scriptoriumadmin/litellm
1
0
Fork 0
mirror of https://github.com/BerriAI/litellm.git synced 2026-09-10 22:41:41 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 83

fix(vertex-ai): classify Lyria model metadata

Browse source
This commit is contained in:
Emerson Gomes 2026-08-12 12:56:06 -05:00
parent e9e2fbb438
commit bd5123564c
No known key found for this signature in database
GPG key ID: D3DF28AB5D1B5E17
4 changed files with 56 additions and 3 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

21
ci_cd/generate_model_prices_schema.py
View file

@ -58,6 +58,11 @@ OBJECT_KEYS: dict[str, JsonSchema] = {
}
ARRAY_KEYS: dict[str, JsonSchema] = {
"supported_audio_formats": {
"type": "array",
"description": "Audio container formats the model can return.",
"items": {"type": "string", "enum": ["mp3", "wav"]},
},
"supported_endpoints": {
"type": "array",
"description": "OpenAI-style API routes this model can be called through, e.g. /v1/chat/completions.",
@ -116,6 +121,10 @@ ARRAY_KEYS: dict[str, JsonSchema] = {
}
INTEGER_KEYS: dict[str, JsonSchema] = {
"max_audio_per_prompt": {
**NONNEG_INTEGER,
"description": "Maximum number of audio outputs accepted or generated per prompt.",
},
"max_tokens": {
**NONNEG_INTEGER,
"description": "Legacy field: max output tokens if the provider specifies it, else max input tokens.",
@ -141,6 +150,14 @@ INTEGER_KEYS: dict[str, JsonSchema] = {
}
NUMBER_KEYS: dict[str, JsonSchema] = {
"audio_seconds_per_prediction": {
**NONNEG_NUMBER,
"description": "Audio duration, in seconds, produced by one prediction.",
},
"max_audio_length_hours": {
**NONNEG_NUMBER,
"description": "Maximum generated audio duration, expressed in hours.",
},
"regional_processing_uplift_multiplier_eu": {
"type": "number",
"minimum": 1,
@ -231,6 +248,10 @@ def string_key_schemas(modes: tuple) -> dict[str, JsonSchema]:
},
"comment": STRING,
"audio_transcription_config": STRING,
"vertex_ai_audio_api": {
"type": "string",
"enum": ["lyria_predict", "lyria_interactions"],
},
}

2
litellm/llms/vertex_ai/text_to_speech/transformation.py
View file

@ -500,7 +500,7 @@ class VertexAILyriaTextToSpeechConfig(VertexAITextToSpeechConfig):
optional_params: dict,
voice: str | dict | None = None,
drop_params: bool = False,
kwargs: dict = {},
kwargs: dict | None = None,
) -> tuple[str | None, dict]:
mapped_params = dict(optional_params)
base_model = model.removeprefix("vertex_ai/")

3
litellm/proxy/pass_through_endpoints/llm_provider_handlers/vertex_passthrough_logging_handler.py
View file

@ -343,8 +343,7 @@ class VertexPassthroughLoggingHandler:
json_response=json_response
)
response_cost: Final = (
VertexPassthroughLoggingHandler._get_audio_prediction_unit_cost(model=model)
or 0.0
VertexPassthroughLoggingHandler._get_audio_prediction_unit_cost(model=model) or 0.0
) * prediction_count
logging_obj.model = model

33
model_prices_and_context_window.schema.json
View file

@ -53,6 +53,11 @@
"type": "number",
"minimum": 0
},
"audio_seconds_per_prediction": {
"type": "number",
"minimum": 0,
"description": "Audio duration, in seconds, produced by one prediction."
},
"audio_transcription_config": {
"type": "string"
},
@ -363,6 +368,16 @@
"type": "string",
"description": "LiteLLM provider slug; one of https://docs.litellm.ai/docs/providers."
},
"max_audio_length_hours": {
"type": "number",
"minimum": 0,
"description": "Maximum generated audio duration, expressed in hours."
},
"max_audio_per_prompt": {
"type": "integer",
"minimum": 0,
"description": "Maximum number of audio outputs accepted or generated per prompt."
},
"max_input_tokens": {
"type": "integer",
"minimum": 0,
@ -603,6 +618,17 @@
"type": "string",
"description": "URL of the provider pricing/model page this entry was taken from."
},
"supported_audio_formats": {
"type": "array",
"description": "Audio container formats the model can return.",
"items": {
"type": "string",
"enum": [
"mp3",
"wav"
]
}
},
"supported_endpoints": {
"type": "array",
"description": "OpenAI-style API routes this model can be called through, e.g. /v1/chat/completions.",
@ -826,6 +852,13 @@
"uses_embed_content": {
"type": "boolean"
},
"vertex_ai_audio_api": {
"type": "string",
"enum": [
"lyria_predict",
"lyria_interactions"
]
},
"web_search_billing_unit": {
"type": "string",
"description": "Whether web search is billed per query or per prompt.",