File size: 120,336 Bytes
96d06f0 724285b 96d06f0 724285b 96d06f0 724285b 96d06f0 724285b 96d06f0 724285b 96d06f0 724285b 96d06f0 724285b 96d06f0 724285b 96d06f0 724285b 96d06f0 724285b 96d06f0 724285b 96d06f0 724285b 96d06f0 724285b 96d06f0 724285b 96d06f0 724285b 96d06f0 724285b 96d06f0 724285b 96d06f0 724285b 96d06f0 724285b 96d06f0 724285b 96d06f0 724285b 96d06f0 724285b 96d06f0 724285b 96d06f0 724285b 96d06f0 724285b 96d06f0 724285b 96d06f0 724285b 96d06f0 724285b 96d06f0 724285b 96d06f0 724285b 96d06f0 |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 |
# 简介
本节将会涵盖一些基础内容,介绍如何利用现有的管线(pipeline)借助 Stable Diffusion 模型来生成和修改图片。我们也将简要介绍管线内的关键组成部分。总的来讲,本节涵盖以下内容:
- 利用 `StableDiffusionPipeline` 根据文字描述生成图片,并探究各个输入参数的意义
- 用代码来了解管线的一些关键组成部分:
- 让这个模型成为“隐编码扩散模型(latent diffusion model)”的可变分自编码器(VAE)
- 处理文本提示的分词器(tokenizer)和文本编码器
- UNet 模型
- 使用的调度器(scheduler),以及其它不同的调度器
- 使用管线的组成部分来复现采样循环
- 用 Img2Img 管线来编辑现有图片
- 使用 inpainting 管线和 Depth2Img 管线
❓如果你有问题,请在 Hugging Face Discord 的 `#diffusion-models-class` 频道提出。如果你还没有 Hugging Face 的账号,你可以在这里注册:https://huggingface.co/join/discord
# 配置
```python
!pip install -Uq diffusers ftfy accelerate
```
```python
# Installing transformers from source for now since we need the latest version for Depth2Img:
!pip install -Uq git+https://github.com/huggingface/transformers
```
```python
import torch
import requests
from PIL import Image
from io import BytesIO
from matplotlib import pyplot as plt
# We'll be exploring a number of pipelines today!
from diffusers import (
StableDiffusionPipeline,
StableDiffusionImg2ImgPipeline,
StableDiffusionInpaintPipeline,
StableDiffusionDepth2ImgPipeline
)
# We'll use a couple of demo images later in the notebook
def download_image(url):
response = requests.get(url)
return Image.open(BytesIO(response.content)).convert("RGB")
# Download images for inpainting example
img_url = "https://raw.githubusercontent.com/CompVis/latent-diffusion/main/data/inpainting_examples/overture-creations-5sI6fQgYIuo.png"
mask_url = "https://raw.githubusercontent.com/CompVis/latent-diffusion/main/data/inpainting_examples/overture-creations-5sI6fQgYIuo_mask.png"
init_image = download_image(img_url).resize((512, 512))
mask_image = download_image(mask_url).resize((512, 512))
```
```python
# Set device
device = (
"mps"
if torch.backends.mps.is_available()
else "cuda"
if torch.cuda.is_available()
else "cpu"
)
```
# 从文本生成图像
我们先载入 Stable Diffusion 的管线。现有的 Stable Diffusion 模型有好多不同版本,截至本书编写时的最新版本是第2.1版。修改版本只需要在 `model_id` 处修改即可(比如你可以试试将其改成 `CompVis/stable-diffusion-v1-4` 或从[dreambooth concepts library](https://huggingface.co/sd-dreambooth-library)选一个模型)。
```python
# Load the pipeline
model_id = "stabilityai/stable-diffusion-2-1-base"
pipe = StableDiffusionPipeline.from_pretrained(model_id).to(device)
```
如果你的GPU内存不够用,可以尝试这些方法减少内存使用:
- 载入 FP16 精度的版本(但并不是所有的系统上都支持),此时你也需要保证所有的张量也被改成了 torch.float16 精度:
`pipe = StableDiffusionPipeline.from_pretrained(model_id, revision="fp16", torch_dtype=torch.float16).to(device)`
- 开启注意力机制切分(attention slicing)。这会牺牲一点点速度来减少GPU内存的使用:
`pipe.enable_attention_slicing()`
- 降低生成结果的图片尺寸
当管线加载好了以后,我们可以用以下代码去使用文字提示生成图片:
```python
# Set up a generator for reproducibility
generator = torch.Generator(device=device).manual_seed(42)
# Run the pipeline, showing some of the available arguments
pipe_output = pipe(
prompt="Palette knife painting of an autumn cityscape", # What to generate
negative_prompt="Oversaturated, blurry, low quality", # What NOT to generate
height=480, width=640, # Specify the image size
guidance_scale=8, # How strongly to follow the prompt
num_inference_steps=35, # How many steps to take
generator=generator # Fixed random seed
)
# View the resulting image:
pipe_output.images[0]
```
0%| | 0/35 [00:00<?, ?it/s]
![png](01_stable_diffusion_introduction_CN_files/01_stable_diffusion_introduction_CN_10_1.png)
**练习:** 用上面的代码块做实验,使用你自己的文字提示,反复调整各项设置看看这些设置如何影响生成效果。使用不同的随机种子或者移除掉 `generator` 这个输入参数,看看能获取什么不同的效果。
主要的要调节参数介绍:
- `width` 和 `height` 指定了生成图片的尺寸。它们必须是可被 8 整除的数字,只有这样我们的可变分自编码器(VAE)才能正常工作(原因后面会将)。
- 步数 `num_inference_steps` 也会影响生成的质量。默认设成 50 就行,但你也可以试试设成 20,看看设置小了后效果会如何。
- 使用 `negative_prompt` 来强调不希望生成的内容,一般会在无分类器引导(classifier-free guidance)的情况下用到,这种添加额外控制的方式特别管用。列出一些不想要的特性对生成更好结果很有帮助。
- `guidance_scale` 这个参数决定了无分类器引导(CFG)的影响强度有多大。增大这个值会使得生成的内容更接近文字提示;但这个值如果过大,可能会使得结果变得过饱和、不好看。
如果你想为文字提示找点灵感,可以参考这里:[Stable Diffusion Prompt Book](https://stability.ai/sdv2-prompt-book)
下面展示了增大 `guidance_scale` 这个参数所带来的作用:
```python
#@markdown comparing guidance scales:
cfg_scales = [1.1, 8, 12] #@param
prompt = "A collie with a pink hat" #@param
fig, axs = plt.subplots(1, len(cfg_scales), figsize=(16, 5))
for i, ax in enumerate(axs):
im = pipe(prompt, height=480, width=480,
guidance_scale=cfg_scales[i], num_inference_steps=35,
generator=torch.Generator(device=device).manual_seed(42)).images[0]
ax.imshow(im); ax.set_title(f'CFG Scale {cfg_scales[i]}');
```
0%| | 0/35 [00:00<?, ?it/s]
0%| | 0/35 [00:00<?, ?it/s]
0%| | 0/35 [00:00<?, ?it/s]
![png](01_stable_diffusion_introduction_CN_files/01_stable_diffusion_introduction_CN_12_3.png)
当然了,如何解读这些参数是个很主观的事情,但总的来说,设置值到 8 到 12 这个区间内是个不错的选择。
# 管线的组成部分
这里我们用的 `StableDiffusionPipeline` 比前面几个单元的 `DDPMPipeline` 要复杂一点。除了 UNet 和调度器之外,管线内还有很多其它的组成部分:
```python
print(list(pipe.components.keys())) # List components
```
['vae', 'text_encoder', 'tokenizer', 'unet', 'scheduler', 'safety_checker', 'feature_extractor']
为了更好地理解管线如何工作,我们简要地逐个了解各个组成部分。然后可以自己用代码实现,并把它们组合在一起,以此复现出整个管线功能。
### 可变分自编码器(VAE)
![vae_diagram.png](data:image/png;base64,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)
