Spaces:
Running
on
CPU Upgrade
Running
on
CPU Upgrade
File size: 23,610 Bytes
e7abd9e 0e60add e7abd9e 0e60add e7abd9e 0e60add e7abd9e 0e60add e7abd9e 0e60add e7abd9e |
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 |
from datetime import datetime, timezone
from typing import Dict, Any, Optional, List
import json
import os
from pathlib import Path
import logging
import aiohttp
import asyncio
import time
from huggingface_hub import HfApi, CommitOperationAdd
from huggingface_hub.utils import build_hf_headers
import datasets
from datasets import load_dataset, disable_progress_bar
import sys
import contextlib
from concurrent.futures import ThreadPoolExecutor
import tempfile
from app.config import (
QUEUE_REPO,
HF_TOKEN,
EVAL_REQUESTS_PATH
)
from app.config.hf_config import HF_ORGANIZATION
from app.services.hf_service import HuggingFaceService
from app.utils.model_validation import ModelValidator
from app.services.votes import VoteService
from app.core.cache import cache_config
from app.utils.logging import LogFormatter
# Disable datasets progress bars globally
disable_progress_bar()
logger = logging.getLogger(__name__)
# Context manager to temporarily disable stdout and stderr
@contextlib.contextmanager
def suppress_output():
stdout = sys.stdout
stderr = sys.stderr
devnull = open(os.devnull, 'w')
try:
sys.stdout = devnull
sys.stderr = devnull
yield
finally:
sys.stdout = stdout
sys.stderr = stderr
devnull.close()
class ProgressTracker:
def __init__(self, total: int, desc: str = "Progress", update_frequency: int = 10):
self.total = total
self.current = 0
self.desc = desc
self.start_time = time.time()
self.update_frequency = update_frequency # Percentage steps
self.last_update = -1
# Initial log with fancy formatting
logger.info(LogFormatter.section(desc))
logger.info(LogFormatter.info(f"Starting processing of {total:,} items..."))
sys.stdout.flush()
def update(self, n: int = 1):
self.current += n
current_percentage = (self.current * 100) // self.total
# Only update on frequency steps (e.g., 0%, 10%, 20%, etc.)
if current_percentage >= self.last_update + self.update_frequency or current_percentage == 100:
elapsed = time.time() - self.start_time
rate = self.current / elapsed if elapsed > 0 else 0
remaining = (self.total - self.current) / rate if rate > 0 else 0
# Create progress stats
stats = {
"Progress": LogFormatter.progress_bar(self.current, self.total),
"Items": f"{self.current:,}/{self.total:,}",
"Time": f"⏱️ {elapsed:.1f}s elapsed, {remaining:.1f}s remaining",
"Rate": f"🚀 {rate:.1f} items/s"
}
# Log progress using tree format
for line in LogFormatter.tree(stats):
logger.info(line)
sys.stdout.flush()
self.last_update = (current_percentage // self.update_frequency) * self.update_frequency
def close(self):
elapsed = time.time() - self.start_time
rate = self.total / elapsed if elapsed > 0 else 0
# Final summary with fancy formatting
logger.info(LogFormatter.section("COMPLETED"))
stats = {
"Total": f"{self.total:,} items",
"Time": f"{elapsed:.1f}s",
"Rate": f"{rate:.1f} items/s"
}
for line in LogFormatter.stats(stats):
logger.info(line)
logger.info("="*50)
sys.stdout.flush()
class ModelService(HuggingFaceService):
_instance: Optional['ModelService'] = None
_initialized = False
def __new__(cls):
if cls._instance is None:
logger.info(LogFormatter.info("Creating new ModelService instance"))
cls._instance = super(ModelService, cls).__new__(cls)
return cls._instance
def __init__(self):
if not hasattr(self, '_init_done'):
logger.info(LogFormatter.section("MODEL SERVICE INITIALIZATION"))
super().__init__()
self.validator = ModelValidator()
self.vote_service = VoteService()
self.eval_requests_path = cache_config.eval_requests_file
logger.info(LogFormatter.info(f"Using eval requests path: {self.eval_requests_path}"))
self.eval_requests_path.parent.mkdir(parents=True, exist_ok=True)
self.hf_api = HfApi(token=HF_TOKEN)
self.cached_models = None
self.last_cache_update = 0
self.cache_ttl = cache_config.cache_ttl.total_seconds()
self._init_done = True
logger.info(LogFormatter.success("Initialization complete"))
async def _download_and_process_file(self, file: str, session: aiohttp.ClientSession, progress: ProgressTracker) -> Optional[Dict]:
"""Download and process a file asynchronously"""
try:
# Build file URL
url = f"https://huggingface.co/datasets/{QUEUE_REPO}/resolve/main/{file}"
headers = build_hf_headers(token=self.token)
