Spaces:
Running
Running
File size: 7,329 Bytes
b7cc217 914fbd3 b7cc217 914fbd3 b7cc217 914fbd3 b7cc217 914fbd3 b7cc217 914fbd3 b7cc217 |
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 |
"""
Models service for Video Model Studio
Handles the model history tracking and management
"""
import logging
import json
from pathlib import Path
from typing import Dict, Any, List, Optional, Tuple
from dataclasses import dataclass
from datetime import datetime
from vms.config import (
STORAGE_PATH, MODEL_TYPES, TRAINING_TYPES
)
logger = logging.getLogger(__name__)
logger.setLevel(logging.INFO)
@dataclass
class Model:
"""Class for tracking model metadata"""
id: str
status: str # 'draft', 'training', 'trained', 'error'
model_type: str # Base model family (e.g. 'hunyuan_video', 'ltx_video', 'wan')
model_display_name: str # Display name for the model type
created_at: datetime
updated_at: datetime
training_progress: Optional[float] = 0.0 # Progress as percentage
current_step: Optional[int] = 0
total_steps: Optional[int] = 0
@classmethod
def from_dir(cls, model_dir: Path) -> 'Model':
"""Create a Model instance from a directory"""
model_id = model_dir.name
# Default values
status = 'draft'
model_type = ''
model_display_name = ''
created_at = datetime.fromtimestamp(model_dir.stat().st_ctime)
updated_at = datetime.fromtimestamp(model_dir.stat().st_mtime)
training_progress = 0.0
current_step = 0
total_steps = 0
# Check for UI state file
ui_state_file = model_dir / "output" / "ui_state.json"
if ui_state_file.exists():
try:
with open(ui_state_file, 'r') as f:
ui_state = json.load(f)
status = ui_state.get('project_status', 'draft')
# Get model type from UI state
model_type_value = ui_state.get('model_type', '')
# First check if model_type_value is a display name
display_name_found = False
for display_name, internal_name in MODEL_TYPES.items():
if display_name == model_type_value:
model_type = internal_name
model_display_name = display_name
display_name_found = True
break
# If not a display name, check if it's an internal name
if not display_name_found:
for display_name, internal_name in MODEL_TYPES.items():
if internal_name == model_type_value:
model_type = internal_name
model_display_name = display_name
break
except Exception as e:
logger.error(f"Error loading UI state for model {model_id}: {str(e)}")
# Check for status file to get training progress
status_file = model_dir / "output" / "status.json"
if status_file.exists():
try:
with open(status_file, 'r') as f:
status_data = json.load(f)
if status_data.get('status') == 'training':
status = 'training'
current_step = status_data.get('step', 0)
total_steps = status_data.get('total_steps', 0)
if total_steps > 0:
training_progress = (current_step / total_steps) * 100
elif status_data.get('status') == 'completed':
status = 'trained'
training_progress = 100.0
elif status_data.get('status') == 'error':
status = 'error'
except Exception as e:
logger.error(f"Error loading status for model {model_id}: {str(e)}")
# Check for pid file to determine if training is active
pid_file = model_dir / "output" / "training.pid"
if pid_file.exists():
status = 'training'
# Check for model weights to determine if trained
model_weights = model_dir / "output" / "pytorch_lora_weights.safetensors"
if model_weights.exists() and status != 'training':
status = 'trained'
training_progress = 100.0
return cls(
id=model_id,
status=status,
model_type=model_type,
model_display_name=model_display_name,
created_at=created_at,
updated_at=updated_at,
training_progress=training_progress,
current_step=current_step,
total_steps=total_steps
)
class ModelsService:
"""Service for tracking and managing model history"""
def __init__(self, app_state=None):
"""Initialize the models service
Args:
app_state: Reference to main application state
"""
self.app = app_state
def get_all_models(self) -> List[Model]:
"""Get a list of all models
Returns:
List of Model objects
"""
models_dir = STORAGE_PATH / "models"
if not models_dir.exists():
return []
models = []
for model_dir in models_dir.iterdir():
if not model_dir.is_dir():
continue
try:
model = Model.from_dir(model_dir)
models.append(model)
except Exception as e:
logger.error(f"Error loading model from {model_dir}: {str(e)}")
# Sort models by updated_at (newest first)
return sorted(models, key=lambda m: m.updated_at, reverse=True)
def get_draft_models(self) -> List[Model]:
"""Get a list of draft models
Returns:
List of Model objects with 'draft' status
"""
return [m for m in self.get_all_models() if m.status == 'draft']
def get_training_models(self) -> List[Model]:
"""Get a list of models currently in training
Returns:
List of Model objects with 'training' status
"""
return [m for m in self.get_all_models() if m.status == 'training']
def get_trained_models(self) -> List[Model]:
"""Get a list of completed trained models
Returns:
List of Model objects with 'trained' status
"""
return [m for m in self.get_all_models() if m.status == 'trained']
def delete_model(self, model_id: str) -> bool:
"""Delete a model by ID
Args:
model_id: The model ID to delete
Returns:
True if deletion was successful
"""
if not model_id:
return False
model_dir = STORAGE_PATH / "models" / model_id
if not model_dir.exists():
return False
try:
import shutil
shutil.rmtree(model_dir)
return True
except Exception as e:
logger.error(f"Error deleting model {model_id}: {str(e)}")
return False |