[Feature]: Native GGUF Quantization Support for DiT (#1285)

Signed-off-by: David Chen <530634352@qq.com>
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
Signed-off-by: WeiQing Chen <40507679+david6666666@users.noreply.github.com>
Signed-off-by: Hongsheng Liu <liuhongsheng4@huawei.com>
Co-authored-by: Isotr0py <mozf@mail2.sysu.edu.cn>
Co-authored-by: Isotr0py <2037008807@qq.com>
Co-authored-by: Hongsheng Liu <liuhongsheng4@huawei.com>
This commit is contained in:
authored and GitHub committed 2026-02-27 11:39:14 +08:00
1 parent b19c227fb3
commit 0b884ebc6c
23 files changed
+1104 -67

No files matched your search

@@ -28,6 +28,7 @@ from vllm.logger import init_logger
from .base import DiffusionQuantizationConfig
from .fp8 import DiffusionFp8Config
from .gguf import DiffusionGgufConfig
if TYPE_CHECKING:
from vllm.model_executor.layers.quantization.base_config import (
@@ -40,6 +41,7 @@ logger = init_logger(__name__)
# To add a new method, create a new config class and register it here
_QUANT_CONFIG_REGISTRY: dict[str, type[DiffusionQuantizationConfig]] = {
"fp8": DiffusionFp8Config,
"gguf": DiffusionGgufConfig,
}
SUPPORTED_QUANTIZATION_METHODS = list(_QUANT_CONFIG_REGISTRY.keys())
@@ -108,6 +110,7 @@ def get_vllm_quant_config_for_layers(
__all__ = [
"DiffusionQuantizationConfig",
"DiffusionFp8Config",
"DiffusionGgufConfig",
"get_diffusion_quant_config",
"get_vllm_quant_config_for_layers",
"SUPPORTED_QUANTIZATION_METHODS",