Skip to content

GitLab

Commit c80f5968 authored by 王敏's avatar 王敏
Browse files

Merge remote-tracking branch 'origin/v0.15.1-dev' into v0.15.1-dev 

# Conflicts:
#	vllm/model_executor/layers/fused_moe/config.py
#	vllm/model_executor/layers/fused_moe/layer.py
#	vllm/model_executor/layers/quantization/compressed_tensors/compressed_tensors_marlin.py
parents 74306deb 530e785f
Hide whitespace changes
Inline Side-by-side
Showing with 43 additions and 43 deletions
vllm/model_executor/models/ouro.py
View file @ c80f5968
...@@ -357,7 +357,7 @@ class OuroModel(nn.Module): ...@@ -357,7 +357,7 @@ class OuroModel(nn.Module):
def forward( def forward(
self, self,
input_ids: torch.Tensor | None, input_ids: torch.Tensor,
positions: torch.Tensor, positions: torch.Tensor,
intermediate_tensors: IntermediateTensors | None = None, intermediate_tensors: IntermediateTensors | None = None,
inputs_embeds: torch.Tensor | None = None, inputs_embeds: torch.Tensor | None = None,
...@@ -482,7 +482,7 @@ class OuroForCausalLM(nn.Module, SupportsLoRA): ...@@ -482,7 +482,7 @@ class OuroForCausalLM(nn.Module, SupportsLoRA):
def forward( def forward(
self, self,
input_ids: torch.Tensor | None, input_ids: torch.Tensor,
positions: torch.Tensor, positions: torch.Tensor,
intermediate_tensors: IntermediateTensors | None = None, intermediate_tensors: IntermediateTensors | None = None,
inputs_embeds: torch.Tensor | None = None, inputs_embeds: torch.Tensor | None = None,
...@@ -504,4 +504,4 @@ class OuroForCausalLM(nn.Module, SupportsLoRA): ...@@ -504,4 +504,4 @@ class OuroForCausalLM(nn.Module, SupportsLoRA):
self, self,
skip_prefixes=(["lm_head."] if self.config.tie_word_embeddings else None), skip_prefixes=(["lm_head."] if self.config.tie_word_embeddings else None),
) )
return loader.load_weights(weights) return loader.load_weights(weights)
\ No newline at end of file
vllm/model_executor/models/ovis.py
View file @ c80f5968
...@@ -525,7 +525,7 @@ class Ovis(nn.Module, SupportsMultiModal, SupportsPP): ...@@ -525,7 +525,7 @@ class Ovis(nn.Module, SupportsMultiModal, SupportsPP):
def forward( def forward(
self, self,
input_ids: torch.Tensor | None, input_ids: torch.Tensor,
positions: torch.Tensor, positions: torch.Tensor,
intermediate_tensors: IntermediateTensors | None = None, intermediate_tensors: IntermediateTensors | None = None,
inputs_embeds: torch.Tensor | None = None, inputs_embeds: torch.Tensor | None = None,
...@@ -552,4 +552,4 @@ class Ovis(nn.Module, SupportsMultiModal, SupportsPP): ...@@ -552,4 +552,4 @@ class Ovis(nn.Module, SupportsMultiModal, SupportsPP):
def load_weights(self, weights: Iterable[tuple[str, torch.Tensor]]) -> set[str]: def load_weights(self, weights: Iterable[tuple[str, torch.Tensor]]) -> set[str]:
loader = AutoWeightsLoader(self) loader = AutoWeightsLoader(self)
return loader.load_weights(weights) return loader.load_weights(weights)
\ No newline at end of file
vllm/model_executor/models/ovis2_5.py
View file @ c80f5968
...@@ -632,7 +632,7 @@ class Ovis2_5(nn.Module, SupportsMultiModal, SupportsPP): ...@@ -632,7 +632,7 @@ class Ovis2_5(nn.Module, SupportsMultiModal, SupportsPP):
def forward( def forward(