可变分自编码器(VAE)是一种模型,它可以将输入图像进行编码,得到一个“压缩过”的信息,再把这个“隐式的”压缩信息给解码出来,得到某种接近输入的输出。当我们使用 Stable Diffusion 生成图片时,我们先在VAE的“隐空间”应用扩散过程**生成隐编码**,然后**在结尾对它们解码**来得到最终结果图片。这样UNet的输入就不是完整的图片、而是缩小版的特征,可以极大地减少计算资源的使用。
下面代码就使用了 VAE,把输入图片编码成隐式表示形式,再对它解码:
```python
# Create some fake data (a random image, range (-1, 1))
images = torch.rand(1, 3, 512, 512).to(device) * 2 - 1
print("Input images shape:", images.shape)
# Encode to latent space
with torch.no_grad():
latents = 0.18215 * pipe.vae.encode(images).latent_dist.mean
print("Encoded latents shape:", latents.shape)
# Decode again
with torch.no_grad():
decoded_images = pipe.vae.decode(latents / 0.18215).sample
print("Decoded images shape:", decoded_images.shape)
```
Input images shape: torch.Size([1, 3, 512, 512])
Encoded latents shape: torch.Size([1, 4, 64, 64])
Decoded images shape: torch.Size([1, 3, 512, 512])
在这个例子中,原本 512x512 尺寸的图片被压缩到 64x64的隐式表示形式(有四个通道)中。每一空间维度上都被压缩到了原有的八分之一,这也是上文中为什么我们设定 `width` 和 `height` 时需要它们是 8 的倍数。
使用这些信息量充裕的 4x64x64 隐编码可比使用 512px 大小的图片要高效多了,可以让我们的扩散模型更快,并使用更少的资源来训练和使用。VAE的解码过程并不是完美的,但即使损失了一点点质量,总的来说也足够好了。
注意:上面的代码例子包含了一个值为 0.18215 的缩放因子,以此来适配stable diffusion训练时的处理流程。
### 分词器(Tokenizer)和文本编码器(Text Encoder)
![text_encoder.png](data:image/png;base64,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)
文本编码器的作用是将输入的字符串(文本提示)转化成数值表示形式,这样才能输入进 UNet 作为条件。文本首先要被管线中的分词器(tokenizer)转换成一系列的分词(token)。文本编码器有大约五万分词的词汇量 —— 它可以将一个长句子分解为一个个的文本最小单元,这些最小单元被称为token,在英文中,一般一个单词就可以作为一个分词,而在汉语中,一般一个或多个汉字组成的词语才是一个分词。这些分词的词嵌入(embedding)会被送入文本编码器模型中 —— 文本编码器是一个transformer模型,被训练作为 CLIP 这个模型的文本编码器。在这里,这个预训练过的 transformer 模型已经有了足够好的文本表示能力,能够很好地把文本描述映射到特征空间。
我们这里通过对一个文本描述进行编码,验证一下这个过程。首先,我们手动进行分词,并将它输入到文本编码器中,再使用管线的 `_encode_prompt` 方法,调用一下这个编码过程,该过程包括补全或截断分词串的长度,使得分词串的长度等于最大长度 77 :
```python
# Tokenizing and encoding an example prompt manualy:
# Tokenize
input_ids = pipe.tokenizer(["A painting of a flooble"])['input_ids']
print("Input ID -> decoded token")
for input_id in input_ids[0]:
print(f"{input_id} -> {pipe.tokenizer.decode(input_id)}")
# Feed through CLIP text encoder
input_ids = torch.tensor(input_ids).to(device)
with torch.no_grad():
text_embeddings = pipe.text_encoder(input_ids)['last_hidden_state']
print("Text embeddings shape:", text_embeddings.shape)
```
Input ID -> decoded token
49406 -> <|startoftext|>
320 -> a
3086 -> painting
539 -> of
320 -> a
4062 -> floo
1059 -> ble
49407 -> <|endoftext|>
Text embeddings shape: torch.Size([1, 8, 1024])
```python
# Get the final text embeddings using the pipeline's _encode_prompt function:
text_embeddings = pipe._encode_prompt("A painting of a flooble", device, 1, False, '')
text_embeddings.shape
```
torch.Size([1, 77, 1024])
这些文本嵌入(text embedding),也即文本编码器中最后一个transformer模块的“隐状态(hidden state)”,将会被送入 UNet 中作为 `forward` 函数的一个额外输入,稍后我们会详细看到。
### UNet
![unet.png](data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAjEAAAEFCAYAAAASbLbSAAAAAXNSR0IArs4c6QAABVp0RVh0bXhmaWxlACUzQ214ZmlsZSUyMGhvc3QlM0QlMjJhcHAuZGlhZ3JhbXMubmV0JTIyJTIwbW9kaWZpZWQlM0QlMjIyMDIyLTEyLTE5VDA0JTNBMzMlM0E0Ny44NTVaJTIyJTIwYWdlbnQlM0QlMjI1LjAlMjAoWDExJTNCJTIwTGludXglMjB4ODZfNjQpJTIwQXBwbGVXZWJLaXQlMkY1MzcuMzYlMjAoS0hUTUwlMkMlMjBsaWtlJTIwR2Vja28pJTIwQ2hyb21lJTJGMTA3LjAuMC4wJTIwU2FmYXJpJTJGNTM3LjM2JTIyJTIwdmVyc2lvbiUzRCUyMjIwLjcuMiUyMiUyMGV0YWclM0QlMjJjWFZ5OHhoenR5WUJIc3NFeXZKbiUyMiUyMHR5cGUlM0QlMjJkZXZpY2UlMjIlM0UlM0NkaWFncmFtJTIwaWQlM0QlMjI0WTc4ZDVQQlQ1OTlxOC10cHpIeSUyMiUyMG5hbWUlM0QlMjJQYWdlLTElMjIlM0UzVmZianBzd0VQMmFTTzNEVnR4aEg1dExiMnBYcmJLcjdqNVZEcDZBVzRPUmNSTFNyJTJCOFFUSURDSnEyMDJUUjlzdWVNamNkenp0aG1aRSUyQlM0cTBrV2Z4SlVPQWp5NkRGeUo2T0xNc3lmQWViRXRsV2lPJTJGYkZSQkpSaXZJYklBNSUyQndrYU5EUzZZaFR5emtBbEJGY3M2NEtoU0ZNSVZRY2pVb3BOZDloUzhPNnFHWW1nQjh4RHd2dm9WMFpWckZIVHUyNGM3NEJGc1Y0NnNQektrWkI2c041SkhoTXFOaTNJbm8zc2lSUkNWYjJrbUFBdmsxZm5wWnIzNWhIdlBqQUpxZnFUQ1IlMkJLMTVzdnQlMkZmeiUyRk9IR3VucXZ0a0ZpR1ZlbSUyRmt5dXR2V09nV0lDdENta2lrVWtVc0puRFRxV1lwVlNLRDlyb05XTSUyQlNoRWhxQ0o0SGRRYXF2WkpDc2xFSXBWd3JVWEk1YmJlejElMkZaenlVeGl2THJlMXAwZlpPdDlxcWdpMGpmRFFKOVliRVNvWndZT2V1RmhPUkVhaERHV3E0UXBHRFNBQUR3b2tTT0ZGczNRMkVhTFZGJTJCM0VOSWRqUm5Qd0ZQenJLTmVFcnZaS0NRbjJEWkFHVXNqUWFXUjdINk1jTGliMm83TDNBVUhBdHclMkZlcjFqUXM1MldQNUliQ2tvOU56QlRNTTdKTDJBWUx1VXZYa25FJTJCRVZ6STNWeWJFZ2lXSWVLNWt1SUh0RHhlR01CaXVlZHBEUktqUGN4VVA3RjZncVBMUnA4YmUzdlRxc0s2dE9KV0FYckdpYWd3N1hPVXloTXEzcjhJeGZzOXhhZUM1ZVg2dURydU9SJTJCUXZGTkozZHUzZmNFZmtUakpzJTJCcjJXTEtpcE9zOG12ZTZtcmV0dnVhREFja0hKNU84OHklMkZkRHY0ejNnN0JSZFJLMEw4ZFdBSzVncXpQRyUyQmY0WklMTE9PeWRBZUVQSHZiT3laVHZYZmhoWDclMkJmanl2WVBhZUM2ekJiRXI2N21kMzJrbyUyQlA1NnpzN3JSdGp6T1FLSFFGVWtPZkclMkZ1NHZ2R0Exd3dNNkIxTTZvSSUyRnBQZHJ6N2VKOXpSNiUyRiUyRjJrdHdLM0wzaCUyRlFQQ21lekxGOTUlMkJhNWNVTDViNGtVQllxSnRMJTJGJTJCTzUxZ21lN2ZORnNmdnQydnRiUHN6MzdCUSUzRCUzRCUzQyUyRmRpYWdyYW0lM0UlM0MlMkZteGZpbGUlM0WsljKvAAAgAElEQVR4Xu2dCXhN19fGX5knGQRB1BRBIighVFpDUFU1VNXUqmqJVlutUmpqqzWUoqMOSktNpdqaU2NQMYSYEgkiJAgJiQwyR+J79vHJP1Hh5t5zknPufc/z5CHu3uus9Vvr9rzdZw+V7ty5cwe8SIAESIAESIAESEBjBCpRxGgsY3SXBEiABEiABEhAIkARw0IgARIgARIgARLQJAGKGE2mjU6TAAmQAAmQAAlQxLAGSIAESIAESIAENEmAIkaTaaPTJEACJEACJEACFDGsARIgARIgARIgAU0SoIjRZNroNAmQAAmQAAmQAEUMa4AESIAESIAESECTBChiNJk2Ok0CJEACJEACJEARwxogARIgARIgARLQJAGKGE2mjU6TAAmQAAmQAAlQxLAGSIAESIAESIAENEmAIkaTaaPTJEACJEACJEACFDGsARIgARIgARIgAU0SoIjRZNroNAmQAAmQAAmQAEUMa4AESIAESIAESECTBChiNJk2Ok0CJEACJEACJEARwxogARIgARIgARLQJAGKGE2mjU6TAAmQAAmQAAmUi4i5npKFI1GJiIxNQWJKJnLzCkhepQSsrczh5mIP73ouaOPlhuoudir1lG6RAAmQAAmYOgHFRcy64PM4FHkNPg1rol6tKqjqbA9rKwtT567a+HPzbiMpNROxV28i4vw1tPOuif6dG6rWXzpGAiRAAiRgugQUEzFpmXlYtCECTo52aN+iPoWLBmtMCJoDJy8iLT0LgX184GRvpcEo6DIJkAAJkICxElBMxHyx6hjc3VzQxqeOsbIzmbiORFxCfGIKPhjSymRiZqAkQAIkQALqJ6CIiBGvkG5m3kbnNp7qJ0APdSIQfCQaVewt+GpJJ1psRAIkQAIkUB4EZBcxYhLv3FVheLW3H18hlUcGy+ke4tXS0o2hmDDEl5N9y4k5b0MCJEACJPBwArKLmC0HLuLGrQL4P16f7I2MQMiJi6hW2Rw92zO3RpZahkMCJEACmiQgu4gRc2HaNKsH9+pOmgRCp0snEH89DUfCYzk3hkVCAiRAAiSgCgKyi5jxC//lqyRVpFZ+J+69Upr31lPyG6dFEiABEiABEigjAdlFzDtf7sGYIXzIlTEPmmn+zap/8e3YTprxl46SAAmQAAkYLwGKGOPNrSKRUcQogpVGSYAESIAE9CBAEaMHNFPuQhFjytln7CRAAiSgLgIUMerKh+q9oYhRfYroIAmQAAmYDAGKGJNJtTyBUsTIw5FWSIAESIAEDCdAEWM4Q5OyQBFjUulmsCRAAiSgagIUMapOj/qco4hRX07oEQmQAAmYKgGKGFPNvJ5xU8ToCY7dSIAESIAEZCdAESM7UuM2SBFj3PlldCRAAiSgJQIUMVrKlgp8pYhRQRLoAgmQAAmQgERAtSLmqzkfo027p+DfsateqTK0v643PRF2GBPHvIZtIad17fLAdl7u1jgcmQhHJ+cSn4cdDsH0ye9g465jWPL9fFw4fxYzFywy6F6GdKaIMYQe+5IACZAACchJQLUiZlDvDhg9dgo6dO6uV7yG9tf1puUpYtJSbyIvLw/VqtfQ1T3Z21HEyI6UBkmABEiABPQkoEoRs3LpD5g3YzLcarjjo1lfo61/J3w99xPs3bVVCtPD0wsfz/4GSTcSMfC5p7D8z13w8mmBCe+8CgsLSzRt0apE//YdupSKJzL8BGZ99D4yMtJRCZUw6t0P8cxzLyDiZBimjh8l3Uvc51Z6GoaPeg//bP4TVy5dRLPH22DGvB8hRMyYEQPRsnU7XLxwDvYOlTF9zvdo1KQpCgoKHui3k3MV7Nv9D+Z+OhEWlpZ43Lcd1iz/uWgkZtF3c/HHyl9Q2dEJjZr4IDLi+H9GYnwbuWLIsDdwNioc8ZdjMXjYKLz82ltSnIu/n4c1yxfD3sEBfk90xI6t6xF8NEbPEinZjSJGFow0QgIkQAIkIAMBVYoYEVfxkZTfFn+LiFPHMPvLxTA3N8fPC79A9JnTmPvtUqxdsRjLf1mIV0a8A9Huj60HYGNjW6J/aZyys7PwXKfH8eNv6+HZ2BspN5PQv0d7LF61CZkZGXjx2fbYsOuYJEgmjBmOC9FnsGbzfhQWFqBDy3pYvWEPUlNTMPC5J/HbnzvRtn1HbPprNb7/cha27juF5Uu+e6DfU2d8hQC/hlj6x3b4NG+FP1YuwZRxo3D0bBJOnzom/f3vHaFwdHTGxDHDcSby1H9ETIsGTpjy6XwMeHkE4q/EoXt7bxw5c10SVR9NGI11QQfg6OQi/X1/8HaKGBm+LDRBAiRAAiSgLgKaEDHDBz6D2AvnUdnRUaJXcPs2rG1s8de2w9Lv74wYgL07g7Bx9zHUa+Ap/Zsur5OOHTmIYf27oX7DRkVZuZWWhg+mzUadeh4YM3IQdodGS599+flHyMrMwJTPFki/9+zYHDMX/Cz9fdoHb2DT7uNFNlrUd0TQ/ghMeT/wgX6/O+ETLJg9FRt2hkl9CgsL0ayuAw5GXMWPX3+O27fzMfnT+dJnx0IP4JNJbz9QxAih5F67rtROiJpdh8/h54XzYGVphXFTZkr/fi4qAqOG9qGIUdf3jt6QAAmQAAnIQEATIuaV/t3Qq99gvDjkNSlkMYIiBIVr1erIy8tF/x5PIPnGdUyd+RV69Oqvs4g5eng/3n/jZew7HluE8kbiNThXccXZyHCMG/1K0YRdIWJyc3Pw4cdz/yNiPpvyLv7851CRjZYNXbD9QBTGjR76QL/Fq6r5s6ZIwuTeJYTPvycuSSImPz+vSCydOn5Eeq11/8ReIVp2HjpbND/m3u9Lf/oalSpVwvips+6KmDOnMerl3hQxMnxZaIIESIAESEBdBFQrYl5+PgCvBr6Lrj36SK+Ptq5fixV/75bmnHzy4dtIS03Blz+uxGdT3oOY8Br49gS8OqA7/gg6II1OFO9fGvLMjFvo2q4xJk2fh94vDEHcxfPo170t1m7Zj+ysLJ1FjLiXGBURozd7dwXhy8+nYf2Oo6X6/enc7xHg54lFKzZKc2m2bliL90YNkV4niVdHE95+VXqd5FKlKqZ/+A7CjoToLGJiL0Rj0nsj8Oc/ByHm3gg+u7dvRvCR87JUHufEyIKRRkiABEiABGQgoFoR88sPC/Dt/M8wbeZX6N1vCObPnorg7VtgZmYmvf6ZtWARjh89hI8+GI3Ne05IS5PFhNhd/2zEqvV7sOznb4r69xs4rFRUYg7J55+MR0bGLdwpLMQb706SRk/EaImuIzFivoyYN5Oflye9Cvpo1jeoW78hbufnP9BvIS4OH9iLmdPGSqMmYpLwP5vWYXfoeSkOMZdm+ZKFcHCojOat/HDsyAGdRYxYuSRiX7X0J1jbWKNJ0xaIOBEmiSw5LooYOSjSBgmQAAmQgBwEVCti5AjOFG2ICdBHD/0rjWKJ69efvsLJsMP4atFqWXBQxMiCkUZIgARIgARkIGD0ImZn0Aas/u2nB6Lq1qMvBr0SKANG9ZgQr8jEhOLos6elUZ6a7nWkpeBuNd1lcZIiRhaMNEICJEACJCADAaMXMTIwooliBChiWA4kQAIkQAJqIUARo5ZMaMQPIWK+e7+zRrwt3c1OnTohODhY83EwABIgARIwZQIUMaacfT1iN5aRGPGq7c6dO3oQYBcSIAESIAG1EKCIUUsmNOIHRYxGEkU3SYAESMAECFDEmECS5QyRIkZOmrRFAiRAAiRgCAHNi5jZH49HQPde0rlF965pH7yJ2nXqYdQ7Ex/KRmzJ/2JP/xJtsrMyseCHFUhLS8HcTz8s+kzsIZOTk439Jy6heo1aj2QuzlDasG4FFq/aUtR2z86tmDdjErKzs+Hd7HHMWvCzdMij2E9m5rT3ceDf3VLbJt7N8OkX30ub1d27xKGTz3fzk85oEoddZtxKx3ujBuP7pX/Bysr6kf7I1YAiRi6StEMCJEACJGAoAU2LmCOH/pUOffx28VqJg9hxd/qkdxB2OASj35/ySBFzPzxxeva2zX/h19+DYG5hUeLj9wIHo4FnE4z54OOHMhfiQhxR8NeaZWjV5gksWX335O2byTfwbIfm+H3TPul8p/mzpkoHTorlz7/++CVC9u3Cj8v+lu778cS3pPkan33xg9RXHHcw7MWnceb0KazesFcSMeL6Y9UvuBx3Ee9P+szQOtC5P0WMzqjYkARIgARIQGECmhYxQ/t1wZvvTUb7Dl0kTJ9Pn4DGXj44dfwoatRyL5OIuXD+DF7qG4D1O478Z0+Vdat/xe+/LZJOsBanaD/s+nvtb4i7GCMJlU1/rSoSMRv/XCWdcP3zyk1S94RrV9DdvymOn0/ByWOh0nEKYtdfcf35+1Js3/I3flq+Qfp98tiRaOXXHgvnz5BGXu6JGHGGlDg2YevekyVGbZSsGYoYJenSNgmQAAmQQFkIaFbEpKYko33z2gi/eOs/oybTJ40ps4gZ+VIvtHniKekMpuKXEAoBbRrih2V/43Hftjqz/Wfzn/hj5ZIiEfPTt3MQfzkO4twkcYmRlsY1LXHkzPUSAkSMzgx87imMmzwT3Z/rJ9kQp23P/moxOrf2KCFihJ0RQ56Tjkno0/8lnX0zpCFFjCH02JcESIAESEBOApoVMeLMo7GjhjzwdOayipjos5EY+NyTCDl5GbZ29iX4rvz1e2zb8jd+W7ejTNzvFzE/fj0bV+Mv/0fEhJ1LlubFiCs25hxGD++PF4cMx/A3xkIcIfDJxLekgy9tbGwfKGLEAZGOzi4Y++GnZfJP38YUMfqSYz8SIAESIAG5CWhWxIhXMB+8PQzbD0T9h0lZRYyYn5J0PUEa7bj/EuJm8Cuj0HfA0DKxv1/EbFi3Ujqt+t4rosRr8ejRoRmEiBF7lvwbvB1Tx42SDrwUJ3eLS0z2Xf/HClhaWkq/i3k1jk4umPXlz+j6TO+iNkIEPWquTpmcf0hjihi5SNIOCZAACZCAoQQ0K2LSUm+iQ6v6OHr2BiwtrUpwuF/E5Ofn4eqVS9LJ0g+6BvT0x8ChI/HCoFdLfCxWKrXyrIIdB8+gdp36RZ89yp5oeL+ISU66Lk3sXbUhGB6eXlgwexquxV/CF98tk14XiYnDv6wJQsNGXqXm9EGvk94c9jx69h2I554fZGgt6NSfIkYnTGxEAiRAAiRQDgQ0K2IEG7FiRyyjvjex9x6v+0VMVMRJ9Onqi6grOf+ZPyP6tG5cFUv/2A6f5q1KIBergQb36Yhj0Tel0ZJ716PsPUjEiH/buysIX8yYhNycHElQzVu4DM4urhjU6ymcPH6kxFLpps1bYtX6PSX8uV/EiFVLAX6e2Lb/NBwqO5ZDuQAUMeWCmTchARIgARLQgYCmRYxYYi2WJ4sVO4+6Jr034oGvix7Vr7TP5banjx9i0u+luIsYN3mGPt316kMRoxc2diIBEiABElCAgKZFjOAxY+pYaQ5JO/9OpeIRr57EROCOXXrIglBue/o4JfajGTNyoCTgxKTf8rooYsqLNO9DAiRAAiTwKAKaFzGPCpCfy0uAIkZenrRGAiRAAiSgPwGKGP3ZmWRPihj1pL0gvwDrf9iF65eS1eMUPTEKAtXruKLvm11gbvnwzT2NIlgGoWkCFDGaTl/5O08RU/7MH3bH9Qt3orZnTVSv7aoux+iNpgkc2X4Krbv7oK7Xo8+J03SgdF7zBChiNJ/C8g2AIqZ8eT/qbkd3RCAlIR2NfRs8qik/JwGdCcScugQrWwv49ym5YlNnA2xIAuVEgCKmnEAby20oYtSVyYTYJASvPQz/53zV5Ri90TSBlOtpOHP0AgaOl2cxhKZh0HlVE6CIUXV61OccRYz6crJk6jp06OsHa7uSmz6qz1N6pCUCO1btx+AJPWHnWH6rH7XEh76qgwBFjDryoBkvKGLUl6qgX/bBuZoT3D3c1OccPdIsgeN7I+HdtgE8W9XTbAx03PgJUMQYf45ljZAiRlacshiLCInGpTPX4PNEI1ns0QgJCAJxUfHIy8tHl8HtCIQEVEtAdhEzfuG/eLW3H6ytLFQbNB3Tj0Bu3m0s3RiKeW89pZ8BFfUSx0jcuXNHRR7p70pKYjo2LQpGpxfa6m+EPUngPgIZqVkI2x2OoVPvHkjLiwTUSEB2EfPFqmNo06we3Ks7qTFe+mQAgfjraTgSHosPhmh/xYIxiRiR0uUzNsA3wAcOzvYGZJhdSaAkgT1/HkavwM5wcSufs9nInwTKSkB2EbPlwEXcuFUA/8f/d+pzWZ1ie3USCDlxEdUqm6Nne+3n1thEzK7Vh2BlZYm6Xu7qLB56pUkCEQfPoU6TmvDx99Sk/3Ta+AnILmKup2Rh7qowvlIystq59yppwhBfVHex03x0xiZioo/FIvLwBbTs6K353DAA9RCIj0lEyo00PPtaB/U4RU9IoBgB2UWMsL0u+DxuZt5G5zZU78ZSbcFHolHF3gL9Ozc0ipCMTcRkpWdj9dwt6DbkSaPID4NQB4Hc7Dzs+zsUr8/orw6H6AUJ3EdAEREj7iHmxri7uaCNTx1C1ziBIxGXEJ+YYhRzYe6lwthEjIhrzbwgNGndAC6cj6bxb5y63A/ZHIbOA9qiRr2q6nKM3pAAAMVETFpmHhZtiICTox3at6jP1UoaLDfxCunAyYtIS89CYB8fONkbz2ZqxihiQjYcQ152Pjya19VgtdFltRIQO/dWqemI1t181Ooi/TJhAoqJmHtMxaulQ5HX4NOwJurVqoKqzvYUNCouOCFcklIzEXv1JiLOX0M775pG8wqpOHZjFDGXoq4idFs4/J5uoeIKo2taI3D9SjKuRF9D37e6as11+msCBBQXMYKhmOx7JCoRkbEpSEzJRG5egQmg1WaI1lbmcHOxh3c9F7TxcjOKSbwPyoQxipiC/AIsmrQWPYZ1hJm5mTYLkF6rjkBhQSGClu1F4OwBMLc0V51/dMi0CZSLiDFtxIxejQSMUcQIzusX7kRtz5qoXttVjdjpk0YJhG4/iTbdm6GuVy2NRkC3jZUARYyxZpZxPZSAsYqYozsicPNaGpq09mAFkIBsBGJOxcHK1hL+fbS/0aVsUGhIFQQoYlSRBjpR3gSMVcQkxCYheM1h+PfyLW+kvJ8RE0i5noYzR2MwcPyzRhwlQ9MiAYoYLWaNPhtMwFhFjACzZOo6dOjrB2s741lNZnDCacBgAjtW7sfgiT1h52hrsC0aIAG5CFDEyEWSdjRFwJhFzNZf9sGlmhPcPdw0lRM6q24Cx/dEwrtdA3i2qqduR+mdSRGgiDGpdDPYewSMWcREhETjUtQ1+LRvxISTgGwE4qLikZeXjy6D28lmk4ZIwFACFDGGEmR/TRIwZhGTkpiOTYuC0emFtprMDZ1WJ4GM1CyE7Q7H0Kl91OkgvTJJAhQxJpl2Bm3MIkZkd/mMDfAN8IGDsz2TTQKyEdjz52H0CuwMFzdH2WzSEAkYQoAixhB67KtZAsYuYnavPgRLK0vU9XLXbI7ouPoIRBw4hzpeNeHjz8N91Zcd0/SIIsY0827yURu7iIk+FovIQxfQspO3yeeaAOQjEB+TiJQbaXj2tQ7yGaUlEjCAAEWMAfDYVbsEjF3EZKVnY/WcLej20pPaTRI9Vx2B3Kw87Fsfitdn9Fedb3TINAlQxJhm3k0+amMXMSLBa+ZtlXbudanuZPL5JgD5CIRsCkPngW1Ro15V+YzSEgnoSYAiRk9w7KZtAqYgYkI2HENuVj4atqir7WTRe1UREDv3utZ0gm83H1X5RWdMkwBFjGnm3eSjNgURcynqKkK3hcPv6RYmn28CkI/A9SvJuBJ9DX3f6iqfUVoiAT0JUMToCY7dtE3AFERMQX4BFk1aix7DOsLM3EzbCaP3qiFQWFCIoGV7ETh7AMwtzVXjFx0xTQIUMaaZd5OP2hREjEjy+oU7UduzJqrXdjX5nBOAfARCt5+EX/dmqONVSz6jtEQCehCgiNEDGrton4CpiJiwnRFIvpomTfDlRQJyETh/Mg7Wdpbw79NKLpO0QwJ6EaCI0QsbO2mdgKmImITYJASvOQz/Xr5aTxn9VxGBlOtpEBN8B45/VkVe0RVTJEARY4pZZ8wwFREjUr1k6jp06OsHazsrZp4EZCOwY+V+DJ7YE3aOtrLZpCESKCsB1YuYd77cU9aY2P4+At+O7UQm9xEwJRET9Ms+OFdzgruHG+uABGQjcHxPJLzbNYBnq3qy2VS7IT6PDM+Q3M8jTYiYMUOeMpyciVr4ZtW/kLtojAGlKYmYiJBoiOXWPu0bG0PqGINKCMRFxSM/Px8Bg9qpxCPl3RAihs8j/Tkr8TyiiNE/H5roqUTRaCLwRzhpSiImJTEdm37ajU79TedhYww1qvYYMlIzEbY7AkOn9lG7q7L5RxFjGEolnkcUMYblRPW9lSga1Qetg4OmJGIEjuUzNsA3wAcOzvY60GETEtCNwJ51h9BrVABc3Bx166DxVhQxhiVQiecRRYxhOVF9byWKRvVB6+CgMYqYkSNH4tSpU1i4cCFat25dgsLu3w/B0tISdb3cdaDDJiSgG4GIA2elvWJ8/D1166DxVhQxhiVQiecRRYxhOVF9byWKRvVB6+CgMYqYFStWIDAwUFp51bVrV0ybNq1IzEQfi0XkoRi07NRUBzpsQgK6EYiPSUTqjTT0eK2Dbh003ooixrAEKvE8oogxLCeq761E0ag+aB0cNEYRI8Ju3Lgxzp07JxGwtbVFt27dJDHj3agpVs/Zgm4vPakDHTYhAd0I5GblYd/6ULw+o79uHTTeiiLGsAQq8TyiiDEsJ6rvrUTRqD5oHRwUIuaTTz7RoaW2mojXSVu3bkVOTk6R40LMNGjQANOGzUWTNh5wqe6kraDoraoJhGwKQ+eBbVGjXlVV+ymHcxQxhlFU4nlEEWNYTlTfW4miUX3QOjg4ffp0HVppr0l4eDi2bNnyQBHzw2dLkZ9TgIYt6movMHqsWgJi517XWk7w7eqjWh/lcowixjCSSjyPKGIMy4nqeytRNKoP2oQdLO11kpjoK/aKCd0WDr+nW5gwIYYuN4HrV5JxJfoa+r7VVW7TqrNHEWNYSpR4HlHEGJYT1fdWomhUH7SJOigm9ooVSuJV2b25MMVXKRXkF2DRpLXoMawjzMzNTJQSw5abQGFBIYKW7UXg7AEwtzSX27yq7FHEGJYOJZ5HFDGG5UT1vZUoGtUHbaIOjhgxAuJ10oOWWN9Dsn7hTtT2rInqtV1NlBLDVoJA6PaT8OveTFpubcwXRYxh2VXieUQRY1hOVN9biaJRfdB0sFQCYTtPI/lqKpq09iAlEpCNwPmTcbCxt0T73q1ks6lGQxQxhmVFiecRRYxhOVF9byWKRvVB08FSCSTEJiF4zWH49/IlJRKQjUDK9TScOXoBA8f3kM2mGg1RxBiWFSWeRxQxhuVE9b2VKBrVB00HH0pgydR16NDXD9Z2ViRFArIR2LFyPwZP7Ak7R1vZbKrNEEWMYRlR4nlEEWNYTlTfW4miUX3QdPChBIJ+2Qfnak5w93AjKRKQjcDxPafh3c4Dnq3qyWZTbYYoYgzLiBLPI4oYw3Ki+t5KFI3qg6aDDyUQERKNuKiraNa+MUmRgGwE4qLikZ+fj4BBxntaOkWMYeWixPOIIsawnKi+txJFo/qg6eBDCaQkpmPTT7vRqb/xPmxYAuVPICM1E2G7IzB0ap/yv3k53ZEixjDQSjyPKGIMy4nqeytRNKoPmg4+ksDyGRvgG9AMDs52j2zLBiSgK4E96w6h16gAuLg56tpFU+0oYgxLlxLPI4oYw3Ki+t5KFI3qg6aDjySw+/dDsLS0RF0v90e2ZQMS0JVA+IGzqOtVCz7+nrp20VQ7ihjD0qXE84gixrCcqL63EkWj+qDp4CMJRB+LReShGLTs1PSRbdmABHQlEB+TiNSkNPQY3kHXLppqRxFjWLqUeB5RxBiWE9X3VqJoVB80HXwkgaz0bKyeswXdXnrykW3ZgAR0JZCblYd960Px+oz+unbRVDuKGMPSpcTziCLGsJyovrcSRaP6oOmgTgTWzAtCk9YN4FLdSaf2bEQCuhAI2RSGzgPboka9qro011QbNYuYiJNheG/UEOw8dFa1TJV4HlHEqDbd8jimRNHI4xmtVDSBAxuPIyczDw1b1K1oV3h/IyJw5kgMXN2d4dvV+F5VqlnE5Obm4HrCVTxWt4Fqq0mJ5xFFjGrTLY9jShSNPJ7RSkUTEHvFHNkWDr+nW1S0K7y/ERG4fiUZV6Kvoe9bXY0oqruhlJeIWfL9fJw6fgRCmKSlpcDSwhJffLcMbjXdcezIQXz+yXhkZWVK//7uxOno1PVZFB+JOXfmNKaNH4X8vDwU3inE0NffxguDXkVBQQG+nvsJ9u7aKsXj4emFj2d/AyfnKuWSKyWeRxQx5ZK6iruJEkVTcdHwznISKMgvwKJJa9FjWEeYmZvJaZq2TJhAYUEhgpbtReDsATC3NDcqEuUpYn5f/jM2BR+HjY0tJr03AtXcamLkW+PR3d8b3yxeg9Ztn8SF82cwqFdHrN2yHxm30oteJ00YMxzNH2+Nl197C/FX4vDV5x9JIui3xd8i4tQxzP5yMczNzfHzwi8QfeY05n67tFzypMTziCKmXFJXcTdRomgqLhreWW4C6xfuRG3Pmqhe21Vu07RnwgRCt5+EX/dmqONVy6golKeIiT4bic+/XiLxEyMzl+IuoEev/pgzfQL+3nGkiOvbr7+I9k91QfOWbYpEzM6gDZg0diTaPxWAJ54KQPfn+sGlSlUMH/gMYi+cR2XHu/v4FNy+DWsbW/y17XC55EmJ5xFFTLmkruJuokTRVFw0vLPcBMJ2nkZSfCq82njIbZr2TJjA+ZNxsLG3QvveLY2KQnmKmLjYGHw69/siESN+796zH+bPnIy/tocWcX3rtf5o274jWrVpX2Ji783kG9i/ZwdC9u7E/j3bsXnPCbwbOBi9+g3Gi1cLWs4AACAASURBVENek/pnZ2chKzMDrlWrl0uelHgeUcSUS+oq7iZKFE3FRcM7y00gITYJwWsOw7+Xr9ymac+ECaRcT8OZsAsYOK6HUVGoaBEzfsosPN3eC9/98of0Oik25hxe6PEEVm/ch7zcnCIRM2bEQHTo8gz6Dx4uzavp2q4xflj2Nw7+uwtb16/Fir93w96hMj758G2kpabgyx9XlkuelHgemZyIkWsZWtjhEHw04U1s2Xvqocn/as7HaNPuKfh31G+S2507dzCsfzf8vGozrK1tylxoShRNmZ1gB1UTWDJ1HTo87wdrWytV+0nntEVgx8r9GDyxJ+wcbbXl+EO8rWgRI0Zmjh7eL71Sys7KRCUzM7z53mQ82/vFkhN7oyIwdfwo5ORkw8zMDJ269sR7E6fjdn4+5s+eiuDtW6R/r9+wEWYtWMSJvUpWqNxFI9cyNF1FzKDeHTB67BR06NxdL0zinaVXbRuEx2VQxOhFkJ0eRSDo131wruoEdw+3RzXl5ySgM4Fje06jaTsPeLaqp3MftTeU+3mk9njl9k+J/6nW9EiMEsvQ+g0chu5PNsXHs74pGj2Z8n4gGnn5YNjIMUU5LS5ibqWnYdoHbyL+cixSbibDpYorvvppFfbsCsK8GZPhVsMdH836Gm39O5W6vM23kSuGDHsDZ6PCJTuDh42SZpZPHjsS61b/Kk3aWrJ6CxZ/vwDBOzZLS+vEbPU53/wiTdgq7VKiaOQubNqrWAIRIdGIi7yKZv6NK9YR3t2oCMRFxSM/Px8Bg4zntHSKGMNKVInnkeZFjBLL0JYu+hrHjxzE1z//jsyMW+jSthG2H4iCo5PzA0XMn78vxdnIcEz+dL70+cQxr6G6W02MmzITxUdiHra8rUUDJ0z5dD4GvDxCWhLXvb03jpy5Disr66KRmBuJ1zCg55P493gczC0s8P2XM9Hs8TZ4qvPTFDGGfbdMundKYjo2LdqNTi8Yz8PGpBOqkuAzUjMRtjsCQ6f2UYlHhrtBEWMYQ4qY+/iJkRgllqGlp6UiwK+hJFz+2fQnwk8cxeyvFpe4+/2vk8Ta+5Nhh3Ap9gL27AySRnHE6EtxEfOw5W1CxGzddwrute/unip+33X4HFxcXItEjIW5BYb07SRtctSh8zPo2LUH/J54+EFrShSNYWXM3moksHzGBvgGNIODs50a3aNPGiWw589D6BUYABe3u0t6tX5RxBiWQSWeR5ofiVFiGZp4PTN1/Bvw8GyCTX+txieffye9zil+FRcxi7+fhw3rVuLl4aPRsJE3gndsQWZmhrQTYnER80r/bqUubxOiRZx5Ua16jSIRI36vUqVqiTkxYqLvyWOh0izzDetW4bnnB+LtcdM4EmPYd8vke+/+/RAsLS1R18vd5FkQgHwEwkPOoq53Lfj4e8pntAItUcQYBp8i5gEjMQ8SMYYuQ/Np3gpnTp/Cm6/2g7OzS4mNhe65UFzEvPx8AHr2HSjNYxGvn4b264JmLdtg+pyFEJ+9GvguuvboI+2OWNryttJETNVqbmhSywpHztyQNjuaPHYE1mzeD1tbO6z4ZSH+Dd6On5ZvoIgx7Ltl8r2jj8Ui8lAMWnYyvvNuTD65FQggPiYRqUlp6DH84SPGFehimW5NEVMmXP9pTBGjo4gxdBnavdv07dZammwr5qncfxUXMUJIzJz2PpycnaWJbF5NW+DK5Vgs+2M7fvlhAb6d/xmmzfwKvfsNKXV5W2kiRozMiM2MDu0PxrqgA/h77Qr8s/lPONg7wMbODp/O/QENG3lRxBj23TKZ3vPnz0dgYCAqV65cIuas9GysnrMF3V56UvMstu36B/O++Rw7NuwpEcvId4ajTSs/BA5/E936dEL9eg2w6OtfitrEXDyPPoN7IuLQWWRmZcK7TUNYWf132fm2v4Mx84vp2Hfgrv28vDxYWFjCzKyS9PvBnUdR1bWa5jnKEUBuVh72rQ/F6zP6y2FOcRulfT/u3VhtIsbL3RqHIxNxKTbmkSdYizbffDEd8xb+VmI5tuJQi92AIqYcaYuEv/x8F2w7ECmNemj1UqJotMqCfgO2trbSIXDjxo3D5MmTS4iZtfOD0Ni3AVyqO2kala4iJvbSRXz84Wd4eeArUrwPEjFH955EtUfsZto2oCW+nrMQ7dq01zQ3pZwP2RSGzgPboka90ldRKnXvstp92PdD2FKriLG2sXnkCdYH9u3Cl59Pwx9bD0gb4FXEiddKPI80PSemrAWqa3txyuea5T9LWz6L10BavpQoGi3zMHXfv/76a3z44YcoLCxEpUqVMHbs2CIxc2DjceRk5qFhi7uTy7V66Spi+vcZgG9++gobVm9BwwaeFDEKJTzqSAyqujvDt6v6X1U+7PthiIgRm6xOeGe4NGp+B3eQdCMRUz77EmLqwk/fzoEY2RcHMTZt0QpfL/q91K049u3+B3M/nQgLS0s87ttOek7dPxKTn5+Hzz+ZANFWjBD6te+ASZ98geefbotrVy9LJ16PGD2uaORG7EX2+fQJ0hEF5uZm0orXKZ8tgENlR5S29Ye+paLE84giRt9saKSfEkWjkdDpZikEXF1dcfPmTelT8brknpgZ9uIIROw7D7+nW2iana4i5rOpsxB24ig2BW3Ext+34nL8pf+8TnJwqCzxuXc1qNsAG9cEleDDkZiHl8v1K8m4cv4a+o7Wb9fy8i7G0r4fYuRy8uIwjBnyVJldEiLmxZ7+2LjrGDwbeyP04D6I/cfECthF383FmuWLsTn4uCROfv9t0QNPmp464ytp1ezSP7ZL4uePlUswZdwoHD2bVOJ1kph7Kc5LWrRiI8zNLfDO6y/iuX6D4excpWgkpvjO9T98NQsnwg5LRxmI9h9NGC0dDClW5Ja29YetnX2ZGYgOSjyPKGL0SoV2Oomi+e79ztpxmJ5WCAGxMqlf3/548clX0bKDNyysLCrEDzluuiN4G+Z8OQs7N+4tYe610a+gfbsnMeKVQGlOjBAxfr7tMOT1AfBq7C29Vrp/TgxfJxmekdt5txG88QDGzB6OrJwsww1WgAXx/ejduzdq+r+tt4iZMi4QG3aGFXkvRjn+3h6KoE3rEHMuCnO/XSp9VtpWHO9O+AQLZk8tsiFGU5vVdcDBiKslRIzo3/fFoejT/6USpIq/TiouYsS2HSNGj0fA089J7c9FRWDYi09Ldkvb+kMsONHnoojRh5qJ91GiaEwcqebDr1q1KpKTk6U4jHEkJuzEEYhJvMf+jSiRq2f7d8PrQwPxQp8Xi0SMmMeSkJiAZ/oF4O1R7+Gr7+eXmNhLEWN4uWttJKa074ehIzGT3x8pjcSIS2yV0aK+I4L+Dcfm9WuQcDVe2pJDXKVtxSGEx/xZU4psiLbCxr8nLpUQMSOGPIfeLwyRfsR143qC9Kd4XXVvTkxxETOo11MYNeZDdO7WU2onNm4Vq2qPnL0hiZgHbf1xbyuQslaHEs8jjsSUNQsaa69E0WgMAd0tRkC88580aZI0uddY58SI1ULtuvhixLBASbSIa8u2TZg0/QPs2XoANd1qlhAx4vNtO4Pw9gdvSOeTFV+dRBFj+NdHa3NiSvt+CBL6TuwVoqF/jyewdkuItOeY2OX91x+/xOY9J6U5McVFTGlbcYg5mgF+ntJropat22HrhrXSvJb7XycJuyH7duLHZeulAyLfeKUvOnV7VtrDbOa0sZIIKi5iFi6YgVPHj0ivk8Qcmo8nvoX01BR8tWg1RYzh5a9/0chxb2OwQRFjDFmULwZTWJ0kaJ06fRIzv/gU4ZGncKewEI0beWHCux+ifdu7S8jvvU4qvqJo8vQJ2Bi0oYSIsbWxLTEnRvSd+9kC9Co24Z9zYh5en1ydBEk0vDHseTR/vLU0alLZyRkz5/+EBg2b/EfEPOyk6cMH9kpCRPwPiJiA+8+mddgder7ESIzoP/ezDxGydwfu3IG0e/yk6fOQlnoTA3vd3a9nwffLiyb25uXlYu6nH0K8brpdcBvNWvhKu807OVehiJHjP736Kl+RSLFZ3RffLYWzi2uRK2I/F3HA4vdL/3qke+IogRlT3oU44NHeoTI+/3qJVHTFL6XtCaUtjlCv10C/HS8pYh6ZZpNqYAr7xJhUQlUeLPeJuZsgIWLGjX4F20JOqzxjyrqnxPPIaF8niSEycWDj0NffLsrK1o1/4NNJY9CqzROPFDHi/KSn23vhm8VrpPOJ1q5YjL/W/obfN+4rV3uxF6IxaewIrN5QcpKirqWmRNHoem+20w4B7tirnVxpyVPu2EsRU7xelXgeGaWIybiVju7+3tgWEimtdRdXTHSUtE6/38BXpOVnjxqJEWchiXOTFq/aLPUXa+9jL5yXlseVtz0xU1zMHn/YadWl/YdNiaLR0n9E6atuBHh2km6c2KpsBHh2Utl4GXtrJZ5HRilitm/5G78t/hYr/t4t1URWZgZe7tdFWsJ26ngodgZteKSIERveiddOuTk5OB1+HO6P1YVYpy9ETHnbExO9YmOiMXPBojLXuBJFU2Yn2EH1BHiKtepTpEkHeYq1JtOmmNNKPI+MUsSI2d4Xos9izjd3z0V5d+QgdHmmt7Tk7K81y3QSMfNnTpE2Hfrtz51o0rS59Drpp2/nYndodLnbE/sILPv5mxKvsnStMiWKRtd7s502CKQkpmPTot3o9EI7bThMLzVBICM1E2G7IzB0qrZ3PS8OW985mppIWDk4qcTzyChFjNgB8VLsBcyY9yOSk66jS9tGsPv/HQZzcrKRl5uLZi1bP3SeyfIl32H39s34dc0/UmrFxkJNH7PF5j0n8MIzT5SrPTFy9OtPX2Hl+uAyl5kSRVNmJ9hB1QQiQqIRF3kVzfwbq9pPOqctAnFR8dKBuAGDjEccU8QYVoNKPI+MUsSIh/5vS77Db+t2/If4/SMxYiMgseGXWE5W/BKvkoRYWb1xD+p7NIZ4RSWWre04eKbEkkul7QmfhIC5cP4sPvvihzJXkBJFU2Yn2EHVBIJ+3Qfnqk5w99BvF05VB0fnKozAsT2n0bSdBzxb1aswH+S+MUWMYUSVeB4ZpYjJzs5CQJuG0k6DYml08et+0TH61X7wbOKDsR9++p/siAOx5s+cjOzsbDhUrowZ836SXi2Vpz1xr9cG9ZAm9rbv0KXMFaRE0ZTZCXZQNYElU9ehw/N+sLa1UrWfdE5bBHas3I/BE3vCztFWW44/xFt9RQy3/LgLVYnnkVGKGAHru/mfSSuTXg1896FfoONHDyH6TAQGvDxCli+a3PZiY85h0vsjucRaluzQyP0EEmKTELzmMPx7+RIOCchGIOV6Gs6EXcDAcT1ks6kGQ/qKGG75QRFT5voVuxC+Oex5zFv4G1yqVC21v3j1JEY47OwdynyPB3WQ2957gYMxZsLH/9lkT1dnlVC+ut6b7dRPIGznaSTFp8KrjYf6naWHmiFw/mQcbOyt0L53S834rIuj+ogYbvnxP7JKPI+MdiRGl4I0hTZKFI0pcDOVGNcv3InanjVRvfb/drU2ldgZp3IEQrefRJvuzVDXq5ZyN6kAy/qIGG75QRGj19HnFVDfqrwlRYwq06IKpwryC7Bo0lr0GNYRZuZmqvCJTmifQGFBIYKW7UXg7AEwtzTXfkDFItBHxHDLD4oYihgD/jNAEWMAPCPvGhd1FUe2hcPv6RZGHinDK08C168k40r0NfR9q2t53rZc7qWPiOGWHxQxFDEGfD0pYgyAZ+RdQzYeR25mHhq2qGvkkTK88iRw5kgMXN2d4du1aXnetlzupY+I4ZYfFDEUMQZ8PSliDIBn5F3XzAtCk9YN4FLdycgjZXjlSSBkUxg6D2yLGvVKX1BRnv7IeS99RAy3/KCIoYgx4FtIEWMAPCPumpWejdVztqDbS08acZQMrbwJ5GblYd/6ULw+o39537pc7qePiBGOccuPu+lR4nnE1UnlUvoVdxMliqbiouGd5SIQfSwWkYdi0LKT8Q35y8WIdspOID4mEalJaegxvEPZO2ugh74ihlt+UMRooLzV6SJFjDrzUtFe7f79ECwtLVHXy72iXeH9jYhA+IGz0rJqH39PI4rqf6HoK2KMEoYeQSnxPOJIjB6J0FIXJYpGS/HT1wcTWD5jA3wDmsHB2Y6ISEA2AnvWHUKvUQFwcXOUzaaaDFHEGJYNJZ5HFDGG5UT1vZUoGtUHTQcfSiAlMR2bftqNTv2N53RhprziCWSkZiJsdwSGTu1T8c4o5AFFjGFglXgeUcQYlhPV91aiaFQfNB18KIGIkGiIPWKatW9MUiQgG4G4qHjk5+cjYJDximOKGMPKRYnnEUWMYTlRfW8likb1QdPBhxII+mUfnKs5wd3DjaRIQDYCx/echnc7D3i2qiebTbUZoogxLCNKPI8oYgzLiep7K1E0qg+aDj6UwJKp69DheT9Y21qRFAnIRmDHyv0YPLEn7BxtZbOpNkMUMYZlRInnEUWMYTlRfW8likb1QdPBUgkkxCYheM1h+PfyJSUSkI1AyvU0nDl6AQPH95DNphoNUcQYlhUlnkcUMYblRPW9lSga1QdNB0slELbzNJKvpqJJaw9SIgHZCJw/GQdre0v4924lm001GqKIMSwrSjyPKGIMy4nqeytRNKoP2kQdHDlyJE6dOoWFCxeidevWD6SwfuFO1Pasieq1XU2UEsNWgkDo9pNo072ZtEeMMV8UMYZlV4nnEUWMYTlRfW8likb1QZuogytWrEBgYCAqVaqErl27Ytq0aSXETEF+ARZNWosewzrCzNzMRCkxbLkJFBYUImjZXgTOHgBzS3O5zavKHkWMYelQ4nlEEWNYTlTfW4miUX3QJuxg48aNce7cOYmAra0tunXrViRmxLLqI9vC4fd0CxMmxNDlJnD9SjKuRF9D37e6ym1adfYoYgxLiRLPI4oYw3Ki+t5KFI3qg9bBwenTp+vQSntNxOukrVu3Iicnp8h5IWYaNGiAHz5bivycAjRsUVd7gdFj1RI4czQGrrWc4NvVR7U+yuUYRYxhJJV4HlHEGJYT1fdWomhUH7QODopXLp988okOLbXVJDw8HFu2bHmgiJk2bC6atPGAS3UnbQVFb1VNIGRTGDoPbIsa9aqq2k85nKOIMYyiEs8jihjDcqL63koUjeqD1sFBIWLu3LmjQ0ttNSntdZJ3o6ZYPWcLur30pLYCoreqJpCblYd960Px+oz+qvZTLucoYgwjqcTziCLGsJyovrcSRaP6oHVw0BhFjJjYK1YoidiKz4UROKKPxSLyUAxadmqqAx02IQHdCMTHJCL1Rhp6vNZBtw4ab0URY1gClXgeUcQYlhPV91aiaFQftA4OGqOIGTFiBMTrpActsd69+hAsrSxR18tdBzpsQgK6EYg4cBZ1vGrBx99Ttw4ab0URY1gClXgeUcQYlhPV91aiaFQftA4OGqOIeVjYy2dsgG9AMzg42+lAh01IQDcCe9YdQq9RAXBxc9Stg8ZbUcQYlkAlnkcUMYblRPW9lSga1Qetg4OmJGJSEtOx6afd6NTfeE8X1iHlbCIzgYzULITtDsfQqX1ktqxecxQxhuVGiecRRYxhOVF9byWKRvVB6+CgKYmYiJBoXIq6Cp/2jXUgwyYkoBuBuKh45OflI2Cw6YhjihjdaqO0Vko8jzQhYgzDxt7fju1ECPcRMCURE/TLPjhXc4K7hxvrgARkI3B8TyS82zWAZ6t6stlUuyEhYngZRkDu55HqRYxhuNibBB5MwJREzJKp69Chrx+s7axYDiQgG4EdK/dj8MSesHO0lc0mDZFAWQlQxJSVGNsbBQFTETEJsUkIXnMY/r18jSJvDEIdBFKup0Hs1Dtw/LPqcIhemCwBihiTTb1pB24qIiZsZwSSr6ahSWsP0044o5eVwPmTcbC2s4R/n1ay2qUxEigrAYqYshJje6MgYCoiZv3CnajtWRPVa7saRd4YhDoIhG4/iTbdm6GuVy11OEQvTJYARYzJpt60AzcFEVOQX4BFk9aix7COMDM3M+2EM3rZCBQWFCJo2V4Ezh4Ac0tz2ezSEAnoQ4AiRh9q7KN5AqYgYuKiruLItnD4Pd1C8/liAOohcP1KMq5EX0Pft7qqxyl6YrIEKGJMNvWmHbgpiJiQDceQm5WPhi3qmnayGb2sBMSEXteaTvDt5iOrXRojAX0IUMToQ419NE/AFETMmnlbpQm9LtWdNJ8vBqAeAiGbwtB5YFvUqFdVPU7RE5MlQBFjsqk37cCNXcRkpWdj9Zwt6PbSk6adaEYvK4Hc7Dzs+zsUr8/oL6tdGiMBfQlQxOhLjv00TcDYRUz0sVhEHrqAlp28NZ0nOq8uAvExiUi5kYZnX+ugLsfojckSoIgx2dSbduDGLmJ2rz4ESytL1PVyN+1EM3pZCUQcOIc6XjXh4+8pq10aIwF9CVDE6EuO/TRNwNhFzPIZG+Ab4AMHZ3tN54nOq4vAnj8Po1dgZ7i4OarLMXpjsgQoYkw29aYduDGLmJTEdGxaFIxOL7Q17SQzelkJZKRmIWx3OIZO7SOrXRojAUMIUMQYQo99NUvAmEVMREg0LkVdg0/7RprNDx1XH4G4qHjk5eWjy+B26nOOHpksAYoYk029aQduzCJm6y/74FLNCe4ebqadZEYvK4HjeyPh3bYBPFvVk9UujZGAIQQoYgyhx76aJWDMImbJ1HXo0NcP1nZWms0PHVcfgR0r92PwxJ6wc7RVn3P0yGQJUMSYbOpNO3BjFTEJsUkIXnsY/s/5mnaCGb2sBFKup0Hs1Dtw/LOy2qUxEjCUAEWMoQTZX5MEjFXEHN0RgZvX0qSdenmRgFwEYk7FwcrGEv59W8llknZIQBYCFDGyYKQRrREwVhGzfuFO1Pasieq1XbWWEvqrYgKh20+iTfdmqOtVS8Ve0jVTJGCyIiYtIxdpmXm4lZWHjKx8ZGTn41Z2PrJybiMrNx/ZubeRm1eA3PwC5OUXIv92AW4XFKKg8I70U1h4B3fu3CmqGfFQNDOrBPP//7EwN4OlhTmsLM1gbWkOaytz2FpbwM7aEnY2FqhsawkH8WNnicp2VnCyt4KTg7Up1mCFxGyMIqYgvwCLJq1Fj2EdYWZuViFceVPjI1BYUIigZXsROHsAzC3NjS9ARqRpAkYrYoQIuZGaXewnBzfTs5FyKxdCwFhZmsPBzgr2Nlaws7GEjY0VrK0sYGNlCRsrC+nvoo0QIuLHwsIMQpiYm5lJYqW0S4ibgsJCSfDcvn1X/IifvPwC5ObdRo70k3/37zl5yMrJR2aOEFJ5UhshZFwqW6OKoy2qOdugmrP48+6PEEG85CFgjCImLuoqjmwLh9/TLeSBRCskAOD6lWRcib6Gvm91JQ8SUB0BoxAxCTezEH8jA1duZOJqUgYSkrMkoeLqbAuXyrZwqmyLyvY2cHKwgaO9DSrbW0M8xNR2iZGdW5m5SM/MQVpGDm6JP28J4ZWN5NRsSeDUcLVDraoOqF3NHu7VHFCjip3awtCEP8YoYkI2HENedj48mtfVRA7opDYInAm7gCo1HNG6m482HKaXJkVAcyJGjGRcvJqOC9fScfFaOi4n3pJGSqpXcUBVZ3u4OttLfzo62BhdItMzcpCUmonk1Ezpz+s3M6RRnsfcKqN+TUc0qOmI+rUcpZEkXg8nYIwiZs28IDRp3QAu1Z2YfhKQjUDI5jB0HtAWNepVlc0mDZGAXAQ0IWLOX0nFmUupOHc5BRevpuExNyfUrOYIN1dH1KhaGbbWlnLx0Jyd7Nx8JCTdQmJyOq7dEKIuDfVrOaHRYy5oUscZDWs7ay6m8nDY2ERMVno2Vs/dgm5DniwPfLyHiRDIzc7Dvr9D8fqM/iYSMcPUGgFVipj824WIuJiM8JhkRMXehL2dFerUcEZtNxc8VoMP5UcV2eWEVFxJTMGlhFRkZuXBq14VNPNwhU99V1hacMKn4GdsIib6WCwiD19Ay47ejyoPfk4COhOIj0lEyo00PPtaB537sCEJlCcBVYmY0xeTcfTsDZw6fwO1qjmivrsr6rtXgYMdV+3oWxQZWbm4GH8TF+OTcfVGOpo3rIbWjauhaX3TXoJrbCJm1+pDsLKyRF0vd31Lhf1I4D8EIg6eQ50mNeHj70k6JKBKAhUuYlIzcnEwIgGHIxOkFUGedaujUd1qJv2KSKlKEa+ezsXdQHTcdWl1VFvvGnjCpwacTXBpt7GJmOUzNsA3wAcOzvZKlQ/tmiCBPX8eRq/AznBxczTB6BmyFghUmIi5mpSJPcfjcTDiKlo0qgmvBjWkybm8yoeAmBQcdSEBJ89dwxM+tdCppTtqVTWdB6AxiZiUxHRsWhSMTi+0LZ/i4V1MgkBGahbCdodj6NQ+JhEvg9QmgXIXMUlp2dgeehlhZxPh6+WO5o3dYW3J1TQVVT65+bdx6mw8wqLi4dvYDU/7PYaqTsZ/wJsxiZiIkGhcOnMNPk80qqgy4n2NkEBcVDzy8vLRZXA7I4yOIRkLgXIVMUGH4rDtcCxae9dG66Z1YM5dRVVTRwUFhTh6+hKORl5B97b10KOdce81YkwiJuiXfXCu5gR3DzfV1BMd0T6B43sj4d22ATxb1dN+MIzAaAmUi4iJS0jHH8ExsLa2xBMt6kubzvFSJwGxyd7BkxeRm5uPFzt7oG4N43wXbkwiZsnUdejQ1w/WdlbqLCp6pUkCO1btx+AJPWHnaPwjs5pMEJ2WCCguYg6EX8Pvu86ii58nvD1qELtGCETGJGBXaDQGdWmM9s1qasRr3d00FhGTEJuE4LWH4f+cr+7BsyUJPIJAyvU0nDl6AQPH9yArElA1AUVFzD+HL+FA+FV09/fipF1Vl8GDnROTf7eFRMG/WS10b1tHgxGU7rKxiJijOyKQkpCOxr4NjCo/DKZiCcScugQrWwv492lVsY7w7iTwCAKKiZidR6/gQMQ19O7kA3tbDnNrtRIzs/OwcU8E2vvURNfWWCCCLQAACwZJREFUtbUaxn/8NhYRs37hTtT2rInqtU173x+jKUyVBHJk+ym07u6Dul61VOIR3SCBBxNQRMRExt7E0q1RGND9cc5/MYLKE/Nk1m47gVef9YJ3vSpGEJFx7NhbkF+A9T/swvVLyUaREwahHgLV67ii75tdYG5prh6n6AkJPICAIiJm5m9H4OtdBx6PqefAsK/mfIw27Z6Cf8euWPL9fFw4fxYzFyxSrCjEidTD+nfDz6s2w9pa+xOZYy4nISzyEqa80kYxZuVp2FhGYsqTGe9FAiRAAmojILuIOXj6Gg5H3kDPDk1VFeug3h0weuwUdOjcHWmpN5GXl4dq1ZWbaFxw+za8atsgPC7DKESMSOaWfafR1rsanmiq/Ym+FDGq+nrSGRIgARLQi4DsIua7P0/Bo64bGqpoFGbl0h8wb8ZkuNVwx0ezvkZUxImikRjfRq4YNDQQZyJPIjHhKoa+9haOHt6Pq1cuoaCwAIuWb4SjkzMiw09g1kfvIyMjHZVQCaPe/RDPPPcCMjNu4YN3XkX85VjpUMFWbfyle0weOxLrVv+K5i3bYMnqLYi/fOmB/SNOhmHCO8PRsJEX7uAOkm4kYspnX8Knufom1J2/nISYuES8/UJzvYpNTZ0oYtSUDfpCAiRAAvoRkF3ETPw+BC8956u6s4+Kj8QUf53UooETJn40B0NefQN//r4Un04ag52Hz0mjNKNf7YeOXXqgd/+X8Fynx/Hjb+vh2dgbKTeT0L9HeyxetQnHjhzEnp1b8e3itcjPz8OUcaPw/qTPUK1ajaKRmMLCwlL7Z2Zk4MWe/ti465hkO/TgPkx5PxDbD0RJokhNV05uPlZsDsOc0f5qcksvXyhi9MLGTiRAAiSgKgKyi5hDUUlwreKiqiCFMw8TMUH/hqOWex0c2LcLn0//QBIU4po+aQxq1HJHm3YdpPkt9Rv+b1v3W2lp+GDabPi08MVLzwegQcPGaP9UAJ7u+TzqezRG8ddJp08dL7V/nXoemDIuEBt2hhUxE6NDf28PhfhMbdfevfswcmCA2twqsz+dOnVCcHBwmfuxAwmQAAmQgHoIyC5iJv4Qgpd6amskZuehs9LIixAx82ZMwl/bQ0uIGF8/f7z/xsvYdzy2KHM3Eq/BuYorLC2tkJubg0P7g3FwfzA2/LECC35YAb8nOhSNxISfOFpq/7OR4Zj8/sgi4SQmBLeo7wghrNwfU9d238Y0EqOeryA9IQESIAES0JeA7CLmu79OwaOOuubECDgvPx+AVwPfRdcefUqsThKvkx4lYl4ePhpd2zXGpOnz0PuFIYi7eB79urfF2i37EbJ3pzRf5vOvl0g5eGfEALT0bYfhb4xFk1pWOHLmBszNzUvtn52Vhf49nsDaLSHS/BnxSuvXH7/E5j0n9c2pYv2kOTGXEvF2P+3PiVEMEg2TAAmQAAmUGwHZRYzY4C40Sn2rk375YQG+nf8Zps38CinJSUUTe3URMaPemYgTYYfx+SfjkZFxC3cKC/HGu5PQq99gZNxKx4fvvY6Yc2dgY2MjvQKa9eXPsHeojLde6y+N0KwLOoC01NQH9hcTe98Y9jyaP94al2JjUNnJGTPn/4QGDZuUWxHoeiOxOsmvSTWjPIZAVwZsRwIkQAIkoB4CsosYEZrYJ6aVdx1VrVBSD/KSnggRM270K9gWclqtLkp+iVGYY0a0T4yqYdM5EiABEiABnQgoImJOX7yJZUHcsVeXDGhBxNzbsXfYs15oaiQ79uqSG7YhARIgARJQNwFFRIwIeVfYFYSE8+wkdaf/0d7dOzvJv1lNdPE1nrOTHh05W5AACZAACaidgGIiRgTOU6zVnv6H+3fvFOv2zWrhGSM7xVrbmaH3JEACJEACgoCiIkbcQEz0/X3nWQT4eaKph3Lb/DOd8hI4HZOA3aHRGNS1sXSCNS8SIAESIAESUBsBxUWMCDguIR1/BMfA2toST7Soz5Ot1VYFxfwR818OnryI3Nx8vNjZA3VrOKrYW7pGAiRAAiRgygTKRcTcA/zPoTj8czgWrb1ro3XTOjA3NzNl9qqKvaCgEEdPX8LRyCt4pm09PNOurqr8ozMkQAIkQAIkcD+BchUx4uZJadnYFnoZx84mwtfLHc0bu8Pa0oKZqSACufm3cepsPMKi4tGqsRu6+z2Gqk62FeQNb0sCJEACJEACuhModxFzz7WrSZnYcyIeB8OvokWjmvBqUAPVqzjo7jlbGkRATNqNupCAk+eu4YlmtdDpcXfUqmpvkE12JgESIAESIIHyJFBhIuZekKkZuTgQkYDQyARYW1mgUd3q8KxbTXWnYJdnUpS6V3ZuPqLjbuBc3HXk5t2Gn3cNtPepAWcHa6VuSbskQAIkQAIkoBiBChcxxSM7fTEZR8/ewKnzN1CrmiPqu7uivnsVONjxIatvBWRk5eJC/E3Exifj6o10NG9YDa0bV0PT+q76mmQ/EiABEiABElAFAVWJmHtE8m8XIuJiMk7FJONM7E042FmhTg1nuLu54LEazqoAp2YnLiekIj4xBZcSUpGRlYcm9aqguYcrfOq7wtKCk6nVnDv6RgIkQAIkoDsBVYqY+90/fyUVZy+l4uzlFFy8mobH3JxQo5ojarg6okbVyib96km8IkpIuoWE5HQk3EjH5cQ01K/lhMaPuaBxHWc0rE3Rp/vXgS1JgARIgAS0REATIqY40Jy827h4NR0XrqXj4rV0XEm8BQsLc2lScFUXe7g62aOqsz0cHWy0lAedfE3PyEFSaiaS0zKRnJKJxJsZuH27ALXdKqN+TUc0qOmI+rUcYWPF1V46AWUjEiABEiABTRPQnIh5EO2Em1mIv5GB+BuZiE/KQEJyFtIycuHqbAuXyrZwFD/2NtIme+LPyvbWqFSpkuoSd+fOHdzKzEV6Zg7EpnNCtKRnZCPlVjaSU7Ph5GCNGq52cK/qAPdq9nCv5oAaVexUFwcdIgESIAESIIHyIGAUIuZBoLJzb+NGanaxnxzcTBeCIFcSOFaW5tJcG3tbK9jZWMHG2lIawbC2uvenhdTG0uLuj4WFGSzMzWBuZgYzs9IFUGHhHRQUFuJ2QSFu3y5E/u0C6Scvv0BaESRGknLz8qU/c3LzkJWTD3HIopi7ItoIoeJS2RpVHG1Qzdm2xI+tNUdYyuNLwXuQAAmQAAlog4DRiphH4RdCJi0zD7eyhIDIR0b23Z/MnNvIys2HEEG5eQXIzRcC5K4YEcKkQBIpdyDEihg5uXeJkR0hbsz//0cIHiF+rCzNYG1pDmsrcwgRYmdtCXsbCzjYWt79sbNEZTsrONlbSQKGFwmQAAmQAAmQgG4ETFbE6Ian9FaSiAEkISMEjBibedgIjaH3Y38SIAESIAESIIGSBChiWBEkQAIkQAIkQAKaJEARo8m00WkSIAESIAESIAGKGNYACZAACZAACZCAJglQxGgybXSaBEiABEiABEiAIoY1QAIkQAIkQAIkoEkCFDGaTBudJgESIAESIAESoIhhDZAACZAACZAACWiSAEWMJtNGp0mABEiABEiABChiWAMkQAIkQAIkQAKaJEARo8m00WkSIAESIAESIAGKGNYACZAACZAACZCAJglQxGgybXSaBEiABEiABEiAIoY1QAIkQAIkQAIkoEkCFDGaTBudJgESIAESIAESoIhhDZAACZAACZAACWiSAEWMJtNGp0mABEiABEiABChiWAMkQAIkQAIkQAKaJEARo8m00WkSIAESIAESIAGKGNYACZAACZAACZCAJgn8H2uWjx1sO5DOAAAAAElFTkSuQmCC)
在扩散模型中,UNet 模型接收一个带噪的输入,并预测噪声,以此实现去噪。但与以往例子不同的是,这里的输入并不是图片了,而是图片的隐式表示形式(latent representation)。此外,除了把用于暗示带噪程度的 timestep 输入进 UNet 作为条件外,这里模型也把文字提示(prompt)的文本嵌入(text embeddings)也作为输入。这里我们用假数据试着让它预测一下,请注意这一过程中各个输入输出的形状大小:
```python
# Dummy inputs:
timestep = pipe.scheduler.timesteps[0]
latents = torch.randn(1, 4, 64, 64).to(device)
text_embeddings = torch.randn(1, 77, 1024).to(device)
# Model prediction:
with torch.no_grad():
unet_output = pipe.unet(latents, timestep, text_embeddings).sample
print('UNet output shape:', unet_output.shape) # Same shape as the input latents
```
UNet output shape: torch.Size([1, 4, 64, 64])
### 调度器(Scheduler)
调度器保存了如何加噪这一信息,管理着如何基于模型的预测更新带噪样本。默认的调度器是 `PNDMScheduler` 调度器,但你也可以用其它的(比如 `LMSDiscreteScheduler` 调度器),只要它们用相同的配置初始化。
我们可以画出图像来观察如何随着 timestep 添加噪声,看看噪声水平(基于$\bar{\alpha}$这个参数)是怎样变动的:
```python
plt.plot(pipe.scheduler.alphas_cumprod, label=r'$\bar{\alpha}$')
plt.xlabel('Timestep (high noise to low noise ->)');
plt.title('Noise schedule');plt.legend();
```
![png](01_stable_diffusion_introduction_CN_files/01_stable_diffusion_introduction_CN_35_0.png)
如果你想尝试不同的调度器,可以参考下面代码修改:
```python
from diffusers import LMSDiscreteScheduler
# Replace the scheduler
pipe.scheduler = LMSDiscreteScheduler.from_config(pipe.scheduler.config)
# Print the config
print('Scheduler config:', pipe.scheduler)
# Generate an image with this new scheduler
pipe(prompt="Palette knife painting of an winter cityscape", height=480, width=480,
generator=torch.Generator(device=device).manual_seed(42)).images[0]
```
Scheduler config: LMSDiscreteScheduler {
"_class_name": "LMSDiscreteScheduler",
"_diffusers_version": "0.11.1",
"beta_end": 0.012,
"beta_schedule": "scaled_linear",
"beta_start": 0.00085,
"clip_sample": false,
"num_train_timesteps": 1000,
"prediction_type": "epsilon",
"set_alpha_to_one": false,
"skip_prk_steps": true,
"steps_offset": 1,
"trained_betas": null
}
0%| | 0/50 [00:00<?, ?it/s]
![png](01_stable_diffusion_introduction_CN_files/01_stable_diffusion_introduction_CN_37_2.png)
在[这里](https://huggingface.co/docs/diffusers/api/pipelines/stable_diffusion_2#how-to-load-and-use-different-schedulers)你可以了解更多使用不同调度器的信息。
### DIY一个采样循环
现在我们已经逐个了解这些组成部分了,我们可以把它们拼装到一起,来复现一下整个管线的功能:
```python
guidance_scale = 8 #@param
num_inference_steps=30 #@param
prompt = "Beautiful picture of a wave breaking" #@param
negative_prompt = "zoomed in, blurry, oversaturated, warped" #@param
# Encode the prompt
text_embeddings = pipe._encode_prompt(prompt, device, 1, True, negative_prompt)
# Create our random starting point
latents = torch.randn((1, 4, 64, 64), device=device, generator=generator)
latents *= pipe.scheduler.init_noise_sigma
# Prepare the scheduler
pipe.scheduler.set_timesteps(num_inference_steps, device=device)
# Loop through the sampling timesteps
for i, t in enumerate(pipe.scheduler.timesteps):
# expand the latents if we are doing classifier free guidance
latent_model_input = torch.cat([latents] * 2)
# Apply any scaling required by the scheduler
latent_model_input = pipe.scheduler.scale_model_input(latent_model_input, t)
# predict the noise residual with the unet
with torch.no_grad():
noise_pred = pipe.unet(latent_model_input, t, encoder_hidden_states=text_embeddings).sample
