[Frontend] Added chat-style multimodal support to /classify. (#27516)
Signed-off-by: WorldExplored <srreyansh.sethi@gmail.com> Signed-off-by: Srreyansh Sethi <107075589+WorldExplored@users.noreply.github.com> Signed-off-by: vnadathur <glvikramn@gmail.com> Signed-off-by: wang.yuqi <yuqi.wang@daocloud.io> Co-authored-by: vnadathur <236933696+vnadathur@users.noreply.github.com> Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com> Co-authored-by: vnadathur <glvikramn@gmail.com> Co-authored-by: wang.yuqi <noooop@126.com> Co-authored-by: wang.yuqi <yuqi.wang@daocloud.io>
This commit is contained in:
co-authored by
vnadathur
gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
vnadathur
wang.yuqi
wang.yuqi
parent
ecf8230d4d
commit
360bd8762f
@@ -43,6 +43,8 @@ from vllm.entrypoints.openai.protocol import (
|
||||
ChatCompletionNamedToolChoiceParam,
|
||||
ChatCompletionRequest,
|
||||
ChatCompletionResponse,
|
||||
ClassificationChatRequest,
|
||||
ClassificationCompletionRequest,
|
||||
ClassificationRequest,
|
||||
ClassificationResponse,
|
||||
CompletionRequest,
|
||||
@@ -114,13 +116,16 @@ CompletionLikeRequest: TypeAlias = (
|
||||
| DetokenizeRequest
|
||||
| EmbeddingCompletionRequest
|
||||
| RerankRequest
|
||||
| ClassificationRequest
|
||||
| ClassificationCompletionRequest
|
||||
| ScoreRequest
|
||||
| TokenizeCompletionRequest
|
||||
)
|
||||
|
||||
ChatLikeRequest: TypeAlias = (
|
||||
ChatCompletionRequest | EmbeddingChatRequest | TokenizeChatRequest
|
||||
ChatCompletionRequest
|
||||
| EmbeddingChatRequest
|
||||
| TokenizeChatRequest
|
||||
| ClassificationChatRequest
|
||||
)
|
||||
SpeechToTextRequest: TypeAlias = TranscriptionRequest | TranslationRequest
|
||||
AnyRequest: TypeAlias = (
|
||||
@@ -814,7 +819,11 @@ class OpenAIServing:
|
||||
if not hasattr(request, "messages"):
|
||||
return message_types
|
||||
|
||||
for message in request.messages:
|
||||
messages = request.messages
|
||||
if messages is None or isinstance(messages, (str, bytes)):
|
||||
return message_types
|
||||
|
||||
for message in messages:
|
||||
if (
|
||||
isinstance(message, dict)
|
||||
and "content" in message
|
||||
@@ -907,7 +916,8 @@ class OpenAIServing:
|
||||
EmbeddingCompletionRequest,
|
||||
ScoreRequest,
|
||||
RerankRequest,
|
||||
ClassificationRequest,
|
||||
ClassificationCompletionRequest,
|
||||
ClassificationChatRequest,
|
||||
),
|
||||
):
|
||||
# Note: input length can be up to the entire model context length
|
||||
@@ -915,7 +925,8 @@ class OpenAIServing:
|
||||
if token_num > self.max_model_len:
|
||||
operations: dict[type[AnyRequest], str] = {
|
||||
ScoreRequest: "score",
|
||||
ClassificationRequest: "classification",
|
||||
ClassificationCompletionRequest: "classification",
|
||||
ClassificationChatRequest: "classification",
|
||||
}
|
||||
operation = operations.get(type(request), "embedding generation")
|
||||
raise ValueError(
|
||||
|
||||
Reference in New Issue
Block a user