self, self,
input_ids: torch.Tensor | None, input_ids: torch.Tensor,
positions: torch.Tensor, positions: torch.Tensor,
intermediate_tensors: IntermediateTensors | None = None, intermediate_tensors: IntermediateTensors | None = None,
inputs_embeds: torch.Tensor | None = None, inputs_embeds: torch.Tensor | None = None,
...@@ -659,4 +659,4 @@ class Ovis2_5(nn.Module, SupportsMultiModal, SupportsPP): ...@@ -659,4 +659,4 @@ class Ovis2_5(nn.Module, SupportsMultiModal, SupportsPP):
def load_weights(self, weights: Iterable[tuple[str, torch.Tensor]]) -> set[str]: def load_weights(self, weights: Iterable[tuple[str, torch.Tensor]]) -> set[str]:
loader = AutoWeightsLoader(self) loader = AutoWeightsLoader(self)
return loader.load_weights(weights) return loader.load_weights(weights)
\ No newline at end of file
vllm/model_executor/models/paddleocr_vl.py
View file @ c80f5968
...@@ -1159,7 +1159,7 @@ class PaddleOCRVLForConditionalGeneration(nn.Module, SupportsMultiModal, Support ...@@ -1159,7 +1159,7 @@ class PaddleOCRVLForConditionalGeneration(nn.Module, SupportsMultiModal, Support
def forward( def forward(
self, self,
input_ids: torch.Tensor | None, input_ids: torch.Tensor,
positions: torch.Tensor, positions: torch.Tensor,
intermediate_tensors: IntermediateTensors | None = None, intermediate_tensors: IntermediateTensors | None = None,
inputs_embeds: torch.Tensor | None = None, inputs_embeds: torch.Tensor | None = None,
...@@ -1227,4 +1227,4 @@ class PaddleOCRVLForConditionalGeneration(nn.Module, SupportsMultiModal, Support ...@@ -1227,4 +1227,4 @@ class PaddleOCRVLForConditionalGeneration(nn.Module, SupportsMultiModal, Support
def load_weights(self, weights: Iterable[tuple[str, torch.Tensor]]) -> set[str]: def load_weights(self, weights: Iterable[tuple[str, torch.Tensor]]) -> set[str]:
loader = AutoWeightsLoader(self) loader = AutoWeightsLoader(self)
autoloaded_weights = loader.load_weights(weights, mapper=self.hf_to_vllm_mapper) autoloaded_weights = loader.load_weights(weights, mapper=self.hf_to_vllm_mapper)
return autoloaded_weights return autoloaded_weights
\ No newline at end of file
vllm/model_executor/models/paligemma.py
View file @ c80f5968
...@@ -389,7 +389,7 @@ class PaliGemmaForConditionalGeneration( ...@@ -389,7 +389,7 @@ class PaliGemmaForConditionalGeneration(
def forward( def forward(
self, self,
input_ids: torch.Tensor | None, input_ids: torch.Tensor,
positions: torch.Tensor, positions: torch.Tensor,
intermediate_tensors: IntermediateTensors | None = None, intermediate_tensors: IntermediateTensors | None = None,
inputs_embeds: torch.Tensor | None = None, inputs_embeds: torch.Tensor | None = None,
...@@ -425,4 +425,4 @@ class PaliGemmaForConditionalGeneration( ...@@ -425,4 +425,4 @@ class PaliGemmaForConditionalGeneration(
return num_image_tokens return num_image_tokens
def get_num_mm_connector_tokens(self, num_vision_tokens: int) -> int: def get_num_mm_connector_tokens(self, num_vision_tokens: int) -> int:
return num_vision_tokens return num_vision_tokens
\ No newline at end of file
vllm/model_executor/models/persimmon.py
View file @ c80f5968
