together 1.2.10__tar.gz → 1.2.12__tar.gz

{together-1.2.10 → together-1.2.12}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: together
-Version: 1.2.10
+Version: 1.2.12
 Summary: Python client for Together's Cloud Platform!
 Home-page: https://github.com/togethercomputer/together-python
 License: Apache-2.0

{together-1.2.10 → together-1.2.12}/pyproject.toml

@@ -12,7 +12,7 @@ build-backend = "poetry.masonry.api"
 
 [tool.poetry]
 name = "together"
-version = "1.2.10"
+version = "1.2.12"
 authors = [
     "Together AI <support@together.ai>"
 ]
@@ -55,7 +55,7 @@ types-tqdm = "^4.65.0.0"
 types-tabulate = "^0.9.0.3"
 pre-commit = "3.5.0"
 types-requests = "^2.31.0.20240218"
-pyarrow-stubs = "^10.0.1.7"
+pyarrow-stubs = ">=10.0.1.7,<20240831.0.0.0"
 mypy = "^1.9.0"
 
 [tool.poetry.group.tests]

{together-1.2.10 → together-1.2.12}/src/together/legacy/complete.py

@@ -14,7 +14,7 @@ class Complete:
     def create(
         cls,
         prompt: str,
-        **kwargs,
+        **kwargs: Any,
     ) -> Dict[str, Any]:
         """Legacy completion function."""
 
@@ -36,7 +36,7 @@ class Complete:
     def create_streaming(
         cls,
         prompt: str,
-        **kwargs,
+        **kwargs: Any,
     ) -> Iterator[Dict[str, Any]]:
         """Legacy streaming completion function."""
 
@@ -59,7 +59,7 @@ class Completion:
     def create(
         cls,
         prompt: str,
-        **kwargs,
+        **kwargs: Any,
     ) -> CompletionResponse | Iterator[CompletionChunk]:
         """Completion function."""
 
@@ -79,7 +79,7 @@ class AsyncComplete:
     async def create(
         cls,
         prompt: str,
-        **kwargs,
+        **kwargs: Any,
     ) -> CompletionResponse | AsyncGenerator[CompletionChunk, None]:
         """Async completion function."""
 

{together-1.2.10 → together-1.2.12}/src/together/legacy/embeddings.py

@@ -11,7 +11,7 @@ class Embeddings:
     def create(
         cls,
         input: str,
-        **kwargs,
+        **kwargs: Any,
     ) -> Dict[str, Any]:
         """Legacy embeddings function."""
 

{together-1.2.10 → together-1.2.12}/src/together/legacy/images.py

@@ -11,7 +11,7 @@ class Image:
     def create(
         cls,
         prompt: str,
-        **kwargs,
+        **kwargs: Any,
     ) -> Dict[str, Any]:
         """Legacy image function."""
 

{together-1.2.10 → together-1.2.12}/src/together/resources/chat/completions.py

@@ -32,6 +32,7 @@ class ChatCompletions:
         frequency_penalty: float | None = None,
         min_p: float | None = None,
         logit_bias: Dict[str, float] | None = None,
+        seed: int | None = None,
         stream: bool = False,
         logprobs: int | None = None,
         echo: bool | None = None,
@@ -40,6 +41,7 @@ class ChatCompletions:
         response_format: Dict[str, str | Dict[str, Any]] | None = None,
         tools: Dict[str, str | Dict[str, Any]] | None = None,
         tool_choice: str | Dict[str, str | Dict[str, str]] | None = None,
+        **kwargs: Any,
     ) -> ChatCompletionResponse | Iterator[ChatCompletionChunk]:
         """
         Method to generate completions based on a given prompt using a specified model.
@@ -78,6 +80,7 @@ class ChatCompletions:
             logit_bias (Dict[str, float], optional): A dictionary of tokens and their bias values that modify the
                 likelihood of specific tokens being sampled. Bias values must be in the range [-100, 100].
                 Defaults to None.
+            seed (int, optional): A seed value to use for reproducibility.
             stream (bool, optional): Flag indicating whether to stream the generated completions.
                 Defaults to False.
             logprobs (int, optional): Number of top-k logprobs to return
@@ -123,6 +126,7 @@ class ChatCompletions:
             frequency_penalty=frequency_penalty,
             min_p=min_p,
             logit_bias=logit_bias,
+            seed=seed,
             stream=stream,
             logprobs=logprobs,
             echo=echo,
@@ -131,6 +135,7 @@ class ChatCompletions:
             response_format=response_format,
             tools=tools,
             tool_choice=tool_choice,
+            **kwargs,
         ).model_dump(exclude_none=True)
 
