File size: 8,768 Bytes
57a645e | 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 | """
Database models and initialization for Universal Model Trainer
"""
from sqlalchemy import (
Column, Integer, String, Text, Float, Boolean, DateTime,
ForeignKey, JSON, Enum, create_engine
)
from sqlalchemy.ext.asyncio import AsyncSession, create_async_engine
from sqlalchemy.orm import sessionmaker, relationship, declarative_base
from datetime import datetime
import enum
import os
from app.config import settings
# Create async engine
DATABASE_PATH = settings.DATABASE_URL.replace("sqlite:///./", "").replace("sqlite://", "")
os.makedirs(os.path.dirname(DATABASE_PATH) if os.path.dirname(DATABASE_PATH) else ".", exist_ok=True)
# Use async engine for SQLite
ASYNC_DB_URL = settings.DATABASE_URL.replace("sqlite://", "sqlite+aiosqlite://")
engine = create_async_engine(ASYNC_DB_URL, echo=settings.DEBUG)
AsyncSessionLocal = sessionmaker(
engine, class_=AsyncSession, expire_on_commit=False
)
Base = declarative_base()
class JobStatus(str, enum.Enum):
"""Training job status enum."""
PENDING = "pending"
QUEUED = "queued"
RUNNING = "running"
COMPLETED = "completed"
FAILED = "failed"
CANCELLED = "cancelled"
PAUSED = "paused"
class TaskType(str, enum.Enum):
"""Supported task types."""
CAUSAL_LM = "causal-lm"
SEQ2SEQ = "seq2seq"
TOKEN_CLASSIFICATION = "token-classification"
SEQUENCE_CLASSIFICATION = "sequence-classification"
QUESTION_ANSWERING = "question-answering"
SUMMARIZATION = "summarization"
TRANSLATION = "translation"
TEXT_CLASSIFICATION = "text-classification"
MASKED_LM = "masked-lm"
VISION_CLASSIFICATION = "vision-classification"
VISION_SEGMENTATION = "vision-segmentation"
AUDIO_CLASSIFICATION = "audio-classification"
AUDIO_TRANSCRIPTION = "audio-transcription"
class TrainingJob(Base):
"""Model for training jobs."""
__tablename__ = "training_jobs"
id = Column(Integer, primary_key=True, index=True)
job_id = Column(String(36), unique=True, index=True, nullable=False)
name = Column(String(255), nullable=False)
description = Column(Text, nullable=True)
# Task configuration
task_type = Column(String(50), nullable=False)
base_model = Column(String(255), nullable=False)
output_model_name = Column(String(255), nullable=True)
# Dataset configuration
dataset_source = Column(String(50), default="huggingface")
dataset_name = Column(String(255), nullable=True)
dataset_config = Column(String(100), nullable=True)
dataset_split = Column(String(50), default="train")
custom_dataset_path = Column(String(512), nullable=True)
# Training arguments
training_args = Column(JSON, default=dict)
peft_config = Column(JSON, nullable=True)
deepspeed_config = Column(JSON, nullable=True)
# Status and progress
status = Column(String(20), default=JobStatus.PENDING.value)
progress = Column(Float, default=0.0)
current_epoch = Column(Integer, default=0)
total_epochs = Column(Integer, default=3)
current_step = Column(Integer, default=0)
total_steps = Column(Integer, default=0)
# Metrics
train_loss = Column(Float, nullable=True)
eval_loss = Column(Float, nullable=True)
learning_rate = Column(Float, nullable=True)
metrics = Column(JSON, default=dict)
# Output
output_path = Column(String(512), nullable=True)
hub_model_id = Column(String(255), nullable=True)
model_card = Column(Text, nullable=True)
# Error handling
error_message = Column(Text, nullable=True)
traceback = Column(Text, nullable=True)
retry_count = Column(Integer, default=0)
max_retries = Column(Integer, default=3)
# Timestamps
created_at = Column(DateTime, default=datetime.utcnow)
started_at = Column(DateTime, nullable=True)
completed_at = Column(DateTime, nullable=True)
updated_at = Column(DateTime, default=datetime.utcnow, onupdate=datetime.utcnow)
# User info
created_by = Column(String(100), nullable=True)
tags = Column(JSON, default=list)
# Relationships
checkpoints = relationship("Checkpoint", back_populates="job", cascade="all, delete-orphan")
logs = relationship("TrainingLog", back_populates="job", cascade="all, delete-orphan")
def to_dict(self):
return {
"id": self.id,
"job_id": self.job_id,
"name": self.name,
"description": self.description,
"task_type": self.task_type,
"base_model": self.base_model,
"output_model_name": self.output_model_name,
"dataset_name": self.dataset_name,
"status": self.status,
"progress": self.progress,
"current_epoch": self.current_epoch,
"total_epochs": self.total_epochs,
"current_step": self.current_step,
"total_steps": self.total_steps,
"train_loss": self.train_loss,
"eval_loss": self.eval_loss,
"metrics": self.metrics,
"output_path": self.output_path,
"hub_model_id": self.hub_model_id,
"error_message": self.error_message,
"created_at": self.created_at.isoformat() if self.created_at else None,
"started_at": self.started_at.isoformat() if self.started_at else None,
"completed_at": self.completed_at.isoformat() if self.completed_at else None,
"tags": self.tags
}
class Checkpoint(Base):
"""Model for training checkpoints."""
