File size: 6,450 Bytes
fd4a82d
82b1671
fd4a82d
82b1671
fd4a82d
 
82b1671
 
fd4a82d
82b1671
fd4a82d
 
82b1671
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
fd4a82d
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
82b1671
 
 
 
 
fd4a82d
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
82b1671
 
 
 
fd4a82d
 
 
82b1671
fd4a82d
 
 
 
 
 
 
 
82b1671
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
fd4a82d
 
 
82b1671
 
 
 
 
 
 
 
 
fd4a82d
 
 
 
 
 
 
 
 
 
 
82b1671
 
 
 
 
 
 
 
fd4a82d
 
 
 
 
 
 
 
 
 
 
 
82b1671
 
fd4a82d
82b1671
fd4a82d
82b1671
 
fd4a82d
 
 
82b1671
 
fd4a82d
 
 
82b1671
fd4a82d
82b1671
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
# By lllyasviel
# WindowsとHugging Face Space環境の両方に対応した DynamicSwap + zeroGPU 対応バージョン

import os
import torch

# Hugging Face Space環境で実行されているかどうかを確認
IN_HF_SPACE = os.environ.get('SPACE_ID') is not None

# CPU デバイスを設定
cpu = torch.device('cpu')

# ステートレスGPU環境では、メインプロセスでCUDAを初期化しない
def get_gpu_device():
    if IN_HF_SPACE:
        # Spacesではデバイスの初期化を遅延させる
        return 'cuda'
    try:
        if torch.cuda.is_available():
            return torch.device(f'cuda:{torch.cuda.current_device()}')
        else:
            print("CUDAが利用できません。デフォルトデバイスとしてCPUを使用します")
            return torch.device('cpu')
    except Exception as e:
        print(f"CUDAデバイスの初期化中にエラーが発生しました: {e}")
        print("CPUデバイスにフォールバックします")
        return torch.device('cpu')

# GPUデバイスを取得(文字列または実際のデバイスオブジェクト)
gpu = get_gpu_device()

# 完全にGPUにロードされたモジュールのリスト
gpu_complete_modules = []

class DynamicSwapInstaller:
    @staticmethod
    def _install_module(module: torch.nn.Module, **kwargs):
        original_class = module.__class__
        module.__dict__['forge_backup_original_class'] = original_class

        def hacked_get_attr(self, name: str):
            if '_parameters' in self.__dict__:
                _parameters = self.__dict__['_parameters']
                if name in _parameters:
                    p = _parameters[name]
                    if p is None:
                        return None
                    if p.__class__ == torch.nn.Parameter:
                        return torch.nn.Parameter(p.to(**kwargs), requires_grad=p.requires_grad)
                    else:
                        return p.to(**kwargs)
            if '_buffers' in self.__dict__:
                _buffers = self.__dict__['_buffers']
                if name in _buffers:
                    return _buffers[name].to(**kwargs)
            return super(original_class, self).__getattr__(name)

        module.__class__ = type(
            'DynamicSwap_' + original_class.__name__,
            (original_class,),
            {'__getattr__': hacked_get_attr}
        )

    @staticmethod
    def _uninstall_module(module: torch.nn.Module):
        if 'forge_backup_original_class' in module.__dict__:
            module.__class__ = module.__dict__.pop('forge_backup_original_class')

    @staticmethod
    def install_model(model: torch.nn.Module, **kwargs):
        for m in model.modules():
            DynamicSwapInstaller._install_module(m, **kwargs)

    @staticmethod
    def uninstall_model(model: torch.nn.Module):
        for m in model.modules():
            DynamicSwapInstaller._uninstall_module(m)


def fake_diffusers_current_device(model: torch.nn.Module, target_device):
    # 文字列デバイスをtorch.deviceに変換
    if isinstance(target_device, str):
        target_device = torch.device(target_device)
    if hasattr(model, 'scale_shift_table'):
        model.scale_shift_table.data = model.scale_shift_table.data.to(target_device)
        return
    for _, p in model.named_modules():
        if hasattr(p, 'weight'):
            p.to(target_device)
            return


def get_cuda_free_memory_gb(device=None):
    if device is None:
        device = gpu
    if isinstance(device, str):
        device = torch.device(device)
    if device.type != 'cuda':
        # CUDAでない場合はデフォルト値
        return 6.0
    try:
        stats = torch.cuda.memory_stats(device)
        active = stats['active_bytes.all.current']
        reserved = stats['reserved_bytes.all.current']
        free_cuda, _ = torch.cuda.mem_get_info(device)
        inactive = reserved - active
        available = free_cuda + inactive
        return available / (1024 ** 3)
    except Exception as e:
        print(f"CUDAメモリ情報取得エラー: {e}")
        return 6.0


def move_model_to_device_with_memory_preservation(model, target_device, preserved_memory_gb=0):
    print(f"{model.__class__.__name__}{target_device} に移動します。保持メモリ: {preserved_memory_gb} GB")
    if isinstance(target_device, str):
        target_device = torch.device(target_device)
    # CPUまたはGPU未使用時は直接移動
    if target_device.type == 'cpu':
        model.to(device=target_device)
        if torch.cuda.is_available():
            torch.cuda.empty_cache()
        return
    for m in model.modules():
        if get_cuda_free_memory_gb(target_device) <= preserved_memory_gb:
            torch.cuda.empty_cache()
            return
        if hasattr(m, 'weight'):
            m.to(device=target_device)
    model.to(device=target_device)
    torch.cuda.empty_cache()


def offload_model_from_device_for_memory_preservation(model, target_device, preserved_memory_gb=0):
    print(f"メモリ保持のため {model.__class__.__name__}{target_device} からオフロードします: {preserved_memory_gb} GB")
    if isinstance(target_device, str):
        target_device = torch.device(target_device)
    if target_device.type == 'cpu':
        model.to(device=cpu)
        if torch.cuda.is_available():
            torch.cuda.empty_cache()
        return
    for m in model.modules():
        if get_cuda_free_memory_gb(target_device) >= preserved_memory_gb:
            torch.cuda.empty_cache()
            return
        if hasattr(m, 'weight'):
            m.to(device=cpu)
    model.to(device=cpu)
    torch.cuda.empty_cache()


def unload_complete_models(*args):
    for m in gpu_complete_modules + list(args):
        if m is None:
            continue
        m.to(device=cpu)
        print(f"{m.__class__.__name__} を完全にアンロードしました")
    gpu_complete_modules.clear()
    if torch.cuda.is_available():
        torch.cuda.empty_cache()


def load_model_as_complete(model, target_device, unload=True):
    if isinstance(target_device, str):
        target_device = torch.device(target_device)
    if unload:
        unload_complete_models()
    model.to(device=target_device)
    print(f"{model.__class__.__name__}{target_device} に完全にロードしました")
    gpu_complete_modules.append(model)