File size: 15,467 Bytes
d919881 59d4479 d919881 59d4479 d919881 b1acf7e d919881 472585e d919881 472585e d919881 472585e d919881 472585e d919881 472585e d919881 472585e d919881 472585e d919881 |
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 |
import os
import gdown
from pathlib import Path
import logging
from typing import Tuple, Any
import torch
import torch.nn as nn
from torchvision import models
from dotenv import load_dotenv
load_dotenv()
# Configure logging
logging.basicConfig(level=logging.INFO)
logger = logging.getLogger(__name__)
class SimpleModelManager:
"""Simple model manager that downloads models from Google Drive using gdown"""
def __init__(self, model_dir: str = "model_weights", cache_models: bool = True):
"""
Initialize simple model manager
Args:
model_dir: Local directory to store models
cache_models: Whether to cache models locally
"""
self.model_dir = Path(model_dir)
self.model_dir.mkdir(exist_ok=True)
self.cache_models = cache_models
# Load model links from environment variables
self.model_links = {
"vision": {
"url": os.getenv("VISION_MODEL_DRIVE_ID", ""),
"filename": os.getenv("VISION_MODEL_FILENAME", "resnet50_model.pth"),
"description": "Vision sentiment analysis model",
},
"audio": {
"url": os.getenv("AUDIO_MODEL_DRIVE_ID", ""),
"filename": os.getenv("AUDIO_MODEL_FILENAME", "wav2vec2_model.pth"),
"description": "Audio sentiment analysis model",
},
}
# Validate that environment variables are set
self._validate_environment()
def _validate_environment(self):
"""Validate that required environment variables are set"""
missing_vars = []
if not self.model_links["vision"]["url"]:
missing_vars.append("VISION_MODEL_DRIVE_ID")
if not self.model_links["audio"]["url"]:
missing_vars.append("AUDIO_MODEL_DRIVE_ID")
if missing_vars:
logger.warning(f"Missing environment variables: {', '.join(missing_vars)}")
logger.warning("Please set these in your .env file or environment")
logger.warning("Models will not be available until these are configured")
def download_from_google_drive(self, share_url: str, filename: str) -> str:
"""
Download file from Google Drive share link using gdown
Args:
share_url: Google Drive share link
filename: Name to save the file as
Returns:
Path to downloaded file
"""
try:
local_path = self.model_dir / filename
if local_path.exists() and self.cache_models:
logger.info(f"Model already cached: {local_path}")
return str(local_path)
logger.info(f"Downloading {filename} from Google Drive using gdown...")
# Use gdown to download the file
# gdown automatically handles virus scan warnings and other Google Drive issues
output_path = str(local_path)
# Download with progress bar
gdown.download(
id=share_url,
output=output_path,
quiet=False, # Show progress bar
fuzzy=True, # Handle various Google Drive URL formats
)
# Verify the file was downloaded
if not Path(output_path).exists():
raise FileNotFoundError(f"Download failed: {output_path} not found")
file_size = Path(output_path).stat().st_size
if file_size == 0:
raise ValueError(f"Downloaded file is empty: {output_path}")
logger.info(f"Successfully downloaded {filename} ({file_size} bytes)")
return output_path
except Exception as e:
logger.error(f"Failed to download {filename}: {e}")
raise
def load_vision_model(self) -> Tuple[Any, torch.device, int]:
"""Load vision sentiment model"""
try:
model_info = self.model_links["vision"]
# Check if URL is configured
if not model_info["url"]:
raise ValueError("VISION_MODEL_DRIVE_ID environment variable not set")
model_path = self.download_from_google_drive(
model_info["url"], model_info["filename"]
)
# Validate the downloaded file
if not Path(model_path).exists():
raise FileNotFoundError(f"Model file not found at {model_path}")
file_size = Path(model_path).stat().st_size
if file_size == 0:
raise ValueError(f"Model file is empty: {model_path}")
# Check file header to see what type of file it is
with open(model_path, "rb") as f:
header = f.read(100) # Read first 100 bytes
logger.info(f"File size: {file_size} bytes")
logger.info(f"File header (first 100 bytes): {header[:50]}...")