         response, _, _ = requestor.request(
@@ -169,6 +174,7 @@ class AsyncChatCompletions:
         frequency_penalty: float | None = None,
         min_p: float | None = None,
         logit_bias: Dict[str, float] | None = None,
+        seed: int | None = None,
         stream: bool = False,
         logprobs: int | None = None,
         echo: bool | None = None,
@@ -177,6 +183,7 @@ class AsyncChatCompletions:
         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,
+        **kwargs: Any,
     ) -> AsyncGenerator[ChatCompletionChunk, None] | ChatCompletionResponse:
         """
         Async method to generate completions based on a given prompt using a specified model.
@@ -215,6 +222,7 @@ class AsyncChatCompletions:
             logit_bias (Dict[str, float], optional): A dictionary of tokens and their bias values that modify the
                 likelihood of specific tokens being sampled. Bias values must be in the range [-100, 100].
                 Defaults to None.
+            seed (int, optional): A seed value to use for reproducibility.
             stream (bool, optional): Flag indicating whether to stream the generated completions.
                 Defaults to False.
             logprobs (int, optional): Number of top-k logprobs to return
@@ -260,6 +268,7 @@ class AsyncChatCompletions:
             frequency_penalty=frequency_penalty,
             min_p=min_p,
             logit_bias=logit_bias,
+            seed=seed,
             stream=stream,
             logprobs=logprobs,
             echo=echo,
@@ -268,6 +277,7 @@ class AsyncChatCompletions:
             response_format=response_format,
             tools=tools,
             tool_choice=tool_choice,
+            **kwargs,
         ).model_dump(exclude_none=True)
 
         response, _, _ = await requestor.arequest(

{together-1.2.10 → together-1.2.12}/src/together/resources/completions.py

@@ -1,6 +1,6 @@
 from __future__ import annotations
 
-from typing import AsyncGenerator, Dict, Iterator, List
+from typing import AsyncGenerator, Dict, Iterator, List, Any
 
 from together.abstract import api_requestor
 from together.together_response import TogetherResponse
@@ -32,11 +32,13 @@ class Completions:
         frequency_penalty: float | None = None,
         min_p: float | None = None,
         logit_bias: Dict[str, float] | None = None,
+        seed: int | None = None,
         stream: bool = False,
         logprobs: int | None = None,
         echo: bool | None = None,
         n: int | None = None,
         safety_model: str | None = None,
+        **kwargs: Any,
     ) -> CompletionResponse | Iterator[CompletionChunk]:
         """
         Method to generate completions based on a given prompt using a specified model.
@@ -74,6 +76,7 @@ class Completions:
             logit_bias (Dict[str, float], optional): A dictionary of tokens and their bias values that modify the
                 likelihood of specific tokens being sampled. Bias values must be in the range [-100, 100].
                 Defaults to None.
+            seed (int, optional): Seed value for reproducibility.
             stream (bool, optional): Flag indicating whether to stream the generated completions.
                 Defaults to False.
             logprobs (int, optional): Number of top-k logprobs to return
@@ -106,6 +109,7 @@ class Completions:
             repetition_penalty=repetition_penalty,
             presence_penalty=presence_penalty,
             frequency_penalty=frequency_penalty,
+            seed=seed,
             min_p=min_p,
             logit_bias=logit_bias,
             stream=stream,
@@ -113,6 +117,7 @@ class Completions:
             echo=echo,
             n=n,
             safety_model=safety_model,
+            **kwargs,
         ).model_dump(exclude_none=True)
 
