File size: 19,822 Bytes
1e6762b 6e82103 ad622f9 1e6762b 6e82103 6b20e72 c514fd1 91b54d9 1e6762b 6b20e72 c514fd1 6b20e72 c7c3a3c c514fd1 1e6762b 76c627a a5f9d20 e11abed 8083a61 1e6762b a5f9d20 76c627a 91b54d9 8083a61 1e6762b 91b54d9 76c627a 922bd17 1e6762b 61450e6 1e6762b 61450e6 76c627a 91b54d9 1e6762b 76c627a 1e6762b 76c627a c514fd1 76c627a c514fd1 91b54d9 76c627a 91b54d9 76c627a 91b54d9 76c627a 91b54d9 76c627a 91b54d9 76c627a 1e6762b 707436a 76c627a 8083a61 76c627a 14d6943 76c627a 8083a61 14d6943 6e82103 76c627a 91b54d9 76c627a 8083a61 c70c3a9 76c627a c70c3a9 76c627a c70c3a9 76c627a 8083a61 c70c3a9 1e6762b 76c627a c70c3a9 1e6762b 8083a61 c70c3a9 6b20e72 76c627a 06df4c1 28667ad 06df4c1 76c627a 28667ad 8083a61 28667ad 76c627a 06df4c1 7155028 1e6762b 544cd77 1e6762b 7155028 1e6762b 9c3622c 922bd17 1e6762b 544cd77 1e6762b 544cd77 1e6762b | 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 | # import streamlit as st
