[Core] Remove tokenizer group in vLLM (#24078)

Signed-off-by: Zhuohan Li <zhuohan123@gmail.com>
This commit is contained in:
Zhuohan Li
2025-09-17 08:42:59 +00:00
committed by GitHub
parent c15309a730
commit 6c47f6bfa4
49 changed files with 276 additions and 934 deletions
+3 -4
View File
@@ -76,8 +76,7 @@ class EmbeddingMixin(OpenAIServing):
try:
ctx.lora_request = self._maybe_get_adapters(ctx.request)
tokenizer = await self.engine_client.get_tokenizer(ctx.lora_request
)
tokenizer = await self.engine_client.get_tokenizer()
renderer = self._get_renderer(tokenizer)
if isinstance(ctx.request, EmbeddingChatRequest):
@@ -394,8 +393,8 @@ class EmbeddingMixin(OpenAIServing):
) -> Optional[ErrorResponse]:
"""Collect and aggregate batch results
with support for chunked processing.
For chunked requests, performs online aggregation to
For chunked requests, performs online aggregation to
minimize memory usage.
For regular requests, collects results normally.
"""