# Download file
async with session.get(url, headers=headers) as response:
if response.status != 200:
logger.error(LogFormatter.error(f"Failed to download {file}", f"HTTP {response.status}"))
progress.update()
return None
try:
# First read content as text
text_content = await response.text()
# Then parse JSON
content = json.loads(text_content)
except json.JSONDecodeError as e:
logger.error(LogFormatter.error(f"Failed to decode JSON from {file}", e))
progress.update()
return None
# Get status and determine target status
status = content.get("status", "PENDING").upper()
target_status = None
status_map = {
"PENDING": ["PENDING", "RERUN"],
"EVALUATING": ["RUNNING"],
"FINISHED": ["FINISHED", "PENDING_NEW_EVAL"]
}
for target, source_statuses in status_map.items():
if status in source_statuses:
target_status = target
break
if not target_status:
progress.update()
return None
# Calculate wait time
try:
submit_time = datetime.fromisoformat(content["submitted_time"].replace("Z", "+00:00"))
if submit_time.tzinfo is None:
submit_time = submit_time.replace(tzinfo=timezone.utc)
current_time = datetime.now(timezone.utc)
wait_time = current_time - submit_time
model_info = {
"name": content["model"],
"submitter": content.get("sender", "Unknown"),
"revision": content["revision"],
"wait_time": f"{wait_time.total_seconds():.1f}s",
"submission_time": content["submitted_time"],
"status": target_status,
"precision": content.get("precision", "Unknown")
}
progress.update()
return model_info
except (ValueError, TypeError) as e:
logger.error(LogFormatter.error(f"Failed to process {file}", e))
progress.update()
return None
except Exception as e:
logger.error(LogFormatter.error(f"Failed to load {file}", e))
progress.update()
return None
async def _refresh_models_cache(self):
"""Refresh the models cache"""
try:
logger.info(LogFormatter.section("CACHE REFRESH"))
self._log_repo_operation("read", f"{HF_ORGANIZATION}/requests", "Refreshing models cache")
# Initialize models dictionary
models = {
"finished": [],
"evaluating": [],
"pending": []
}
try:
logger.info(LogFormatter.subsection("DATASET LOADING"))
logger.info(LogFormatter.info("Loading dataset files..."))
# List files in repository
with suppress_output():
files = self.hf_api.list_repo_files(
repo_id=QUEUE_REPO,
repo_type="dataset",
token=self.token
)
# Filter JSON files
json_files = [f for f in files if f.endswith('.json')]
total_files = len(json_files)
# Log repository stats
stats = {
"Total_Files": len(files),
"JSON_Files": total_files,
}
for line in LogFormatter.stats(stats, "Repository Statistics"):
logger.info(line)
if not json_files:
raise Exception("No JSON files found in repository")
# Initialize progress tracker
progress = ProgressTracker(total_files, "PROCESSING FILES")
try:
# Create aiohttp session to reuse connections
async with aiohttp.ClientSession() as session:
# Process files in chunks
chunk_size = 50
for i in range(0, len(json_files), chunk_size):
chunk = json_files[i:i + chunk_size]
chunk_tasks = [
self._download_and_process_file(file, session, progress)
for file in chunk
]
results = await asyncio.gather(*chunk_tasks)
# Process results
for result in results:
if result:
status = result.pop("status")
models[status.lower()].append(result)
finally:
progress.close()
# Final summary with fancy formatting
logger.info(LogFormatter.section("CACHE SUMMARY"))
stats = {
"Finished": len(models["finished"]),
"Evaluating": len(models["evaluating"]),
"Pending": len(models["pending"])
}
for line in LogFormatter.stats(stats, "Models by Status"):
logger.info(line)
logger.info("="*50)
except Exception as e:
logger.error(LogFormatter.error("Error processing files", e))
raise
# Update cache
self.cached_models = models
self.last_cache_update = time.time()
logger.info(LogFormatter.success("Cache updated successfully"))
return models
except Exception as e:
logger.error(LogFormatter.error("Cache refresh failed", e))
raise
async def initialize(self):
"""Initialize the model service"""
if self._initialized:
logger.info(LogFormatter.info("Service already initialized, using cached data"))
return
try:
logger.info(LogFormatter.section("MODEL SERVICE INITIALIZATION"))
# Check if cache already exists
cache_path = cache_config.get_cache_path("datasets")
if not cache_path.exists() or not any(cache_path.iterdir()):
logger.info(LogFormatter.info("No existing cache found, initializing datasets cache..."))
