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

[Core] Logprobs support in Multi-step (#7652)

Browse Source
This commit is contained in:
afeldman-nm
2024-08-29 22:19:08 -04:00
committed by GitHub
parent 4abed65c58
commit 428dd1445e
103 changed files with 872 additions and 376 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
vllm/model_executor/models/ultravox.py
View File

@@ -27,6 +27,7 @@ from vllm.model_executor.layers.activation import SiluAndMul, get_act_fn
from vllm.model_executor.layers.layernorm import RMSNorm
from vllm.model_executor.layers.quantization.base_config import (
QuantizationConfig)
from vllm.model_executor.layers.sampler import SamplerOutput
from vllm.model_executor.model_loader.weight_utils import default_weight_loader
from vllm.model_executor.models.interfaces import SupportsMultiModal
from vllm.model_executor.models.utils import (filter_weights,
@@ -37,7 +38,7 @@ from vllm.multimodal import MULTIMODAL_REGISTRY
from vllm.multimodal.base import MultiModalInputs
from vllm.multimodal.utils import (cached_get_tokenizer,
repeat_and_pad_placeholder_tokens)
from vllm.sequence import VLLM_TOKEN_ID_ARRAY_TYPE, SamplerOutput, SequenceData
from vllm.sequence import VLLM_TOKEN_ID_ARRAY_TYPE, SequenceData
from vllm.transformers_utils.configs.ultravox import UltravoxConfig
_AUDIO_PLACEHOLDER_TOKEN = 128002