(chunk_ids)):
ch_start = chunk_ids[i]
ch_end = f if i == len(chunk_ids) - 1 else chunk_ids[i + 1]
frame_ids = [0] + list(range(ch_start, ch_end))
self.generator.manual_seed(seed)
print(f'Processing chunk {i + 1} / {len(chunk_ids)}')
result.append(self.inference_chunk(frame_ids=frame_ids,
image=image,
prompt=[prompt] * f,
negative_prompt=[negative_prompt] * f,
**kwargs).images[1:])
result = np.concatenate(result)
return result
else:
return self.pipe(generator=self.generator, **kwargs).videos[0]
def process_controlnet_canny(self,
video_path,
prompt,
num_inference_steps=20,
controlnet_conditioning_scale=1.0,
guidance_scale=9.0,
seed=42,
eta=0.0,
low_threshold=100,
high_threshold=200,
resolution=512):
video_path = gradio_utils.edge_path_to_video_path(video_path)
if self.model_type != ModelType.ControlNetCanny:
controlnet = ControlNetModel.from_pretrained("lllyasviel/sd-controlnet-canny")
self.set_model(ModelType.ControlNetCanny, model_id="runwayml/stable-diffusion-v1-5", controlnet=controlnet)
self.pipe.scheduler = DDIMScheduler.from_config(self.pipe.scheduler.config)
self.pipe.unet.set_attn_processor(processor=self.controlnet_attn_proc)
self.pipe.controlnet.set_attn_processor(processor=self.controlnet_attn_proc)
# TODO: Check scheduler
added_prompt = 'best quality, extremely detailed'
negative_prompts = 'longbody, lowres, bad anatomy, bad hands, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality'
video, fps = utils.prepare_video(video_path, resolution, self.device, self.dtype, False)
control = utils.pre_process_canny(video, low_threshold, high_threshold).to(self.device).to(self.dtype)
f, _, h, w = video.shape
self.generator.manual_seed(seed)
latents = torch.randn((1, 4, h//8, w//8), dtype=self.dtype, device=self.device, generator=self.generator)
latents = latents.repeat(f, 1, 1, 1)
result = self.inference(image=control,
prompt=prompt + ', ' + added_prompt,
height=h,
width=w,
negative_prompt=negative_prompts,
num_inference_steps=num_inference_steps,
guidance_scale=guidance_scale,
controlnet_conditioning_scale=controlnet_conditioning_scale,
eta=eta,
latents=latents,
seed=seed,
output_type='numpy',
split_to_chunks=True,
chunk_size=8,
)
return utils.create_video(result, fps)
def process_controlnet_pose(self,
video_path,
prompt,
num_inference_steps=20,
controlnet_conditioning_scale=1.0,
guidance_scale=9.0,
seed=42,
eta=0.0,
resolution=512):
video_path = gradio_utils.motion_to_video_path(video_path)
if self.model_type != ModelType.ControlNetPose:
controlnet = ControlNetModel.from_pretrained("fusing/stable-diffusion-v1-5-controlnet-openpose")
self.set_model(ModelType.ControlNetPose, model_id="runwayml/stable-diffusion-v1-5", controlnet=controlnet)
self.pipe.scheduler = DDIMScheduler.from_config(self.pipe.scheduler.config)
self.pipe.unet.set_attn_processor(processor=self.controlnet_attn_proc)
self.pipe.controlnet.set_attn_processor(processor=self.controlnet_attn_proc)
added_prompt = 'best quality, extremely detailed, HD, ultra-realistic, 8K, HQ, masterpiece, trending on artstation, art, smooth'
negative_prompts = 'longbody, lowres, bad anatomy, bad hands, missing fingers, extra digit, fewer difits, cropped, worst quality, low quality, deformed body, bloated, ugly, unrealistic'
video, fps = utils.prepare_video(video_path, resolution, self.device, self.dtype, False, output_fps=4)
control = utils.pre_process_pose(video, apply_pose_detect=False).to(self.device).to(self.dtype)
f, _, h, w = video.shape
self.generator.manual_seed(seed)
latents = torch.randn((1, 4, h//8, w//8), dtype=self.dtype, device=self.device, generator=self.generator)
latents = latents.repeat(f, 1, 1, 1)
result = self.inference(image=control,
prompt=prompt + ', ' + added_prompt,
height=h,
width=w,
negative_prompt=negative_prompts,
num_inference_steps=num_inference_steps,
guidance_scale=guidance_scale,
controlnet_conditioning_scale=controlnet_conditioning_scale,
eta=eta,
latents=latents,
seed=seed,
output_type='numpy',
split_to_chunks=True,
chunk_size=8,
)
