@tduncan / Nvidia-Optimization-Stack-VLLM Mirror · read-only / commits / 44eaa5a5d9
Fork 0

A high-throughput and memory-efficient inference and serving engine for LLMs

Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git

Synced Nov 16, 10:45 UTC

[Frontend] Clarify model_type error messages (#9345)

Steve Grubb committed Oct 15, 2024 at 00:29 UTC 44eaa5a5d966d41c5b19b38d60c41bec02399525
1 file changed +3 -3
vllm/entrypoints/chat_utils.py
+3 -3
@@ -166,15 +166,15 @@ class BaseMultiModalItemTracker(ABC, Generic[_T]):
166 if model_type == "molmo":
167 return ""
168
169 - raise TypeError(f"Unknown model type: {model_type}")
169 + raise TypeError(f"Unknown {modality} model type: {model_type}")
170 elif modality == "audio":
171 if model_type == "ultravox":
172 return "<|reserved_special_token_0|>"
173 - raise TypeError(f"Unknown model type: {model_type}")
173 + raise TypeError(f"Unknown {modality} model type: {model_type}")
174 elif modality == "video":
175 if model_type == "qwen2_vl":
176 return "<|vision_start|><|video_pad|><|vision_end|>"
177 - raise TypeError(f"Unknown model type: {model_type}")
177 + raise TypeError(f"Unknown {modality} model type: {model_type}")
178 else:
179 raise TypeError(f"Unknown modality: {modality}")
180