# import re
# from transformers import pipeline
# from diffusers import StableDiffusionPipeline
# from pydub import AudioSegment
# import zipfile
# from io import BytesIO
# import json
# import torch
# # 初始化模型 (全部使用原始pipeline)
# @st.cache_resource
# def load_pipelines():
# try:
# # 1. Script generation
# script_pipe = pipeline(
# "text2text-generation",
# model="RUCAIBox/mvp-story",
# device=0 if torch.cuda.is_available() else -1
# )
# # 2. Storyboard generation (BART model)
# storyboard_pipe = pipeline(
# "text2text-generation",
# model="Jessiesj/script_gpt2-story",
# device=0 if torch.cuda.is_available() else -1
# )
# # 3. Soundtrack generation (MusicGen)
# music_pipe = pipeline(
# "text-to-audio",
# model="facebook/musicgen-small",
# device_map="auto"
# )
# # 4. Storyboard image generation (Stable Diffusion)
# image_pipe = StableDiffusionPipeline.from_pretrained(
# "prompthero/openjourney-v4",
# safety_checker=None,
# torch_dtype=torch.float16 if torch.cuda.is_available() else torch.float32
# )
# return script_pipe, storyboard_pipe, image_pipe
# # return script_pipe, storyboard_pipe, music_pipe, image_pipe
# except Exception as e:
# st.error(f"Failed to load models: {str(e)}")
# raise
# # Prompt template
# def build_script_prompt(theme, style_or_opening):
# return f"""
# {style_or_opening}
# Generate a micro - movie script with two scenes, with the theme: {theme}.
# ### Format requirements (must be strictly followed):
# - Use Markdown syntax
# - Each scene starts with ### Scene X: [Location] (INT/EXT. Time)
# - Include action descriptions (wrapped in square brackets []) and character dialogues
# ### Example:
# ### Scene 1: Spaceship cockpit (INT. Emergency)
# [Alarm lights are flashing wildly, sparks are coming out of the console]
# Captain (shouting): Start the backup engine!
# Co - pilot (typing on the keyboard): The power system has failed!
# ### Scene 2: Alien jungle (EXT. Dusk)
# [The expedition team pushes aside the dense purple bushes]
# Team member A (pointing into the distance): Look! Is that the entrance to the ruins?
# """
# # Content cleaning
# def format_script(text):
# # Remove invalid symbols
# text = re.sub(r"[():<>]{2,}", "", text)
# # Standardize scene titles
# return re.sub(r"(?<!### )Scene\d+", r"### Scene\g<0>", text)
# # Load pipelines
# script_pipe, storyboard_pipe, image_pipe = load_pipelines()
# # script_pipe, storyboard_pipe, music_pipe, image_pipe = load_pipelines()
# # Streamlit interface
# st.title("🎥 Micro - movie Intelligent Creation System")
# # User input
# user_input = st.text_input(
# "Enter the movie theme (e.g., Sci - fi space adventure)",
# help="Recommended format: Genre + Core conflict, example: 'Jungle exploration + Search for the lost golden city'"
# )
# style_or_opening = st.text_input(
# "Enter the script style or opening",
# help="Provide a style description or the beginning of the script"
# )
# if user_input and style_or_opening:
# # 1. Generate the script
# with st.status("🖋️ Generating the script...", expanded=True) as status:
# try:
# # Generate content
# prompt = build_script_prompt(user_input, style_or_opening)
# response = script_pipe(
# prompt,
# max_length=600,
# temperature=0.8,
# num_beams=4,
# no_repeat_ngram_size=3
# )[0]["generated_text"]
# # Post - processing
# def format_script(text):
# return text
# # Display the result
# st.subheader("Generated Script")
# cleaned = format_script(response) # Ensure cleaned is defined here
# st.markdown(f"```markdown\n{format_script(response)}\n```")
# status.update(label="✅ Generation completed", state="complete")
# except Exception as e:
# status.update(label="❌ Generation failed", status="error")
# st.error(f"Error details: {str(e)}")
# st.stop()
# # 2. Generate the storyboard
# with st.status("🎥 Converting to storyboard script...", expanded=True) as status:
# try:
# # 示例数据模板
# sample_storyboard = """
# {
# "scenes": [
# {
# "scene_number": 1,
# "location": "Abandoned Space Station (INT. Dystopian Future)",
# "characters": ["Eve", "R2-D9"],
# "camera_angle": "Wide shot showing decaying corridors",
# "action": "Eve scans walls with her holographic projector, R2-D9 beeps nervously"
# },
# {
# "scene_number": 2,
# "location": "Alien Wasteland (EXT. Toxic Sunset)",
# "characters": ["Eve", "Mutant Leader"],
# "camera_angle": "Low-angle shot of Eve aiming plasma rifle",
# "action": "Mutant leader growls, acid rain hisses on Eve's protective suit"
# }
# ]
# }
# """
# # JSON模板引导
# json_template = """{
# "scenes": [
# {
# "scene_number": 1,
# "location": "Scene location (INT/EXT)",
# "characters": ["Character 1", "Character 2"],
# "camera_angle": "Camera angle",
# "action": "Main action description"
# }
# ]
# }"""
# # 生成分镜
# # storyboard = storyboard_pipe(
# # f"Convert to storyboard JSON: {cleaned[:800]}",
# # max_length=1500,
# # temperature=0.3
# # )[0]["generated_text"]
# # 增强的格式修复
# # storyboard = (
# # storyboard.replace("'", '"')
# # .replace("None", '""')
# # .replace("True", "true")
# # .replace("False", "false")
# # )
# # storyboard = re.search(r'\{.