return utils.create_gif(result, fps)
def process_controlnet_canny_db(self,
db_path,
video_path,
prompt,
num_inference_steps=20,
controlnet_conditioning_scale=1.0,
guidance_scale=9.0,
seed=42,
eta=0.0,
low_threshold=100,
high_threshold=200,
resolution=512):
db_path = gradio_utils.get_model_from_db_selection(db_path)
video_path = gradio_utils.get_video_from_canny_selection(video_path)
# Load db and controlnet weights
if 'db_path' not in self.states or db_path != self.states['db_path']:
controlnet = ControlNetModel.from_pretrained("lllyasviel/sd-controlnet-canny")
self.set_model(ModelType.ControlNetCannyDB, model_id=db_path, controlnet=controlnet)
self.pipe.scheduler = DDIMScheduler.from_config(self.pipe.scheduler.config)
self.pipe.unet.set_attn_processor(processor=self.controlnet_attn_proc)
self.pipe.controlnet.set_attn_processor(processor=self.controlnet_attn_proc)
self.states['db_path'] = db_path
added_prompt = 'best quality, extremely detailed'
negative_prompts = 'longbody, lowres, bad anatomy, bad hands, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality'
video, fps = utils.prepare_video(video_path, resolution, self.device, self.dtype, False)
control = utils.pre_process_canny(video, low_threshold, high_threshold).to(self.device).to(self.dtype)
f, _, h, w = video.shape
self.generator.manual_seed(seed)
latents = torch.randn((1, 4, h//8, w//8), dtype=self.dtype, device=self.device, generator=self.generator)
latents = latents.repeat(f, 1, 1, 1)
result = self.inference(image=control,
prompt=prompt + ', ' + added_prompt,
height=h,
width=w,
negative_prompt=negative_prompts,
num_inference_steps=num_inference_steps,
guidance_scale=guidance_scale,
controlnet_conditioning_scale=controlnet_conditioning_scale,
eta=eta,
latents=latents,
seed=seed,
output_type='numpy',
split_to_chunks=True,
chunk_size=8,
)
return utils.create_gif(result, fps)
def process_pix2pix(self, video, prompt, resolution=512, seed=0, start_t=0, end_t=-1, out_fps=-1):
if self.model_type != ModelType.Pix2Pix_Video:
self.set_model(ModelType.Pix2Pix_Video, model_id="timbrooks/instruct-pix2pix")
self.pipe.scheduler = EulerAncestralDiscreteScheduler.from_config(self.pipe.scheduler.config)
self.pipe.unet.set_attn_processor(processor=self.pix2pix_attn_proc)
video, fps = utils.prepare_video(video, resolution, self.device, self.dtype, True, start_t, end_t, out_fps)
self.generator.manual_seed(seed)
result = self.inference(image=video,
prompt=prompt,
seed=seed,
output_type='numpy',
num_inference_steps=50,
image_guidance_scale=1.5,
split_to_chunks=True,
chunk_size=8,
)
return utils.create_video(result, fps)
def process_text2video(self, prompt, resolution=512, seed=24, num_frames=8, fps=4, t0=881, t1=941,
use_cf_attn=True, use_motion_field=True, use_foreground_motion_field=False,
smooth_bg=False, smooth_bg_strength=0.4, motion_field_strength=12):
if self.model_type != ModelType.Text2Video:
unet = UNet2DConditionModel.from_pretrained("runwayml/stable-diffusion-v1-5", subfolder="unet")
self.set_model(ModelType.Text2Video, model_id="runwayml/stable-diffusion-v1-5", unet=unet)
self.pipe.scheduler = DDIMScheduler.from_config(self.pipe.scheduler.config)
self.pipe.unet.set_attn_processor(processor=self.text2video_attn_proc)
self.generator.manual_seed(seed)
added_prompt = "high quality, HD, 8K, trending on artstation, high focus, dramatic lighting"
self.generator.manual_seed(seed)
prompt = prompt.rstrip()
if len(prompt) > 0 and (prompt[-1] == "," or prompt[-1] == "."):
prompt = prompt.rstrip()[:-1]
prompt = prompt.rstrip()
prompt = prompt + ", "+added_prompt
result = self.inference(prompt=[prompt],
video_length=num_frames,
height=resolution,
width=resolution,
num_inference_steps=50,
guidance_scale=7.5,
guidance_stop_step=1.0,
t0=t0,
t1=t1,
use_foreground_motion_field=use_foreground_motion_field,
motion_field_strength=motion_field_strength,
use_motion_field=use_motion_field,
smooth_bg=smooth_bg,
smooth_bg_strength=smooth_bg_strength,
seed=seed,
output_type='numpy',
)
return utils.create_video(result, fps)
|