...@@ -271,7 +271,7 @@ class PersimmonModel(nn.Module): ...@@ -271,7 +271,7 @@ class PersimmonModel(nn.Module):
def forward( def forward(
self, self,
input_ids: torch.Tensor | None, input_ids: torch.Tensor,
positions: torch.Tensor, positions: torch.Tensor,
intermediate_tensors: IntermediateTensors | None, intermediate_tensors: IntermediateTensors | None,
inputs_embeds: torch.Tensor | None = None, inputs_embeds: torch.Tensor | None = None,
...@@ -348,7 +348,7 @@ class PersimmonForCausalLM(nn.Module, SupportsPP): ...@@ -348,7 +348,7 @@ class PersimmonForCausalLM(nn.Module, SupportsPP):
def forward( def forward(
self, self,
input_ids: torch.Tensor | None, input_ids: torch.Tensor,
positions: torch.Tensor, positions: torch.Tensor,
intermediate_tensors: IntermediateTensors | None = None, intermediate_tensors: IntermediateTensors | None = None,
inputs_embeds: torch.Tensor | None = None, inputs_embeds: torch.Tensor | None = None,
...@@ -370,4 +370,4 @@ class PersimmonForCausalLM(nn.Module, SupportsPP): ...@@ -370,4 +370,4 @@ class PersimmonForCausalLM(nn.Module, SupportsPP):
def load_weights(self, weights: Iterable[tuple[str, torch.Tensor]]) -> set[str]: def load_weights(self, weights: Iterable[tuple[str, torch.Tensor]]) -> set[str]:
loader = AutoWeightsLoader(self) loader = AutoWeightsLoader(self)
return loader.load_weights(weights) return loader.load_weights(weights)
\ No newline at end of file
vllm/model_executor/models/phi.py
View file @ c80f5968
...@@ -234,7 +234,7 @@ class PhiModel(nn.Module): ...@@ -234,7 +234,7 @@ class PhiModel(nn.Module):
def forward( def forward(
self, self,
input_ids: torch.Tensor | None, input_ids: torch.Tensor,
positions: torch.Tensor, positions: torch.Tensor,
intermediate_tensors: IntermediateTensors | None, intermediate_tensors: IntermediateTensors | None,
inputs_embeds: torch.Tensor | None = None, inputs_embeds: torch.Tensor | None = None,
...@@ -340,7 +340,7 @@ class PhiForCausalLM(nn.Module, SupportsLoRA, SupportsPP): ...@@ -340,7 +340,7 @@ class PhiForCausalLM(nn.Module, SupportsLoRA, SupportsPP):
def forward( def forward(
self, self,
input_ids: torch.Tensor | None, input_ids: torch.Tensor,
positions: torch.Tensor, positions: torch.Tensor,
intermediate_tensors: IntermediateTensors | None = None, intermediate_tensors: IntermediateTensors | None = None,
inputs_embeds: torch.Tensor | None = None, inputs_embeds: torch.Tensor | None = None,
...@@ -360,4 +360,4 @@ class PhiForCausalLM(nn.Module, SupportsLoRA, SupportsPP): ...@@ -360,4 +360,4 @@ class PhiForCausalLM(nn.Module, SupportsLoRA, SupportsPP):
def load_weights(self, weights: Iterable[tuple[str, torch.Tensor]]) -> set[str]: def load_weights(self, weights: Iterable[tuple[str, torch.Tensor]]) -> set[str]:
loader = AutoWeightsLoader(self) loader = AutoWeightsLoader(self)
return loader.load_weights(weights) return loader.load_weights(weights)
\ No newline at end of file
vllm/model_executor/models/phi3v.py
View file @ c80f5968
...@@ -686,7 +686,7 @@ class Phi3VForCausalLM(nn.Module, SupportsMultiModal, SupportsPP, SupportsQuant) ...@@ -686,7 +686,7 @@ class Phi3VForCausalLM(nn.Module, SupportsMultiModal, SupportsPP, SupportsQuant)