*\}', storyboard, re.DOTALL).group()
# # parsed = json.loads(storyboard)
# parsed = json.loads(sample_storyboard)
# except Exception as e:
# # st.warning(f"Using sample storyboard due to error: {str(e)}")
# # parsed = json.loads(sample_storyboard) # 回退到示例数据
# # status.update(label="⚠️ Using sample storyboard", status="warning")
# pass
# finally:
# # status.update(label="Storyboard ready", status="complete")
# # st.subheader("Storyboard Script")
# # st.json(parsed)
# # st.caption("*Note: If using sample data, edit the JSON before proceeding*")
# pass
# # 3. 生成配乐
# # with st.spinner("🎵 正在生成配乐..."):
# # audio = music_pipe(
# # "epic sci-fi background music",
# # max_new_tokens=256
# # )
# # audio_buffer = BytesIO()
# # audio["audio"].export(audio_buffer, format="wav")
# # st.subheader("🎧 背景音乐")
# # st.audio(audio_buffer, format="audio/wav")
# # 4. 生成分镜图（可选）
# # if st.toggle("🎨 生成分镜预览图（需要2-3分钟）"):
# # with st.spinner("🖼️ 正在渲染电影级画面..."):
# # # 从剧本提取关键描述
# # scene_desc = re.search(r"场景\d+：(.*?)\n", script)
# # if scene_desc:
# # scene_desc = scene_desc.group(1)
# # else:
# # scene_desc = f"movie scene of {user_input}"
# # # 精细化图片提示词
# # image_prompt = f"""电影剧照，{scene_desc}，
# # 超高清8K分辨率，
# # 电影级打光，浅景深效果，
# # Unreal Engine 5渲染，
# # 电影胶片颗粒感"""
# # # 优化生成参数
# # image = image_pipe(
# # image_prompt,
# # num_inference_steps=50,
# # guidance_scale=9.5,
# # width=1024,
# # height=576 # 16:9电影比例
# # ).images[0]
# # st.subheader("🎞️ 首场景预览")
# # st.image(image, caption=scene_desc)
# # except Exception as e:
# # st.error("生成过程中发生错误")
# # st.code(f"错误详情：{str(e)}")
# # st.button("点击查看技术详情",
# # help="将以下信息提供给技术人员：\n" + str(e))
# st.subheader("🎧 Background Music")
# for i, scene in enumerate(parsed["scenes"]):
# action = scene["action"]
# with st.spinner(f"🎵 Generating soundtrack for scene {i + 1}..."):
# try:
# audio = music_pipe(
# action,
# max_new_tokens=256
# )
# audio_buffer = BytesIO()
# audio["audio"].export(audio_buffer, format="wav")
# st.audio(audio_buffer, format="audio/wav", caption=f"Soundtrack for scene {i + 1}")
# except Exception as e:
# st.error(f"Failed to generate soundtrack for scene {i + 1}: {str(e)}")
# st.subheader("🎨 Storyboard Images")
# for i, scene in enumerate(parsed["scenes"]):
# location = scene["location"]
# if st.toggle(f"Generate preview for scene {i + 1} (takes 2 - 3 minutes)"):
# with st.spinner(f"🖼️ Rendering image for scene {i + 1}..."):
# try:
# # 精细化图片提示词
# image_prompt = f"""Movie still, {location},
# Ultra - high - definition 8K resolution,
# Movie - level lighting, shallow depth - of - field effect,
# Unreal Engine 5 rendering,
# Movie film graininess"""
# # 优化生成参数
# image = image_pipe(
# image_prompt,
# num_inference_steps=50,
# guidance_scale=9.5,
# width=1024,
# height=576 # 16:9 movie ratio
# ).images[0]
# st.image(image, caption=f"Preview for scene {i + 1}")
# except Exception as e:
# st.error(f"Failed to generate image for scene {i + 1}: {str(e)}")
import streamlit as st
import re
from transformers import pipeline, AutoModelForTextToWaveform, AutoTokenizer
from diffusers import StableDiffusionPipeline
from pydub import AudioSegment
import zipfile
from io import BytesIO
import json
import torch
import torchaudio
# 初始化模型 (全部使用原始pipeline)
@st.cache_resource
def load_pipelines():
try:
device = "cuda:0" if torch.cuda.is_available() else "cpu"
torch_dtype = torch.float16 if device == "cuda:0" else torch.float32
# 1. Script generation
script_pipe = pipeline(
"text2text-generation",
model="RUCAIBox/mvp-story",
device=device
)
# 2. Storyboard generation (BART model)
storyboard_pipe = pipeline(
"text2text-generation",
model="Jessiesj/script_gpt2-story",
device=device
)
# 3. Soundtrack generation (MusicGen)
music_model = AutoModelForTextToWaveform.from_pretrained(
"facebook/musicgen-small",
torch_dtype=torch_dtype
).to(device)
music_tokenizer = AutoTokenizer.from_pretrained("facebook/musicgen-small")
music_pipe = pipeline(
"text-to-audio",
model=music_model,
tokenizer=music_tokenizer,
device=device
)
# 4. Storyboard image generation (Stable Diffusion)
image_pipe = StableDiffusionPipeline.from_pretrained(
"prompthero/openjourney-v4",
safety_checker=None,
torch_dtype=torch_dtype
).to(device)
return script_pipe, storyboard_pipe, music_pipe, image_pipe
except Exception as e:
st.error(f"Failed to load models: {str(e)}")
raise
# Prompt template
def build_script_prompt(theme, style_or_opening):
return f"""
{style_or_opening}
Generate a micro - movie script with two scenes, with the theme: {theme}.
### Format requirements (must be strictly followed):
- Use Markdown syntax
- Each scene starts with ### Scene X: [Location] (INT/EXT. Time)
- Include action descriptions (wrapped in square brackets []) and character dialogues
### Example:
### Scene 1: Spaceship cockpit (INT. Emergency)
[Alarm lights are flashing wildly, sparks are coming out of the console]
Captain (shouting): Start the backup engine!
Co - pilot (typing on the keyboard): The power system has failed!
### Scene 2: Alien jungle (EXT. Dusk)
[The expedition team pushes aside the dense purple bushes]
Team member A (pointing into the distance): Look! Is that the entrance to the ruins?