cache_config.flush_cache("datasets")
else:
logger.info(LogFormatter.info("Using existing datasets cache"))
# Ensure eval requests directory exists
self.eval_requests_path.parent.mkdir(parents=True, exist_ok=True)
logger.info(LogFormatter.info(f"Eval requests directory: {self.eval_requests_path}"))
# List existing files
if self.eval_requests_path.exists():
files = list(self.eval_requests_path.glob("**/*.json"))
stats = {
"Total_Files": len(files),
"Directory": str(self.eval_requests_path)
}
for line in LogFormatter.stats(stats, "Eval Requests"):
logger.info(line)
# Load initial cache
await self._refresh_models_cache()
self._initialized = True
logger.info(LogFormatter.success("Model service initialization complete"))
except Exception as e:
logger.error(LogFormatter.error("Initialization failed", e))
raise
async def get_models(self) -> Dict[str, List[Dict[str, Any]]]:
"""Get all models with their status"""
if not self._initialized:
logger.info(LogFormatter.info("Service not initialized, initializing now..."))
await self.initialize()
current_time = time.time()
cache_age = current_time - self.last_cache_update
# Check if cache needs refresh
if not self.cached_models:
logger.info(LogFormatter.info("No cached data available, refreshing cache..."))
return await self._refresh_models_cache()
elif cache_age > self.cache_ttl:
logger.info(LogFormatter.info(f"Cache expired ({cache_age:.1f}s old, TTL: {self.cache_ttl}s)"))
return await self._refresh_models_cache()
else:
logger.info(LogFormatter.info(f"Using cached data ({cache_age:.1f}s old)"))
return self.cached_models
async def submit_model(
self,
model_data: Dict[str, Any],
user_id: str
) -> Dict[str, Any]:
logger.info(LogFormatter.section("MODEL SUBMISSION"))
self._log_repo_operation("write", f"{HF_ORGANIZATION}/requests", f"Submitting model {model_data['model_id']} by {user_id}")
stats = {
"Model": model_data["model_id"],
"User": user_id,
"Revision": model_data["revision"],
"Precision": model_data["precision"],
"Type": model_data["model_type"]
}
for line in LogFormatter.tree(stats, "Submission Details"):
logger.info(line)
# Validate required fields
required_fields = [
"model_id", "base_model", "revision", "precision",
"weight_type", "model_type", "use_chat_template"
]
for field in required_fields:
if field not in model_data:
raise ValueError(f"Missing required field: {field}")
# Get model info and validate it exists on HuggingFace
try:
logger.info(LogFormatter.subsection("MODEL VALIDATION"))
# Get the model info to check if it exists
model_info = self.hf_api.model_info(
model_data["model_id"],
revision=model_data["revision"],
token=self.token
)
if not model_info:
raise Exception(f"Model {model_data['model_id']} not found on HuggingFace Hub")
logger.info(LogFormatter.success("Model exists on HuggingFace Hub"))
except Exception as e:
logger.error(LogFormatter.error("Model validation failed", e))
raise
# Update model revision with commit sha
model_data["revision"] = model_info.sha
# Check if model already exists in the system
try:
logger.info(LogFormatter.subsection("CHECKING EXISTING SUBMISSIONS"))
existing_models = await self.get_models()
# Check in all statuses (pending, evaluating, finished)
for status, models in existing_models.items():
for model in models:
if model["name"] == model_data["model_id"] and model["revision"] == model_data["revision"]:
error_msg = f"Model {model_data['model_id']} revision {model_data["revision"]} is already in the system with status: {status}"
logger.error(LogFormatter.error("Submission rejected", error_msg))
raise ValueError(error_msg)
logger.info(LogFormatter.success("No existing submission found"))
except ValueError:
raise
except Exception as e:
logger.error(LogFormatter.error("Failed to check existing submissions", e))
raise
# Validate model card
valid, error, model_card = await self.validator.check_model_card(
model_data["model_id"]
)
if not valid:
logger.error(LogFormatter.error("Model card validation failed", error))
raise Exception(error)
logger.info(LogFormatter.success("Model card validation passed"))
# Check size limits
model_size, error = await self.validator.get_model_size(