def forward( def forward(
self, self,
input_ids: torch.Tensor | None, input_ids: torch.Tensor,
positions: torch.Tensor, positions: torch.Tensor,
intermediate_tensors: IntermediateTensors | None = None, intermediate_tensors: IntermediateTensors | None = None,
inputs_embeds: torch.Tensor | None = None, inputs_embeds: torch.Tensor | None = None,
...@@ -716,4 +716,4 @@ class Phi3VForCausalLM(nn.Module, SupportsMultiModal, SupportsPP, SupportsQuant) ...@@ -716,4 +716,4 @@ class Phi3VForCausalLM(nn.Module, SupportsMultiModal, SupportsPP, SupportsQuant)
if "embed_tokens.weight" not in autoloaded_weights: if "embed_tokens.weight" not in autoloaded_weights:
self.embed_tokens = self.language_model.model.embed_tokens self.embed_tokens = self.language_model.model.embed_tokens
autoloaded_weights.add("embed_tokens.weight") autoloaded_weights.add("embed_tokens.weight")
return autoloaded_weights return autoloaded_weights
\ No newline at end of file
vllm/model_executor/models/phi4mm.py
View file @ c80f5968
...@@ -1211,7 +1211,7 @@ class Phi4MMForCausalLM(nn.Module, SupportsLoRA, SupportsMultiModal): ...@@ -1211,7 +1211,7 @@ class Phi4MMForCausalLM(nn.Module, SupportsLoRA, SupportsMultiModal):
def forward( def forward(
self, self,
input_ids: torch.Tensor | None, input_ids: torch.Tensor,
positions: torch.Tensor, positions: torch.Tensor,
intermediate_tensors: IntermediateTensors | None = None, intermediate_tensors: IntermediateTensors | None = None,
inputs_embeds: torch.Tensor | None = None, inputs_embeds: torch.Tensor | None = None,
...@@ -1248,4 +1248,4 @@ class Phi4MMForCausalLM(nn.Module, SupportsLoRA, SupportsMultiModal): ...@@ -1248,4 +1248,4 @@ class Phi4MMForCausalLM(nn.Module, SupportsLoRA, SupportsMultiModal):
language_model="model.", language_model="model.",
connector=["audio_projection_for_vision", "audio_projection"], connector=["audio_projection_for_vision", "audio_projection"],
tower_model=["vision_encoder", "embed_tokens_extend"], tower_model=["vision_encoder", "embed_tokens_extend"],
) )
\ No newline at end of file
vllm/model_executor/models/phimoe.py
View file @ c80f5968
...@@ -483,7 +483,7 @@ class PhiMoEModel(nn.Module): ...@@ -483,7 +483,7 @@ class PhiMoEModel(nn.Module):
def forward( def forward(
self, self,
input_ids: torch.Tensor | None, input_ids: torch.Tensor,
positions: torch.Tensor, positions: torch.Tensor,
intermediate_tensors: IntermediateTensors | None, intermediate_tensors: IntermediateTensors | None,
inputs_embeds: torch.Tensor | None = None, inputs_embeds: torch.Tensor | None = None,
...@@ -649,7 +649,7 @@ class PhiMoEForCausalLM(nn.Module, SupportsLoRA, SupportsPP): ...@@ -649,7 +649,7 @@ class PhiMoEForCausalLM(nn.Module, SupportsLoRA, SupportsPP):
def forward( def forward(
self, self,
input_ids: torch.Tensor | None, input_ids: torch.Tensor,
positions: torch.Tensor, positions: torch.Tensor,
intermediate_tensors: IntermediateTensors | None = None, intermediate_tensors: IntermediateTensors | None = None,
inputs_embeds: torch.Tensor | None = None, inputs_embeds: torch.Tensor | None = None,
