[V1] Get supported tasks from model runner instead of model config (#21585)

Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
This commit is contained in:
Cyrus Leung
2025-07-25 05:36:45 -07:00
committed by GitHub
parent 5c3f2628d5
commit 46d81d6951
19 changed files with 200 additions and 54 deletions
+1 -1
View File
@@ -26,8 +26,8 @@ from vllm.model_executor.layers.quantization import QuantizationConfig
from vllm.model_executor.layers.vocab_parallel_embedding import (
VocabParallelEmbedding)
from vllm.model_executor.pooling_metadata import PoolingMetadata
from vllm.pooling_params import PoolingTask
from vllm.sequence import IntermediateTensors
from vllm.tasks import PoolingTask
from .interfaces import SupportsCrossEncoding, SupportsQuant, SupportsV0Only
from .utils import AutoWeightsLoader, WeightsMapper, maybe_prefix
+1 -1
View File
@@ -16,8 +16,8 @@ from vllm.model_executor.layers.pooler import (DispatchPooler, Pooler,
get_prompt_token_ids)
from vllm.model_executor.models.llama import LlamaForCausalLM
from vllm.model_executor.pooling_metadata import PoolingMetadata
from vllm.pooling_params import PoolingTask
from vllm.sequence import PoolerOutput
from vllm.tasks import PoolingTask
from vllm.transformers_utils.tokenizer import cached_tokenizer_from_config
from .interfaces import SupportsV0Only
+1 -1
View File
@@ -23,8 +23,8 @@ from vllm.model_executor.layers.vocab_parallel_embedding import (
VocabParallelEmbedding)
from vllm.model_executor.model_loader.weight_utils import default_weight_loader
from vllm.model_executor.pooling_metadata import PoolingMetadata
from vllm.pooling_params import PoolingTask
from vllm.sequence import IntermediateTensors
from vllm.tasks import PoolingTask
from .interfaces import SupportsCrossEncoding, SupportsV0Only
from .utils import WeightsMapper, maybe_prefix