# Try to detect file type
if header.startswith(b"<"):
raise ValueError(
f"File appears to be HTML/XML, not a PyTorch model: {model_path}"
)
elif header.startswith(b"\x89PNG"):
raise ValueError(f"File appears to be a PNG image: {model_path}")
elif header.startswith(b"\xff\xd8\xff"):
raise ValueError(f"File appears to be a JPEG image: {model_path}")
# For any other file type (including ZIP), try to load it directly as a PyTorch model
logger.info(
f"File appears to be a PyTorch model file, attempting to load directly..."
)
# Load the model
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
try:
# Try loading the file directly as a PyTorch model
checkpoint = torch.load(
model_path, map_location=device, weights_only=False
)
logger.info("Successfully loaded model file directly")
except Exception as load_error:
logger.error(f"Failed to load model directly: {load_error}")
try:
# Try with weights only as fallback
checkpoint = torch.load(
model_path, map_location=device, weights_only=True
)
logger.info("Loaded with weights_only=True (weights only)")
except Exception as fallback_error:
logger.error(
f"Failed to load with weights_only=True: {fallback_error}"
)
raise ValueError(
f"Cannot load model file {model_path}. File may be corrupted or in wrong format."
)
# Initialize ResNet-50 model
model = models.resnet50(weights=None)
num_ftrs = model.fc.in_features
# Determine number of classes from checkpoint
if "fc.weight" in checkpoint:
num_classes = checkpoint["fc.weight"].shape[0]
else:
num_classes = 3 # Default fallback
model.fc = nn.Linear(num_ftrs, num_classes)
model.load_state_dict(checkpoint)
model.to(device)
model.eval()
logger.info(f"Vision model loaded successfully with {num_classes} classes!")
return model, device, num_classes
except Exception as e:
logger.error(f"Failed to load vision model: {e}")
raise
def load_audio_model(self) -> Tuple[Any, torch.device]:
"""Load audio sentiment model"""
try:
model_info = self.model_links["audio"]
# Check if URL is configured
if not model_info["url"]:
raise ValueError("AUDIO_MODEL_DRIVE_ID environment variable not set")
model_path = self.download_from_google_drive(
model_info["url"], model_info["filename"]
)
# Validate the downloaded file
if not Path(model_path).exists():
raise FileNotFoundError(f"Model file not found at {model_path}")
file_size = Path(model_path).stat().st_size
if file_size == 0:
raise ValueError(f"Model file is empty: {model_path}")
# Check file header to see what type of file it is
with open(model_path, "rb") as f:
header = f.read(100) # Read first 100 bytes
logger.info(f"File size: {file_size} bytes")
logger.info(f"File header (first 100 bytes): {header[:50]}...")
# Try to detect file type
if header.startswith(b"<"):
raise ValueError(
f"File appears to be HTML/XML, not a PyTorch model: {model_path}"
)
elif header.startswith(b"\x89PNG"):
raise ValueError(f"File appears to be a PNG image: {model_path}")
elif header.startswith(b"\xff\xd8\xff"):
raise ValueError(f"File appears to be a JPEG image: {model_path}")
# For any other file type (including ZIP), try to load it directly as a PyTorch model
logger.info(
f"File appears to be a PyTorch model file, attempting to load directly..."
)
# Load the model
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
try:
# Try loading the file directly as a PyTorch model
checkpoint = torch.load(
model_path, map_location=device, weights_only=False
)
logger.info("Successfully loaded model file directly")
except Exception as load_error:
logger.error(f"Failed to load model directly: {load_error}")
try:
# Try with weights only as fallback
checkpoint = torch.load(
model_path, map_location=device, weights_only=True
)
logger.info("Loaded with weights_only=True (weights only)")
except Exception as fallback_error:
logger.error(
f"Failed to load with weights_only=True: {fallback_error}"
)
raise ValueError(
f"Cannot load model file {model_path}. File may be corrupted or in wrong format."
