biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

[V1] Make v1 more testable (#9888)

Browse Source 
Signed-off-by: Joe Runde <Joseph.Runde@ibm.com>
This commit is contained in:
Joe Runde
2024-11-06 12:57:35 -07:00
committed by GitHub
parent 87bd7e0515
commit d58268c56a
75 changed files with 243 additions and 165 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
vllm/model_executor/models/llava_next.py
View File

@@ -16,7 +16,7 @@ from vllm.inputs import (INPUT_REGISTRY, DecoderOnlyInputs, DummyData,
InputContext)
from vllm.model_executor.layers.pooler import Pooler, PoolingType
from vllm.model_executor.layers.quantization import QuantizationConfig
from vllm.model_executor.layers.sampler import Sampler, SamplerOutput
from vllm.model_executor.layers.sampler import SamplerOutput, get_sampler
from vllm.model_executor.pooling_metadata import PoolingMetadata
from vllm.model_executor.sampling_metadata import SamplingMetadata
from vllm.multimodal import MULTIMODAL_REGISTRY
@@ -327,7 +327,7 @@ class LlavaNextForConditionalGeneration(nn.Module, SupportsMultiModal,
if hasattr(self.language_model, "sampler"):
return self.language_model.sampler
return Sampler()
return get_sampler()
def _validate_image_sizes(self, data: torch.Tensor) -> torch.Tensor:
expected_dims = (2, )