__tablename__ = "checkpoints"
id = Column(Integer, primary_key=True, index=True)
job_id = Column(Integer, ForeignKey("training_jobs.id"), nullable=False)
checkpoint_name = Column(String(255), nullable=False)
checkpoint_path = Column(String(512), nullable=False)
step = Column(Integer, nullable=False)
epoch = Column(Float, nullable=False)
loss = Column(Float, nullable=True)
metrics = Column(JSON, default=dict)
is_best = Column(Boolean, default=False)
created_at = Column(DateTime, default=datetime.utcnow)
size_mb = Column(Float, nullable=True)
# Relationship
job = relationship("TrainingJob", back_populates="checkpoints")
class TrainingLog(Base):
"""Model for training logs."""
__tablename__ = "training_logs"
id = Column(Integer, primary_key=True, index=True)
job_id = Column(Integer, ForeignKey("training_jobs.id"), nullable=False)
level = Column(String(10), default="INFO")
message = Column(Text, nullable=False)
step = Column(Integer, nullable=True)
epoch = Column(Float, nullable=True)
loss = Column(Float, nullable=True)
learning_rate = Column(Float, nullable=True)
metrics = Column(JSON, nullable=True)
created_at = Column(DateTime, default=datetime.utcnow)
# Relationship
job = relationship("TrainingJob", back_populates="logs")
class ModelRegistry(Base):
"""Registry of trained and available models."""
__tablename__ = "model_registry"
id = Column(Integer, primary_key=True, index=True)
name = Column(String(255), unique=True, nullable=False)
model_id = Column(String(255), nullable=False)
task_type = Column(String(50), nullable=False)
description = Column(Text, nullable=True)
tags = Column(JSON, default=list)
parameters = Column(String(20), nullable=True)
is_local = Column(Boolean, default=False)
local_path = Column(String(512), nullable=True)
hub_url = Column(String(512), nullable=True)
is_trained = Column(Boolean, default=False)
training_job_id = Column(Integer, ForeignKey("training_jobs.id"), nullable=True)
created_at = Column(DateTime, default=datetime.utcnow)
last_used = Column(DateTime, nullable=True)
class DatasetCache(Base):
"""Cache for downloaded datasets."""
__tablename__ = "dataset_cache"
id = Column(Integer, primary_key=True, index=True)
name = Column(String(255), unique=True, nullable=False)
config = Column(String(100), nullable=True)
split = Column(String(50), nullable=True)
local_path = Column(String(512), nullable=False)
size_mb = Column(Float, nullable=True)
num_samples = Column(Integer, nullable=True)
features = Column(JSON, nullable=True)
created_at = Column(DateTime, default=datetime.utcnow)
last_accessed = Column(DateTime, default=datetime.utcnow)
async def init_db():
"""Initialize database tables."""
async with engine.begin() as conn:
await conn.run_sync(Base.metadata.create_all)
async def get_db():
"""Get database session."""
async with AsyncSessionLocal() as session:
try:
yield session
finally:
await session.close() |