"""
# Content cleaning
def format_script(text):
# Remove invalid symbols
text = re.sub(r"[():<>]{2,}", "", text)
# Standardize scene titles
return re.sub(r"(?<!### )Scene\d+", r"### Scene\g<0>", text)
# Load pipelines
script_pipe, storyboard_pipe, music_pipe, image_pipe = load_pipelines()
# Streamlit interface
st.title("🎥 Micro - movie Intelligent Creation System")
# User input
user_input = st.text_input(
"Enter the movie theme (e.g., Sci - fi space adventure)",
help="Recommended format: Genre + Core conflict, example: 'Jungle exploration + Search for the lost golden city'"
)
style_or_opening = st.text_input(
"Enter the script style or opening",
help="Provide a style description or the beginning of the script"
)
if user_input and style_or_opening:
# 1. Generate the script
with st.status("🖋️ Generating the script...", expanded=True) as status:
try:
# Generate content
prompt = build_script_prompt(user_input, style_or_opening)
response = script_pipe(
prompt,
max_length=600,
temperature=0.8,
num_beams=4,
no_repeat_ngram_size=3
)[0]["generated_text"]
# Post - processing
def format_script(text):
return text
# Display the result
st.subheader("Generated Script")
cleaned = format_script(response) # Ensure cleaned is defined here
st.markdown(f"```markdown\n{format_script(response)}\n```")
status.update(label="✅ Generation completed")
except Exception as e:
status.update(label="❌ Generation failed")
st.error(f"Error details: {str(e)}")
st.stop()
# 2. Generate the storyboard
with st.status("🎥 Converting to storyboard script...", expanded=True) as status:
try:
# 示例数据模板
sample_storyboard = """
{
"scenes": [
{
"scene_number": 1,
"location": "Abandoned Space Station (INT. Dystopian Future)",
"characters": ["Eve", "R2-D9"],
"camera_angle": "Wide shot showing decaying corridors",
"action": "Eve scans walls with her holographic projector, R2-D9 beeps nervously"
},
{
"scene_number": 2,
"location": "Alien Wasteland (EXT. Toxic Sunset)",
"characters": ["Eve", "Mutant Leader"],
"camera_angle": "Low-angle shot of Eve aiming plasma rifle",
"action": "Mutant leader growls, acid rain hisses on Eve's protective suit"
}
]
}
"""
# JSON模板引导
json_template = """{
"scenes": [
{
"scene_number": 1,
"location": "Scene location (INT/EXT)",
"characters": ["Character 1", "Character 2"],
"camera_angle": "Camera angle",
"action": "Main action description"
}
]
}"""
# 生成分镜
storyboard = storyboard_pipe(
f"Convert to storyboard JSON: {cleaned[:800]}",
max_length=1500,
temperature=0.3
)[0]["generated_text"]
# 增强的格式修复
storyboard = (
storyboard.replace("'", '"')
.replace("None", '""')
.replace("True", "true")
.replace("False", "false")
)
storyboard = re.search(r'\{.*\}', storyboard, re.DOTALL).group()
parsed = json.loads(storyboard)
except Exception as e:
st.warning(f"Using sample storyboard due to error: {str(e)}")
parsed = json.loads(sample_storyboard)
status.update(label="⚠️ This is sample storyboard")
status.update(label="Storyboard ready")
st.subheader("Storyboard Script")
st.json(parsed)
# st.caption("*Note: If using sample data, edit the JSON before proceeding*")
st.subheader("🎧 Background Music")
for i, scene in enumerate(parsed["scenes"]):
action = scene["action"]
with st.spinner(f"🎵 Generating soundtrack for scene {i + 1}..."):
try:
audio = music_pipe(
action,
max_new_tokens=256
)
audio_buffer = BytesIO()
audio["audio"].export(audio_buffer, format="wav")
st.audio(audio_buffer, format="audio/wav", caption=f"Soundtrack for scene {i + 1}")
except Exception as e:
st.error(f"Failed to generate soundtrack for scene {i + 1}: {str(e)}")
st.subheader("🎨 Storyboard Images")
for i, scene in enumerate(parsed["scenes"]):
location = scene["location"]
if st.toggle(f"Generate preview for scene {i + 1} (takes 2 - 3 minutes)"):
with st.spinner(f"🖼️ Rendering image for scene {i + 1}..."):
try:
# 精细化图片提示词
image_prompt = f"""Movie still, {location},
Ultra - high - definition 8K resolution,
Movie - level lighting, shallow depth - of - field effect,
Unreal Engine 5 rendering,
Movie film graininess"""
# 优化生成参数
image = image_pipe(
image_prompt,
num_inference_steps=50,
guidance_scale=9.5,
width=1024,
height=576 # 16:9 movie ratio
).images[0]
st.image(image, caption=f"Preview for scene {i + 1}")
except Exception as e:
st.error(f"Failed to generate image for scene {i + 1}: {str(e)}")
|