...@@ -668,4 +668,4 @@ class PhiMoEForCausalLM(nn.Module, SupportsLoRA, SupportsPP): ...@@ -668,4 +668,4 @@ class PhiMoEForCausalLM(nn.Module, SupportsLoRA, SupportsPP):
return loader.load_weights(weights) return loader.load_weights(weights)
def get_expert_mapping(self) -> list[tuple[str, str, int, str]]: def get_expert_mapping(self) -> list[tuple[str, str, int, str]]:
return self.model.get_expert_mapping() return self.model.get_expert_mapping()
\ No newline at end of file
vllm/model_executor/models/pixtral.py
View file @ c80f5968
...@@ -479,7 +479,7 @@ class PixtralForConditionalGeneration( ...@@ -479,7 +479,7 @@ class PixtralForConditionalGeneration(
def forward( def forward(
self, self,
input_ids: torch.Tensor | None, input_ids: torch.Tensor,
positions: torch.Tensor, positions: torch.Tensor,
intermediate_tensors: IntermediateTensors | None = None, intermediate_tensors: IntermediateTensors | None = None,
inputs_embeds: torch.Tensor | None = None, inputs_embeds: torch.Tensor | None = None,
...@@ -1413,4 +1413,4 @@ class PixtralHFVisionModel(nn.Module): ...@@ -1413,4 +1413,4 @@ class PixtralHFVisionModel(nn.Module):
weight_loader = getattr(param, "weight_loader", default_weight_loader) weight_loader = getattr(param, "weight_loader", default_weight_loader)
weight_loader(param, loaded_weight) weight_loader(param, loaded_weight)
loaded_params.add(name) loaded_params.add(name)
return loaded_params return loaded_params
\ No newline at end of file
vllm/model_executor/models/plamo2.py
View file @ c80f5968
...@@ -775,7 +775,7 @@ class Plamo2Model(torch.nn.Module): ...@@ -775,7 +775,7 @@ class Plamo2Model(torch.nn.Module):
def forward( def forward(
self, self,
input_ids: torch.Tensor | None, input_ids: torch.Tensor,
positions: torch.Tensor, positions: torch.Tensor,
intermediate_tensors: IntermediateTensors | None = None, intermediate_tensors: IntermediateTensors | None = None,
inputs_embeds: torch.Tensor | None = None, inputs_embeds: torch.Tensor | None = None,
...@@ -852,7 +852,7 @@ class Plamo2ForCausalLM( ...@@ -852,7 +852,7 @@ class Plamo2ForCausalLM(
def forward( def forward(
self, self,
input_ids: torch.Tensor | None, input_ids: torch.Tensor,
positions: torch.Tensor, positions: torch.Tensor,
intermediate_tensors: IntermediateTensors | None = None, intermediate_tensors: IntermediateTensors | None = None,
inputs_embeds: torch.Tensor | None = None, inputs_embeds: torch.Tensor | None = None,
...@@ -996,4 +996,4 @@ class Plamo2ForCausalLM( ...@@ -996,4 +996,4 @@ class Plamo2ForCausalLM(
param = params_dict[name] param = params_dict[name]
weight_loader = getattr(param, "weight_loader", default_weight_loader) weight_loader = getattr(param, "weight_loader", default_weight_loader)
weight_loader(param, loaded_weight) weight_loader(param, loaded_weight)
\ No newline at end of file
vllm/model_executor/models/plamo3.py
View file @ c80f5968
...@@ -342,7 +342,7 @@ class Plamo3Model(nn.Module): ...@@ -342,7 +342,7 @@ class Plamo3Model(nn.Module):
def forward( def forward(
self, self,
input_ids: torch.Tensor | None, input_ids: torch.Tensor,
positions: torch.Tensor, positions: torch.Tensor,
intermediate_tensors: IntermediateTensors | None = None, intermediate_tensors: IntermediateTensors | None = None,
