-
-
Notifications
You must be signed in to change notification settings - Fork 202
Expand file tree
/
Copy path__init__.py
More file actions
45 lines (38 loc) · 1.55 KB
/
__init__.py
File metadata and controls
45 lines (38 loc) · 1.55 KB
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
"""
Low-VRAM Training Framework for GO-1
=====================================
Clean Architecture Package Structure:
low_vram/
├── core/ # Domain Layer (Pure business logic)
│ ├── interfaces/ # Abstract interfaces (Dependency Inversion)
│ ├── entities/ # Domain entities
│ └── use_cases/ # Application business rules
├── adapters/ # Interface Adapters Layer
│ ├── memory/ # Memory management implementations
│ └── training/ # Training strategy implementations
└── infrastructure/ # Frameworks & Drivers Layer
├── pytorch/ # PyTorch-specific implementations
└── config/ # Configuration handling
Design Principles Applied:
- SOLID Principles (especially Dependency Inversion)
- Clean Architecture (layered separation)
- Single Responsibility (one class = one job)
- Interface Segregation (small, focused interfaces)
- DRY (shared utilities extracted)
Security Considerations (CIA Triad):
- Integrity: Checksum validation for cached features
- Availability: Graceful degradation when memory is low
- Confidentiality: No sensitive data logged
"""
__version__ = "0.1.0"
__author__ = "AgiBot-World Contributors"
from go1.tools.low_vram.core.interfaces import (
MemoryManager,
TrainingStrategy,
FeatureCache,
)
__all__ = [
"MemoryManager",
"TrainingStrategy",
"FeatureCache",
]