[Bugfix] Reject pipeline parallelism for DiffusionGemma (#45828)
Signed-off-by: Guan-Ming (Wesley) Chiu <105915352+guan404ming@users.noreply.github.com>
This commit is contained in:
@@ -61,7 +61,6 @@ from vllm.v1.worker.gpu.states import RequestState
|
||||
|
||||
from .interfaces import (
|
||||
SupportsMultiModal,
|
||||
SupportsPP,
|
||||
SupportsQuant,
|
||||
)
|
||||
|
||||
@@ -143,7 +142,6 @@ class DiffusionGemmaForConditionalGeneration(
|
||||
nn.Module,
|
||||
SupportsMultiModal,
|
||||
SupportsQuant,
|
||||
SupportsPP,
|
||||
):
|
||||
"""DiffusionGemma for vLLM.
|
||||
|
||||
@@ -266,10 +264,6 @@ class DiffusionGemmaForConditionalGeneration(
|
||||
eps=getattr(text_config, "rms_norm_eps", 1e-6),
|
||||
)
|
||||
|
||||
self.make_empty_intermediate_tensors = (
|
||||
self.model.make_empty_intermediate_tensors
|
||||
)
|
||||
|
||||
def compute_self_conditioning(
|
||||
self,
|
||||
inputs_embeds: torch.Tensor,
|
||||
|
||||
Reference in New Issue
Block a user