inputs_embeds: torch.Tensor | None = None, inputs_embeds: torch.Tensor | None = None,
...@@ -412,7 +412,7 @@ class Plamo3ForCausalLM(nn.Module, SupportsLoRA, SupportsPP): ...@@ -412,7 +412,7 @@ class Plamo3ForCausalLM(nn.Module, SupportsLoRA, SupportsPP):
def forward( def forward(
self, self,
input_ids: torch.Tensor | None, input_ids: torch.Tensor,
positions: torch.Tensor, positions: torch.Tensor,
intermediate_tensors: IntermediateTensors | None = None, intermediate_tensors: IntermediateTensors | None = None,
inputs_embeds: torch.Tensor | None = None, inputs_embeds: torch.Tensor | None = None,
...@@ -434,4 +434,4 @@ class Plamo3ForCausalLM(nn.Module, SupportsLoRA, SupportsPP): ...@@ -434,4 +434,4 @@ class Plamo3ForCausalLM(nn.Module, SupportsLoRA, SupportsPP):
self, self,
skip_prefixes=(["lm_head."] if self.config.tie_word_embeddings else None), skip_prefixes=(["lm_head."] if self.config.tie_word_embeddings else None),
) )
return loader.load_weights(weights) return loader.load_weights(weights)
\ No newline at end of file
vllm/model_executor/models/qwen.py
View file @ c80f5968
...@@ -243,7 +243,7 @@ class QWenModel(nn.Module): ...@@ -243,7 +243,7 @@ class QWenModel(nn.Module):
def forward( def forward(
self, self,
input_ids: torch.Tensor | None, input_ids: torch.Tensor,
positions: torch.Tensor, positions: torch.Tensor,
intermediate_tensors: IntermediateTensors | None, intermediate_tensors: IntermediateTensors | None,
inputs_embeds: torch.Tensor | None = None, inputs_embeds: torch.Tensor | None = None,
...@@ -425,7 +425,7 @@ class QWenLMHeadModel(QWenBaseModel, SupportsPP, SupportsLoRA): ...@@ -425,7 +425,7 @@ class QWenLMHeadModel(QWenBaseModel, SupportsPP, SupportsLoRA):
def forward( def forward(
self, self,
input_ids: torch.Tensor | None, input_ids: torch.Tensor,
positions: torch.Tensor, positions: torch.Tensor,
intermediate_tensors: IntermediateTensors | None = None, intermediate_tensors: IntermediateTensors | None = None,
inputs_embeds: torch.Tensor | None = None, inputs_embeds: torch.Tensor | None = None,
......
vllm/model_executor/models/qwen2.py
View file @ c80f5968
...@@ -439,7 +439,7 @@ class Qwen2Model(nn.Module): ...@@ -439,7 +439,7 @@ class Qwen2Model(nn.Module):
def forward( def forward(
self, self,
input_ids: torch.Tensor | None, input_ids: torch.Tensor,
positions: torch.Tensor, positions: torch.Tensor,
intermediate_tensors: IntermediateTensors | None = None, intermediate_tensors: IntermediateTensors | None = None,
inputs_embeds: torch.Tensor | None = None, inputs_embeds: torch.Tensor | None = None,
...@@ -659,7 +659,7 @@ class Qwen2ForCausalLM(nn.Module, SupportsLoRA, SupportsPP, SupportsEagle3): ...@@ -659,7 +659,7 @@ class Qwen2ForCausalLM(nn.Module, SupportsLoRA, SupportsPP, SupportsEagle3):
def forward( def forward(
self, self,
input_ids: torch.Tensor | None, input_ids: torch.Tensor,
positions: torch.Tensor, positions: torch.Tensor,
intermediate_tensors: IntermediateTensors | None = None, intermediate_tensors: IntermediateTensors | None = None,
inputs_embeds: torch.Tensor | None = None, inputs_embeds: torch.Tensor | None = None,
......