model_info,
model_data["precision"],
model_data["base_model"],
revision=model_data["revision"]
)
if model_size is None:
logger.error(LogFormatter.error("Model size validation failed", error))
raise Exception(error)
logger.info(LogFormatter.success(f"Model size validation passed: {model_size:.1f}GB"))
# Size limits based on precision
if model_data["precision"] in ["float16", "bfloat16"] and model_size > 100:
error_msg = f"Model too large for {model_data['precision']} (limit: 100GB)"
logger.error(LogFormatter.error("Size limit exceeded", error_msg))
raise Exception(error_msg)
# Chat template validation if requested
if model_data["use_chat_template"]:
valid, error = await self.validator.check_chat_template(
model_data["model_id"],
model_data["revision"]
)
if not valid:
logger.error(LogFormatter.error("Chat template validation failed", error))
raise Exception(error)
logger.info(LogFormatter.success("Chat template validation passed"))
architectures = model_info.config.get("architectures", "")
if architectures:
architectures = ";".join(architectures)
# Create eval entry
eval_entry = {
"model": model_data["model_id"],
"base_model": model_data["base_model"],
"revision": model_info.sha,
"precision": model_data["precision"],
"params": model_size,
"architectures": architectures,
"weight_type": model_data["weight_type"],
"status": "PENDING",
"submitted_time": datetime.now(timezone.utc).strftime("%Y-%m-%dT%H:%M:%SZ"),
"model_type": model_data["model_type"],
"job_id": -1,
"job_start_time": None,
"use_chat_template": model_data["use_chat_template"],
"sender": user_id
}
logger.info(LogFormatter.subsection("EVALUATION ENTRY"))
for line in LogFormatter.tree(eval_entry):
logger.info(line)
# Upload to HF dataset
try:
logger.info(LogFormatter.subsection("UPLOADING TO HUGGINGFACE"))
logger.info(LogFormatter.info(f"Uploading to {HF_ORGANIZATION}/requests..."))
# Construct the path in the dataset
org_or_user = model_data["model_id"].split("/")[0] if "/" in model_data["model_id"] else ""
model_path = model_data["model_id"].split("/")[-1]
relative_path = f"{org_or_user}/{model_path}_eval_request_False_{model_data['precision']}_{model_data['weight_type']}.json"
# Create a temporary file with the request
with tempfile.NamedTemporaryFile(mode='w', suffix='.json', delete=False) as temp_file:
json.dump(eval_entry, temp_file, indent=2)
temp_file.flush()
temp_path = temp_file.name
# Upload file directly
self.hf_api.upload_file(
path_or_fileobj=temp_path,
path_in_repo=relative_path,
repo_id=f"{HF_ORGANIZATION}/requests",
repo_type="dataset",
commit_message=f"Add {model_data['model_id']} to eval queue",
token=self.token
)
# Clean up temp file
os.unlink(temp_path)
logger.info(LogFormatter.success("Upload successful"))
except Exception as e:
logger.error(LogFormatter.error("Upload failed", e))
raise
# Add automatic vote
try:
logger.info(LogFormatter.subsection("AUTOMATIC VOTE"))
logger.info(LogFormatter.info(f"Adding upvote for {model_data['model_id']} by {user_id}"))
await self.vote_service.add_vote(
model_data["model_id"],
user_id,
"up"
)
logger.info(LogFormatter.success("Vote recorded successfully"))
except Exception as e:
logger.error(LogFormatter.error("Failed to record vote", e))
# Don't raise here as the main submission was successful
return {
"status": "success",
"message": "Model submitted successfully and vote recorded"
}
async def get_model_status(self, model_id: str) -> Dict[str, Any]:
"""Get evaluation status of a model"""
logger.info(LogFormatter.info(f"Checking status for model: {model_id}"))
eval_path = self.eval_requests_path
for user_folder in eval_path.iterdir():
if user_folder.is_dir():
for file in user_folder.glob("*.json"):
with open(file, "r") as f:
data = json.load(f)
if data["model"] == model_id:
status = {
"status": data["status"],
"submitted_time": data["submitted_time"],
"job_id": data.get("job_id", -1)
}
logger.info(LogFormatter.success("Status found"))
for line in LogFormatter.tree(status, "Model Status"):
logger.info(line)
return status
logger.warning(LogFormatter.warning(f"No status found for model: {model_id}"))
return {"status": "not_found"} |