MCPcopy Create free account
hub / github.com/huggingface/diffusers / GroupOffloadingConfig

Class GroupOffloadingConfig

src/diffusers/hooks/group_offloading.py:100–112  ·  view source on GitHub ↗

Source from the content-addressed store, hash-verified

98
99@dataclass
100class GroupOffloadingConfig:
101 onload_device: torch.device
102 offload_device: torch.device
103 offload_type: GroupOffloadingType
104 non_blocking: bool
105 record_stream: bool
106 low_cpu_mem_usage: bool
107 num_blocks_per_group: int | None = None
108 offload_to_disk_path: str | None = None
109 stream: torch.cuda.Stream | torch.Stream | None = None
110 block_modules: list[str] | None = None
111 exclude_kwargs: list[str] | None = None
112 module_prefix: str = ""
113
114
115class ModuleGroup:

Callers 1

apply_group_offloadingFunction · 0.85

Calls

no outgoing calls

Tested by

no test coverage detected

Used in the wild real call sites across dependent graphs

searching dependent graphs…