Skip to content

add turboboost_tps param #230

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
zainhas wants to merge 1 commit into from
Closed

add turboboost_tps param #230

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
17 changes: 17 additions & 0 deletions src/together/resources/chat/completions.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -41,6 +41,7 @@ def create(
response_format: Dict[str, str | Dict[str, Any]] | None = None,
tools: List[Dict[str, Any]] | None = None,
tool_choice: str | Dict[str, str | Dict[str, str]] | None = None,
turboboost_tps: float | None = None,
**kwargs: Any,
) -> ChatCompletionResponse | Iterator[ChatCompletionChunk]:
"""
Expand Down Expand Up @@ -103,6 +104,13 @@ def create(
via {"type": "function", "function": {"name": "my_function"}} forces the model to call that function.
Sets to `auto` if None.
Defaults to None.
turboboost_tps (float, optional): A parameter that controls the speed-quality tradeoff between
the draft model (faster, lower quality) and target model (slower, higher quality).
Values range from 0.0 (regular speculative decoding with target model quality) to
1.0 (draft model speed with 100% acceptance rate). Higher values increase speed
while potentially reducing quality.
Defaults to 0.0.


Returns:
ChatCompletionResponse | Iterator[ChatCompletionChunk]: Object containing the completions
Expand Down Expand Up @@ -135,6 +143,7 @@ def create(
response_format=response_format,
tools=tools,
tool_choice=tool_choice,
turboboost_tps=turboboost_tps,
**kwargs,
).model_dump(exclude_none=True)

Expand Down Expand Up @@ -183,6 +192,7 @@ async def create(
response_format: Dict[str, Any] | None = None,
tools: Dict[str, str | Dict[str, str | Dict[str, Any]]] | None = None,
tool_choice: str | Dict[str, str | Dict[str, str]] | None = None,
turboboost_tps: float | None = None,
**kwargs: Any,
) -> AsyncGenerator[ChatCompletionChunk, None] | ChatCompletionResponse:
"""
Expand Down Expand Up @@ -245,6 +255,12 @@ async def create(
via {"type": "function", "function": {"name": "my_function"}} forces the model to call that function.
Sets to `auto` if None.
Defaults to None.
turboboost_tps (float, optional): A parameter that controls the speed-quality tradeoff between
the draft model (faster, lower quality) and target model (slower, higher quality).
Values range from 0.0 (regular speculative decoding with target model quality) to
1.0 (draft model speed with 100% acceptance rate). Higher values increase speed
while potentially reducing quality.
Defaults to 0.0.

Returns:
AsyncGenerator[ChatCompletionChunk, None] | ChatCompletionResponse: Object containing the completions
Expand Down Expand Up @@ -277,6 +293,7 @@ async def create(
response_format=response_format,
tools=tools,
tool_choice=tool_choice,
turboboost_tps=turboboost_tps,
**kwargs,
).model_dump(exclude_none=True)

Expand Down
16 changes: 16 additions & 0 deletions src/together/resources/completions.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -38,6 +38,7 @@ def create(
echo: bool | None = None,
n: int | None = None,
safety_model: str | None = None,
turboboost_tps: float | None = None,
**kwargs: Any,
) -> CompletionResponse | Iterator[CompletionChunk]:
"""
Expand Down Expand Up @@ -88,6 +89,12 @@ def create(
safety_model (str, optional): A moderation model to validate tokens. Choice between available moderation
models found [here](https://docs.together.ai/docs/inference-models#moderation-models).
Defaults to None.
turboboost_tps (float, optional): A parameter that controls the speed-quality tradeoff between
the draft model (faster, lower quality) and target model (slower, higher quality).
Values range from 0.0 (regular speculative decoding with target model quality) to
1.0 (draft model speed with 100% acceptance rate). Higher values increase speed
while potentially reducing quality.
Defaults to 0.0.

Returns:
CompletionResponse | Iterator[CompletionChunk]: Object containing the completions
Expand Down Expand Up @@ -117,6 +124,7 @@ def create(
echo=echo,
n=n,
safety_model=safety_model,
turboboost_tps=turboboost_tps,
**kwargs,
).model_dump(exclude_none=True)

Expand Down Expand Up @@ -162,6 +170,7 @@ async def create(
echo: bool | None = None,
n: int | None = None,
safety_model: str | None = None,
turboboost_tps: float | None = None,
**kwargs: Any,
) -> AsyncGenerator[CompletionChunk, None] | CompletionResponse:
"""
Expand Down Expand Up @@ -212,6 +221,12 @@ async def create(
safety_model (str, optional): A moderation model to validate tokens. Choice between available moderation
models found [here](https://docs.together.ai/docs/inference-models#moderation-models).
Defaults to None.
turboboost_tps (float, optional): A parameter that controls the speed-quality tradeoff between
the draft model (faster, lower quality) and target model (slower, higher quality).
Values range from 0.0 (regular speculative decoding with target model quality) to
1.0 (draft model speed with 100% acceptance rate). Higher values increase speed
while potentially reducing quality.
Defaults to 0.0.

Returns:
AsyncGenerator[CompletionChunk, None] | CompletionResponse: Object containing the completions
Expand Down Expand Up @@ -241,6 +256,7 @@ async def create(
echo=echo,
n=n,
safety_model=safety_model,
turboboost_tps=turboboost_tps,
**kwargs,
).model_dump(exclude_none=True)

Expand Down
Loading