# perform guidance
noise_pred_uncond, noise_pred_text = noise_pred.chunk(2)
noise_pred = noise_pred_uncond + guidance_scale * (noise_pred_text - noise_pred_uncond)
# compute the previous noisy sample x_t -> x_t-1
latents = pipe.scheduler.step(noise_pred, t, latents).prev_sample
# Decode the resulting latents into an image
with torch.no_grad():
image = pipe.decode_latents(latents.detach())
# View
pipe.numpy_to_pil(image)[0]
```
![png](01_stable_diffusion_introduction_CN_files/01_stable_diffusion_introduction_CN_40_0.png)
当然大多数情况下,还是使用现有的管线更方便。如果你想看看实际的代码以及深入了解如何修改这些组成部分,你可以参考 Stable Diffusion Deep Dive 这个[笔记本](https://github.com/fastai/diffusion-nbs/blob/master/Stable%20Diffusion%20Deep%20Dive.ipynb)和[视频](https://m.youtube.com/watch?v=0_BBRNYInx8)。这里有对相关内容更全面彻底的探究。
# 其它的一些管线
除了从文字提示生成图片外,我们还能做点什么呢?其实还有很多!下面还会向你展示几个管线,让你了解一些 Stable Diffusion 的应用。
## Img2Img
直到现在,我们生成的图片还都是完全从随机的隐变量来开始生成的,而且也都使用了完整的扩散模型采样循环。但如果使用 Img2Img 管线,我们其实不必从头开始。Img2Img 这个管线首先将一张已有的图片进行编码,编码成一系列的隐变量,然后在这些隐变量上随机加噪声,以这些作为起始点。噪声加多大量、去噪需要多少步数(timestep)决定了这个 img2img 过程的“强度”。只加一点点噪声(强度低)只会带来微小改变,而加入大量噪声并跑完完整去噪过程又会让生成出几乎完全不像原始图片,仅在整体结构上还有点相似。
这个管线无需什么特殊模型,只要模型的 ID 和我们的文字到图像模型一样就行,没有新的需要下载的文件。
```python
# Loading an Img2Img pipeline
model_id = "stabilityai/stable-diffusion-2-1-base"
img2img_pipe = StableDiffusionImg2ImgPipeline.from_pretrained(model_id).to(device)
```
Fetching 16 files: 0%| | 0/16 [00:00<?, ?it/s]
这里是使用该管线的代码:
```python
# Apply Img2Img
result_image = img2img_pipe(
prompt="An oil painting of a man on a bench",
image = init_image, # The starting image
strength = 0.6, # 0 for no change, 1.0 for max strength
).images[0]
# View the result
fig, axs = plt.subplots(1, 2, figsize=(12, 5))
axs[0].imshow(init_image);axs[0].set_title('Input Image')
axs[1].imshow(result_image);axs[1].set_title('Result');
```
0%| | 0/30 [00:00<?, ?it/s]
![png](01_stable_diffusion_introduction_CN_files/01_stable_diffusion_introduction_CN_47_1.png)
**练习:** 用这个管线做实验,试试你自己的图片,或者试试不同的强度和文字提示。也可以试试不同的图片尺寸、生成步数。
## In-Painting
![inpaint_unet.png](data:image/png;base64,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)
如果我们想在一张图中保留一部分不变而在其它部分生成新东西,那该怎么办呢?这种技术叫 inpainting 。虽然我们可以用 `StableDiffusionInpaintPipelineLegacy` 管线来实现,但其实有更简单的选择。这里的 Stable Diffusion 模型接收一个掩模(mask)作为额外条件性输入。这个掩模图片需要和输入图片尺寸一致,白色区域表示要被替换的部分,黑色区域表示要保留的部分。以下代码就展示了我们如何载入这个管线并如何应用到前面载入的示例图片和掩模上:
```python
# Load the inpainting pipeline (requires a suitable inpainting model)
pipe = StableDiffusionInpaintPipeline.from_pretrained("runwayml/stable-diffusion-inpainting")
pipe = pipe.to(device)
```
```python
# Inpaint with a prompt for what we want the result to look like
prompt = "A small robot, high resolution, sitting on a park bench"
image = pipe(prompt=prompt, image=init_image, mask_image=mask_image).images[0]
# View the result
fig, axs = plt.subplots(1, 3, figsize=(16, 5))
axs[0].imshow(init_image);axs[0].set_title('Input Image')
axs[1].imshow(mask_image);axs[1].set_title('Mask')
axs[2].imshow(image);axs[2].set_title('Result');
```
0%| | 0/50 [00:00<?, ?it/s]
![png](01_stable_diffusion_introduction_CN_files/01_stable_diffusion_introduction_CN_53_1.png)
当和其它可以自动生成掩模的模型结合起来的时候,这个模型将会相当强大。比如[这个示例space](https://huggingface.co/spaces/nielsr/text-based-inpainting)就用了一个名为 CLIPSeg 的模型,它可以根据文字描述自动地用掩模去掉一个物体。
### 题外话:管理你的模型缓存
探索不同的管线和模型可能会占满你的硬盘空间。你可用这个指令看看你都下载了哪些模型到你的硬盘上:
```python
!ls ~/.cache/huggingface/diffusers/ # List the contents of the cache directory
```
models--CompVis--stable-diffusion-v1-4
models--ddpm-bedroom-256
models--google--ddpm-bedroom-256
models--google--ddpm-celebahq-256
models--runwayml--stable-diffusion-inpainting
models--stabilityai--stable-diffusion-2-1-base
看看[缓存相关文档](https://huggingface.co/docs/huggingface_hub/main/en/how-to-cache)来了解如何高效地查看和管理缓存。
## Depth2Image
![depth to image examples](https://camo.githubusercontent.com/5a7cd804dff6283157144f1d54b99a637195fcd25f10c7e9ab7952322d746269/68747470733a2f2f68756767696e67666163652e636f2f73746162696c69747961692f737461626c652d646966667573696f6e2d322d64657074682f7265736f6c76652f6d61696e2f646570746832696d6167652e706e67)
_Input image, depth image and generated examples (image source: StabilityAI)_
Img2Img 已经很厉害了,但有时我们还想借用原始图片的整体结构,但使用不同的颜色或纹理来生成新图片。保留图片整体结构但却不保留原有颜色,是很难通过调节 Img2Img 的“强度”来实现的。
所以这里就需要另一个微调的模型了!这个模型需要输入额外的深度信息作为生成条件。相关管线使用了一个深度预测模型来预测出一个深度图,然后这个深度图会被输入微调过的 UNet 中用以生成图片。我们这里希望生成的图片能够保留原始图片的深度信息和总体结构,同时又在相关部分填入全新的内容。
```python
# Load the Depth2Img pipeline (requires a suitable model)
pipe = StableDiffusionDepth2ImgPipeline.from_pretrained("stabilityai/stable-diffusion-2-depth")
pipe = pipe.to(device)
```
```python
# Inpaint with a prompt for what we want the result to look like
prompt = "An oil painting of a man on a bench"
image = pipe(prompt=prompt, image=init_image).images[0]
# View the result
fig, axs = plt.subplots(1, 2, figsize=(16, 5))
axs[0].imshow(init_image);axs[0].set_title('Input Image')
axs[1].imshow(image);axs[1].set_title('Result');
```
0%| | 0/40 [00:00<?, ?it/s]
![png](01_stable_diffusion_introduction_CN_files/01_stable_diffusion_introduction_CN_62_1.png)
让我们和 Img2Img 的例子做个对比 —— 这里生成结果有着更丰富的颜色变化,但图片的整体结构仍然忠于原始图片。这里的例子其实并不够理想,因为为了匹配上狗的形状,这里生成的人有着怪异的身体构造。但在某些场景下,这个技术还是相当有用的,比如这个例子中的“杀手级应用软件”(查看[推特](https://twitter.com/CarsonKatri/status/1600248599254007810?s=20&t=BlzSK26sfqi2336SN0gKpQ)),它利用深度模型去给一个 3D 场景加入纹理!
# 下一步?
希望这节的内容让你对 Stable Diffusion 能做的事情有了一个认识!如果你玩烦了本节笔记本中的例子,你可以尝试 **DreamBooth hackathon** 这节笔记本,看看怎样自己来微调 Stable Diffusion 模型,来用在文字到图像或图像到图像的相关管线上。
如果你想深入探索不同的组成部分是如何运作的,你可以看看 **Stable Diffusion Deep Dive** 这节笔记本,这里介绍了大量的细节以及额外的一些小技巧。
一定要和我们以及社区分享你的创造啊!
|