vllm/model_executor/models/qwen2_5_omni_thinker.py
View file @ c80f5968
...@@ -1298,7 +1298,7 @@ class Qwen2_5OmniThinkerForConditionalGeneration( ...@@ -1298,7 +1298,7 @@ class Qwen2_5OmniThinkerForConditionalGeneration(
def forward( def forward(
self, self,
input_ids: torch.Tensor | None, input_ids: torch.Tensor,
positions: torch.Tensor, positions: torch.Tensor,
intermediate_tensors: IntermediateTensors | None = None, intermediate_tensors: IntermediateTensors | None = None,
inputs_embeds: torch.Tensor | None = None, inputs_embeds: torch.Tensor | None = None,
...@@ -1330,4 +1330,4 @@ class Qwen2_5OmniThinkerForConditionalGeneration( ...@@ -1330,4 +1330,4 @@ class Qwen2_5OmniThinkerForConditionalGeneration(
language_model="language_model", language_model="language_model",
connector="merger.", connector="merger.",
tower_model=["visual.", "audio_tower."], tower_model=["visual.", "audio_tower."],
) )
\ No newline at end of file
vllm/model_executor/models/qwen2_audio.py
View file @ c80f5968
...@@ -451,7 +451,7 @@ class Qwen2AudioForConditionalGeneration(nn.Module, SupportsMultiModal, Supports ...@@ -451,7 +451,7 @@ class Qwen2AudioForConditionalGeneration(nn.Module, SupportsMultiModal, Supports
def forward( def forward(
self, self,
input_ids: torch.Tensor | None, input_ids: torch.Tensor,
positions: torch.Tensor, positions: torch.Tensor,
intermediate_tensors: IntermediateTensors | None = None, intermediate_tensors: IntermediateTensors | None = None,
inputs_embeds: torch.Tensor | None = None, inputs_embeds: torch.Tensor | None = None,
......
vllm/model_executor/models/qwen2_moe.py
View file @ c80f5968
...@@ -408,7 +408,7 @@ class Qwen2MoeModel(nn.Module): ...@@ -408,7 +408,7 @@ class Qwen2MoeModel(nn.Module):
def forward( def forward(
self, self,
input_ids: torch.Tensor | None, input_ids: torch.Tensor,
positions: torch.Tensor, positions: torch.Tensor,
intermediate_tensors: IntermediateTensors | None = None, intermediate_tensors: IntermediateTensors | None = None,
inputs_embeds: torch.Tensor | None = None, inputs_embeds: torch.Tensor | None = None,
...@@ -633,7 +633,7 @@ class Qwen2MoeForCausalLM(nn.Module, SupportsPP, SupportsLoRA): ...@@ -633,7 +633,7 @@ class Qwen2MoeForCausalLM(nn.Module, SupportsPP, SupportsLoRA):
def forward( def forward(
self, self,
input_ids: torch.Tensor | None, input_ids: torch.Tensor,
positions: torch.Tensor, positions: torch.Tensor,
intermediate_tensors: IntermediateTensors | None = None, intermediate_tensors: IntermediateTensors | None = None,
inputs_embeds: torch.Tensor | None = None, inputs_embeds: torch.Tensor | None = None,
......
vllm/model_executor/models/qwen2_rm.py
View file @ c80f5968
...@@ -79,7 +79,7 @@ class Qwen2RewardBaseModel(nn.Module, SupportsLoRA, SupportsPP): ...@@ -79,7 +79,7 @@ class Qwen2RewardBaseModel(nn.Module, SupportsLoRA, SupportsPP):
def forward( def forward(
self, self,
input_ids: torch.Tensor | None, input_ids: torch.Tensor,
positions: torch.Tensor, positions: torch.Tensor,
intermediate_tensors: IntermediateTensors | None = None, intermediate_tensors: IntermediateTensors | None = None,
inputs_embeds: torch.Tensor | None = None, inputs_embeds: torch.Tensor | None = None,
......
vllm/model_executor/models/qwen2_vl.py
View file @ c80f5968
...@@ -1432,7 +1432,7 @@ class Qwen2VLForConditionalGeneration( ...@@ -1432,7 +1432,7 @@ class Qwen2VLForConditionalGeneration(
def forward( def forward(
self, self,
input_ids: torch.Tensor | None, input_ids: torch.Tensor,
positions: torch.Tensor, positions: torch.Tensor,
intermediate_tensors: IntermediateTensors | None = None, intermediate_tensors: IntermediateTensors | None = None,
inputs_embeds: torch.Tensor | None = None, inputs_embeds: torch.Tensor | None = None,
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment