File size: 19,708 Bytes
05676b4 |
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 |
"""
Model registry for BackgroundFX Pro.
Manages available models, versions, and metadata.
"""
import json
import hashlib
from pathlib import Path
from typing import Dict, List, Optional, Any, Tuple
from dataclasses import dataclass, field, asdict
from enum import Enum
from datetime import datetime
import requests
import yaml
import logging
logger = logging.getLogger(__name__)
class ModelStatus(Enum):
"""Model availability status."""
AVAILABLE = "available"
DOWNLOADING = "downloading"
NOT_DOWNLOADED = "not_downloaded"
CORRUPTED = "corrupted"
DEPRECATED = "deprecated"
class ModelTask(Enum):
"""Model task types."""
SEGMENTATION = "segmentation"
MATTING = "matting"
ENHANCEMENT = "enhancement"
DETECTION = "detection"
BACKGROUND_GEN = "background_generation"
class ModelFramework(Enum):
"""Supported frameworks."""
PYTORCH = "pytorch"
ONNX = "onnx"
TENSORRT = "tensorrt"
COREML = "coreml"
TFLITE = "tflite"
@dataclass
class ModelInfo:
"""Model information and metadata."""
# Basic info
model_id: str
name: str
version: str
task: ModelTask
framework: ModelFramework
# Files and URLs
url: str
mirror_urls: List[str] = field(default_factory=list)
filename: str = ""
file_size: int = 0
sha256: Optional[str] = None
# Model details
description: str = ""
author: str = ""
license: str = ""
paper_url: Optional[str] = None
github_url: Optional[str] = None
# Performance metrics
accuracy: Optional[float] = None
speed_fps: Optional[float] = None
memory_mb: Optional[int] = None
# Requirements
min_gpu_memory_gb: float = 0
min_ram_gb: float = 2
requires_gpu: bool = False
supported_platforms: List[str] = field(default_factory=lambda: ["windows", "linux", "macos"])
# Configuration
input_size: Optional[Tuple[int, int]] = None
batch_size: int = 1
config: Dict[str, Any] = field(default_factory=dict)
# Status
status: ModelStatus = ModelStatus.NOT_DOWNLOADED
local_path: Optional[str] = None
download_date: Optional[datetime] = None
last_used: Optional[datetime] = None
use_count: int = 0
def to_dict(self) -> Dict[str, Any]:
"""Convert to dictionary."""
data = asdict(self)
# Convert enums to strings
data['task'] = self.task.value
data['framework'] = self.framework.value
data['status'] = self.status.value
# Convert datetime to ISO format
if self.download_date:
data['download_date'] = self.download_date.isoformat()
if self.last_used:
data['last_used'] = self.last_used.isoformat()
return data
@classmethod
def from_dict(cls, data: Dict[str, Any]) -> 'ModelInfo':
"""Create from dictionary."""
# Convert string enums
if 'task' in data:
data['task'] = ModelTask(data['task'])
if 'framework' in data:
data['framework'] = ModelFramework(data['framework'])
if 'status' in data:
data['status'] = ModelStatus(data['status'])
# Convert ISO strings to datetime
if 'download_date' in data and data['download_date']:
data['download_date'] = datetime.fromisoformat(data['download_date'])
if 'last_used' in data and data['last_used']:
data['last_used'] = datetime.fromisoformat(data['last_used'])
return cls(**data)
class ModelRegistry:
"""Central registry for all available models."""
# Default model definitions
DEFAULT_MODELS = {
"rmbg-1.4": ModelInfo(
model_id="rmbg-1.4",
name="RMBG v1.4",
version="1.4",
task=ModelTask.SEGMENTATION,
framework=ModelFramework.ONNX,
url="https://huggingface.co/briaai/RMBG-1.4/resolve/main/model.onnx",
filename="rmbg_v1.4.onnx",
file_size=176_000_000, # ~176MB
sha256="d0c3e8c7d98e32b9c30e0c8f228e3c6d1a5e5c8e9f0a1b2c3d4e5f6a7b8c9d0e1",
description="State-of-the-art background removal model",
author="BRIA AI",
license="BRIA RMBG-1.4 Community License",
github_url="https://github.com/bria-ai/RMBG-1.4",
accuracy=0.98,
speed_fps=30,
memory_mb=500,
requires_gpu=False,
input_size=(1024, 1024)
),
"u2net": ModelInfo(
model_id="u2net",
name="U2-Net",
version="1.0",
task=ModelTask.SEGMENTATION,
framework=ModelFramework.PYTORCH,
url="https://github.com/xuebinqin/U-2-Net/releases/download/v1.0/u2net.pth",
filename="u2net.pth",
file_size=176_000_000,
description="Salient object detection for background removal",
author="Xuebin Qin et al.",
license="Apache 2.0",
paper_url="https://arxiv.org/abs/2005.09007",
accuracy=0.95,
speed_fps=20,
memory_mb=800,
requires_gpu=True,
input_size=(320, 320)
),
"u2netp": ModelInfo(
model_id="u2netp",
name="U2-Net Lite",
version="1.0",
task=ModelTask.SEGMENTATION,
framework=ModelFramework.PYTORCH,
url="https://github.com/xuebinqin/U-2-Net/releases/download/v1.0/u2netp.pth",
filename="u2netp.pth",
file_size=4_700_000, # ~4.7MB
description="Lightweight version of U2-Net",
author="Xuebin Qin et al.",
license="Apache 2.0",
accuracy=0.92,
speed_fps=40,
memory_mb=200,
requires_gpu=False,
input_size=(320, 320)
),
"isnet": ModelInfo(
model_id="isnet",
name="IS-Net",
version="1.0",
task=ModelTask.SEGMENTATION,
framework=ModelFramework.PYTORCH,
url="https://github.com/xuebinqin/DIS/releases/download/v1.0/isnet.pth",
filename="isnet.pth",
file_size=450_000_000,
description="Highly accurate salient object detection",
author="Xuebin Qin et al.",
license="Apache 2.0",
paper_url="https://arxiv.org/abs/2203.03041",
accuracy=0.97,
speed_fps=15,
memory_mb=1200,
requires_gpu=True,
min_gpu_memory_gb=4,
input_size=(1024, 1024)
),
"modnet": ModelInfo(
model_id="modnet",
name="MODNet",
version="1.0",
task=ModelTask.MATTING,
framework=ModelFramework.PYTORCH,
url="https://github.com/ZHKKKe/MODNet/releases/download/v1.0/modnet_photographic_portrait_matting.ckpt",
filename="modnet.ckpt",
file_size=25_000_000,
description="Trimap-free portrait matting",
author="Zhanghan Ke et al.",
license="CC BY-NC 4.0",
paper_url="https://arxiv.org/abs/2011.11961",
github_url="https://github.com/ZHKKKe/MODNet",
accuracy=0.94,
speed_fps=25,
memory_mb=400,
requires_gpu=False,
input_size=(512, 512)
),
"robust_video_matting": ModelInfo(
model_id="robust_video_matting",
name="Robust Video Matting",
version="1.0",
task=ModelTask.MATTING,
framework=ModelFramework.ONNX,
url="https://github.com/PeterL1n/RobustVideoMatting/releases/download/v1.0.0/rvm_mobilenetv3.onnx",
filename="rvm_mobilenetv3.onnx",
file_size=14_000_000,
description="Temporal coherent video matting",
author="Shanchuan Lin et al.",
license="GPL-3.0",
paper_url="https://arxiv.org/abs/2108.11515",
github_url="https://github.com/PeterL1n/RobustVideoMatting",
accuracy=0.93,
speed_fps=30,
memory_mb=300,
requires_gpu=False,
config={"temporal": True, "recurrent": True}
),
"selfie_segmentation": ModelInfo(
model_id="selfie_segmentation",
name="MediaPipe Selfie Segmentation",
version="1.0",
task=ModelTask.SEGMENTATION,
framework=ModelFramework.TFLITE,
url="https://storage.googleapis.com/mediapipe-models/selfie_segmentation/selfie_segmentation.tflite",
filename="selfie_segmentation.tflite",
file_size=260_000, # ~260KB
description="Ultra-lightweight real-time segmentation",
author="Google MediaPipe",
license="Apache 2.0",
accuracy=0.88,
speed_fps=60,
memory_mb=50,
requires_gpu=False,
input_size=(256, 256)
)
}
def __init__(self, models_dir: Optional[Path] = None,
config_file: Optional[Path] = None):
"""
Initialize model registry.
Args:
models_dir: Directory to store downloaded models
config_file: Optional config file with custom models
"""
self.models_dir = models_dir or Path.home() / ".backgroundfx" / "models"
self.models_dir.mkdir(parents=True, exist_ok=True)
self.registry_file = self.models_dir / "registry.json"
self.models: Dict[str, ModelInfo] = {}
# Load registry
self._load_registry()
# Load custom config if provided
if config_file:
self._load_custom_config(config_file)
# Update model status
self._update_model_status()
def _load_registry(self):
"""Load model registry from file or create default."""
if self.registry_file.exists():
try:
with open(self.registry_file, 'r') as f:
data = json.load(f)
for model_id, model_data in data.items():
self.models[model_id] = ModelInfo.from_dict(model_data)
logger.info(f"Loaded {len(self.models)} models from registry")
except Exception as e:
logger.error(f"Failed to load registry: {e}")
self._initialize_default_registry()
else:
self._initialize_default_registry()
def _initialize_default_registry(self):
"""Initialize with default models."""
self.models = self.DEFAULT_MODELS.copy()
self._save_registry()
logger.info("Initialized registry with default models")
def _save_registry(self):
"""Save registry to file."""
try:
data = {
model_id: model.to_dict()
for model_id, model in self.models.items()
}
with open(self.registry_file, 'w') as f:
json.dump(data, f, indent=2)
except Exception as e:
logger.error(f"Failed to save registry: {e}")
def _load_custom_config(self, config_file: Path):
"""Load custom model configurations."""
try:
with open(config_file, 'r') as f:
if config_file.suffix == '.yaml':
config = yaml.safe_load(f)
else:
config = json.load(f)
for model_data in config.get('models', []):
model = ModelInfo.from_dict(model_data)
self.models[model.model_id] = model
logger.info(f"Added custom model: {model.name}")
self._save_registry()
except Exception as e:
logger.error(f"Failed to load custom config: {e}")
def _update_model_status(self):
"""Update status of all models based on local files."""
for model_id, model in self.models.items():
model_path = self.models_dir / model.filename
if model_path.exists():
# Verify file integrity
if self._verify_model_file(model_path, model):
model.status = ModelStatus.AVAILABLE
model.local_path = str(model_path)
else:
model.status = ModelStatus.CORRUPTED
logger.warning(f"Model {model_id} file is corrupted")
else:
model.status = ModelStatus.NOT_DOWNLOADED
model.local_path = None
def _verify_model_file(self, file_path: Path, model: ModelInfo) -> bool:
"""Verify model file integrity."""
# Check file size
if model.file_size > 0:
actual_size = file_path.stat().st_size
if abs(actual_size - model.file_size) > 1000: # Allow 1KB difference
logger.warning(f"Size mismatch for {model.model_id}: "
f"expected {model.file_size}, got {actual_size}")
return False
# Check SHA256 if available
if model.sha256:
try:
sha256 = self._calculate_sha256(file_path)
if sha256 != model.sha256:
logger.warning(f"SHA256 mismatch for {model.model_id}")
return False
except Exception as e:
logger.error(f"Failed to verify SHA256: {e}")
return False
return True
def _calculate_sha256(self, file_path: Path) -> str:
"""Calculate SHA256 hash of file."""
sha256_hash = hashlib.sha256()
with open(file_path, "rb") as f:
for byte_block in iter(lambda: f.read(4096), b""):
sha256_hash.update(byte_block)
return sha256_hash.hexdigest()
def register_model(self, model: ModelInfo) -> bool:
"""
Register a new model.
Args:
model: Model information
Returns:
True if registered successfully
"""
try:
self.models[model.model_id] = model
self._save_registry()
logger.info(f"Registered model: {model.name}")
return True
except Exception as e:
logger.error(f"Failed to register model: {e}")
return False
def get_model(self, model_id: str) -> Optional[ModelInfo]:
"""Get model information by ID."""
return self.models.get(model_id)
def list_models(self, task: Optional[ModelTask] = None,
framework: Optional[ModelFramework] = None,
status: Optional[ModelStatus] = None) -> List[ModelInfo]:
"""
List models with optional filtering.
Args:
task: Filter by task type
framework: Filter by framework
status: Filter by status
Returns:
List of matching models
"""
models = list(self.models.values())
if task:
models = [m for m in models if m.task == task]
if framework:
models = [m for m in models if m.framework == framework]
if status:
models = [m for m in models if m.status == status]
return models
def get_best_model(self, task: ModelTask,
prefer_speed: bool = False,
require_gpu: Optional[bool] = None) -> Optional[ModelInfo]:
"""
Get best model for a task.
Args:
task: Task type
prefer_speed: Prefer speed over accuracy
require_gpu: GPU requirement
Returns:
Best matching model
"""
candidates = self.list_models(task=task, status=ModelStatus.AVAILABLE)
if require_gpu is not None:
candidates = [m for m in candidates
if m.requires_gpu == require_gpu]
if not candidates:
return None
# Sort by preference
if prefer_speed:
candidates.sort(key=lambda m: m.speed_fps or 0, reverse=True)
else:
candidates.sort(key=lambda m: m.accuracy or 0, reverse=True)
return candidates[0] if candidates else None
def update_model_usage(self, model_id: str):
"""Update model usage statistics."""
if model_id in self.models:
model = self.models[model_id]
model.use_count += 1
model.last_used = datetime.now()
self._save_registry()
def get_total_size(self, status: Optional[ModelStatus] = None) -> int:
"""Get total size of models in bytes."""
models = self.list_models(status=status)
return sum(m.file_size for m in models)
def cleanup_unused_models(self, days: int = 30) -> List[str]:
"""
Remove models not used in specified days.
Args:
days: Days threshold
Returns:
List of removed model IDs
"""
removed = []
cutoff = datetime.now().timestamp() - (days * 86400)
for model_id, model in self.models.items():
if (model.status == ModelStatus.AVAILABLE and
model.last_used and
model.last_used.timestamp() < cutoff):
# Delete file
if model.local_path:
try:
Path(model.local_path).unlink()
model.status = ModelStatus.NOT_DOWNLOADED
model.local_path = None
removed.append(model_id)
logger.info(f"Removed unused model: {model_id}")
except Exception as e:
logger.error(f"Failed to remove model {model_id}: {e}")
if removed:
self._save_registry()
return removed
def export_registry(self, output_file: Path):
"""Export registry to file."""
data = {
'version': '1.0',
'models': [model.to_dict() for model in self.models.values()]
}
with open(output_file, 'w') as f:
if output_file.suffix == '.yaml':
yaml.dump(data, f, default_flow_style=False)
else:
json.dump(data, f, indent=2)
def get_statistics(self) -> Dict[str, Any]:
"""Get registry statistics."""
total_models = len(self.models)
downloaded = len([m for m in self.models.values()
if m.status == ModelStatus.AVAILABLE])
task_counts = {}
for task in ModelTask:
count = len([m for m in self.models.values() if m.task == task])
if count > 0:
task_counts[task.value] = count
return {
'total_models': total_models,
'downloaded_models': downloaded,
'total_size_mb': self.get_total_size() / (1024 * 1024),
'downloaded_size_mb': self.get_total_size(ModelStatus.AVAILABLE) / (1024 * 1024),
'models_by_task': task_counts,
'most_used': sorted(
[(m.model_id, m.use_count) for m in self.models.values()],
key=lambda x: x[1],
reverse=True
)[:5]
} |