)
# Check if we have a state dict or a full model
if isinstance(checkpoint, dict) and "classifier.weight" in checkpoint:
# This is a state dictionary - we need to initialize the model first
from transformers import AutoModelForAudioClassification
# Determine number of classes from checkpoint
if "classifier.weight" in checkpoint:
num_classes = checkpoint["classifier.weight"].shape[0]
else:
num_classes = 3 # Default fallback
# Initialize Wav2Vec2 model with the correct number of classes
model = AutoModelForAudioClassification.from_pretrained(
"facebook/wav2vec2-base", num_labels=num_classes
)
# Load the state dictionary
model.load_state_dict(checkpoint)
model.to(device)
model.eval()
logger.info(
f"Audio model loaded successfully with {num_classes} classes!"
)
return model, device
else:
# This is a full model object
model = checkpoint
model.to(device)
model.eval()
logger.info("Audio model loaded successfully!")
return model, device
except Exception as e:
logger.error(f"Failed to load audio model: {e}")
raise
def update_model_links(self, vision_url: str = None, audio_url: str = None):
"""Update Google Drive URLs for models (optional override)"""
if vision_url:
self.model_links["vision"]["url"] = vision_url
if audio_url:
self.model_links["audio"]["url"] = audio_url
# Update environment variables if provided
if vision_url:
os.environ["VISION_MODEL_DRIVE_ID"] = vision_url
if audio_url:
os.environ["AUDIO_MODEL_DRIVE_ID"] = audio_url
logger.info("Model links updated!")
def list_cached_models(self) -> list:
"""List all cached models"""
cached_models = []
for file_path in self.model_dir.glob("*.pth"):
cached_models.append(file_path.name)
return cached_models
def clear_cache(self):
"""Clear all cached models"""
for file_path in self.model_dir.glob("*.pth"):
file_path.unlink()
logger.info("Cache cleared!")
def get_model_status(self) -> dict:
"""Get status of all models"""
status = {}
for model_type, info in self.model_links.items():
status[model_type] = {
"configured": bool(info["url"]),
"filename": info["filename"],
"cached": (self.model_dir / info["filename"]).exists(),
"url": info["url"] if info["url"] else "Not configured",
}
return status
# Example usage
if __name__ == "__main__":
# Initialize manager
manager = SimpleModelManager()
# Check model status
status = manager.get_model_status()
print("Model Status:")
for model_type, info in status.items():
print(f" {model_type}: {'✅' if info['configured'] else '❌'} {info['url']}")
if info["cached"]:
print(f" 📁 Cached: {info['filename']}")
# Load models if configured
try:
if status["vision"]["configured"]:
vision_model, device, num_classes = manager.load_vision_model()
print(f"✅ Vision model loaded: {num_classes} classes")
else:
print("❌ Vision model not configured")
if status["audio"]["configured"]:
audio_model, device = manager.load_audio_model()
print("✅ Audio model loaded")
else:
print("❌ Audio model not configured")
if status["vision"]["configured"] and status["audio"]["configured"]:
print("\n🎉 All models loaded successfully!")
else:
print("\n⚠️ Some models are not configured")
print("Please set the following environment variables:")
print(" VISION_MODEL_DRIVE_ID")
print(" AUDIO_MODEL_DRIVE_ID")
except Exception as e:
print(f"Error loading models: {e}")
print("\nFor folder structures:")
print(" 1. Navigate to each subfolder (Audio/Vision)")
print(" 2. Right-click on each .pth file")
print(" 3. Share -> Copy link")
print(" 4. Use those direct file links instead of folder links")
print("\nNote: Downloaded files are used directly as PyTorch models.")
print("\nOr set environment variables in your .env file:")
print(" VISION_MODEL_DRIVE_ID=your_vision_model_file_id")
print(" AUDIO_MODEL_DRIVE_ID=your_audio_model_file_id")
|