[Misc]] Move processing context to multimodal directory (#25548)
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
This commit is contained in:
@@ -12,11 +12,11 @@ from mistral_common.protocol.instruct.request import ChatCompletionRequest
|
||||
from PIL import Image
|
||||
|
||||
from vllm.config import ModelConfig
|
||||
from vllm.inputs import InputProcessingContext
|
||||
from vllm.multimodal import MULTIMODAL_REGISTRY, MultiModalDataDict
|
||||
from vllm.multimodal.cache import MultiModalProcessorOnlyCache
|
||||
from vllm.multimodal.inputs import MultiModalInputs
|
||||
from vllm.multimodal.processing import BaseMultiModalProcessor
|
||||
from vllm.multimodal.processing import (BaseMultiModalProcessor,
|
||||
InputProcessingContext)
|
||||
from vllm.transformers_utils.tokenizer import (AnyTokenizer, MistralTokenizer,
|
||||
cached_tokenizer_from_config,
|
||||
encode_tokens)
|
||||
|
||||
@@ -18,10 +18,10 @@ from vllm.config import ModelConfig, VllmConfig, set_current_vllm_config
|
||||
from vllm.distributed import (cleanup_dist_env_and_memory,
|
||||
init_distributed_environment,
|
||||
initialize_model_parallel)
|
||||
from vllm.inputs import InputProcessingContext
|
||||
from vllm.model_executor.model_loader.utils import set_default_torch_dtype
|
||||
from vllm.multimodal import MULTIMODAL_REGISTRY, BatchedTensorInputs
|
||||
from vllm.multimodal.processing import BaseMultiModalProcessor
|
||||
from vllm.multimodal.processing import (BaseMultiModalProcessor,
|
||||
InputProcessingContext)
|
||||
from vllm.multimodal.utils import group_mm_kwargs_by_modality
|
||||
from vllm.transformers_utils.tokenizer import cached_tokenizer_from_config
|
||||
from vllm.utils import is_list_of
|
||||
|
||||
@@ -11,8 +11,9 @@ import torch.nn.functional as F
|
||||
from transformers import PretrainedConfig
|
||||
|
||||
from vllm.config import ModelConfig, ModelDType, RunnerOption
|
||||
from vllm.inputs import InputContext
|
||||
from vllm.logprobs import Logprob, PromptLogprobs, SampleLogprobs
|
||||
from vllm.multimodal.processing import InputProcessingContext
|
||||
from vllm.transformers_utils.tokenizer import cached_tokenizer_from_config
|
||||
|
||||
from .registry import HF_EXAMPLE_MODELS
|
||||
|
||||
@@ -264,7 +265,7 @@ def build_model_context(
|
||||
limit_mm_per_prompt: Optional[dict[str, int]] = None,
|
||||
mm_processor_cache_gb: int = 0,
|
||||
):
|
||||
"""Creates an InputContext for a given model.
|
||||
"""Creates an InputProcessingContext for a given model.
|
||||
|
||||
Args:
|
||||
model_id: ID of the model being considered.
|
||||
@@ -273,7 +274,7 @@ def build_model_context(
|
||||
limit_mm_per_prompt: Multimodal limits.
|
||||
|
||||
Returns:
|
||||
InputContext for the model being considered.
|
||||
InputProcessingContext for the model being considered.
|
||||
"""
|
||||
model_info = HF_EXAMPLE_MODELS.find_hf_info(model_id)
|
||||
model_info.check_available_online(on_fail="skip")
|
||||
@@ -298,7 +299,11 @@ def build_model_context(
|
||||
enforce_eager=model_info.enforce_eager,
|
||||
**model_config_kwargs,
|
||||
)
|
||||
return InputContext(model_config)
|
||||
|
||||
return InputProcessingContext(
|
||||
model_config,
|
||||
tokenizer=cached_tokenizer_from_config(model_config),
|
||||
)
|
||||
|
||||
|
||||
def check_embeddings_close(
|
||||
|
||||
@@ -8,11 +8,11 @@ import numpy as np
|
||||
import pytest
|
||||
|
||||
from vllm.config import ModelConfig
|
||||
from vllm.inputs import InputProcessingContext
|
||||
from vllm.multimodal import MULTIMODAL_REGISTRY
|
||||
# yapf conflicts with isort for this block
|
||||
# yapf: disable
|
||||
from vllm.multimodal.processing import (PlaceholderFeaturesInfo,
|
||||
from vllm.multimodal.processing import (InputProcessingContext,
|
||||
PlaceholderFeaturesInfo,
|
||||
PromptIndexTargets, PromptInsertion,
|
||||
PromptReplacement, apply_text_matches,
|
||||
apply_token_matches,
|
||||
|
||||
Reference in New Issue
Block a user