         response, _, _ = requestor.request(
@@ -151,11 +156,13 @@ class AsyncCompletions:
         frequency_penalty: float | None = None,
         min_p: float | None = None,
         logit_bias: Dict[str, float] | None = None,
+        seed: int | None = None,
         stream: bool = False,
         logprobs: int | None = None,
         echo: bool | None = None,
         n: int | None = None,
         safety_model: str | None = None,
+        **kwargs: Any,
     ) -> AsyncGenerator[CompletionChunk, None] | CompletionResponse:
         """
         Async method to generate completions based on a given prompt using a specified model.
@@ -193,6 +200,7 @@ class AsyncCompletions:
             logit_bias (Dict[str, float], optional): A dictionary of tokens and their bias values that modify the
                 likelihood of specific tokens being sampled. Bias values must be in the range [-100, 100].
                 Defaults to None.
+            seed (int, optional): Seed value for reproducibility.
             stream (bool, optional): Flag indicating whether to stream the generated completions.
                 Defaults to False.
             logprobs (int, optional): Number of top-k logprobs to return
@@ -227,11 +235,13 @@ class AsyncCompletions:
             frequency_penalty=frequency_penalty,
             min_p=min_p,
             logit_bias=logit_bias,
+            seed=seed,
             stream=stream,
             logprobs=logprobs,
             echo=echo,
             n=n,
             safety_model=safety_model,
+            **kwargs,
         ).model_dump(exclude_none=True)
 
         response, _, _ = await requestor.arequest(

{together-1.2.10 → together-1.2.12}/src/together/resources/embeddings.py

@@ -1,6 +1,6 @@
 from __future__ import annotations
 
-from typing import List
+from typing import List, Any
 
 from together.abstract import api_requestor
 from together.together_response import TogetherResponse
@@ -21,6 +21,7 @@ class Embeddings:
         *,
         input: str | List[str],
         model: str,
+        **kwargs: Any,
     ) -> EmbeddingResponse:
         """
         Method to generate completions based on a given prompt using a specified model.
@@ -40,6 +41,7 @@ class Embeddings:
         parameter_payload = EmbeddingRequest(
             input=input,
             model=model,
+            **kwargs,
         ).model_dump(exclude_none=True)
 
         response, _, _ = requestor.request(
@@ -65,6 +67,7 @@ class AsyncEmbeddings:
         *,
         input: str | List[str],
         model: str,
+        **kwargs: Any,
     ) -> EmbeddingResponse:
         """
         Async method to generate completions based on a given prompt using a specified model.
@@ -84,6 +87,7 @@ class AsyncEmbeddings:
         parameter_payload = EmbeddingRequest(
             input=input,
             model=model,
+            **kwargs,
         ).model_dump(exclude_none=True)
 
         response, _, _ = await requestor.arequest(

{together-1.2.10 → together-1.2.12}/src/together/resources/images.py

@@ -1,5 +1,7 @@
 from __future__ import annotations
 
+from typing import Any
+
 from together.abstract import api_requestor
 from together.together_response import TogetherResponse
 from together.types import (
@@ -25,6 +27,7 @@ class Images:
         height: int | None = 1024,
         width: int | None = 1024,
         negative_prompt: str | None = None,
+        **kwargs: Any,
     ) -> ImageResponse:
         """
         Method to generate images based on a given prompt using a specified model.
@@ -67,6 +70,7 @@ class Images:
             height=height,
             width=width,
             negative_prompt=negative_prompt,
+            **kwargs,
         ).model_dump(exclude_none=True)
 
         response, _, _ = requestor.request(
@@ -98,6 +102,7 @@ class AsyncImages:
         height: int | None = 1024,
         width: int | None = 1024,
         negative_prompt: str | None = None,
+        **kwargs: Any,
     ) -> ImageResponse:
         """
         Async method to generate images based on a given prompt using a specified model.
@@ -140,6 +145,7 @@ class AsyncImages:
             height=height,
             width=width,
             negative_prompt=negative_prompt,
+            **kwargs,
         ).model_dump(exclude_none=True)
 
         response, _, _ = await requestor.arequest(

{together-1.2.10 → together-1.2.12}/src/together/resources/rerank.py

@@ -25,6 +25,7 @@ class Rerank:
         top_n: int | None = None,
         return_documents: bool = False,
         rank_fields: List[str] | None = None,
+        **kwargs: Any,
     ) -> RerankResponse:
         """
         Method to generate completions based on a given prompt using a specified model.
@@ -52,6 +53,7 @@ class Rerank:
             top_n=top_n,
             return_documents=return_documents,
             rank_fields=rank_fields,
+            **kwargs,
         ).model_dump(exclude_none=True)
 
         response, _, _ = requestor.request(
@@ -81,6 +83,7 @@ class AsyncRerank:
         top_n: int | None = None,
         return_documents: bool = False,
         rank_fields: List[str] | None = None,
+        **kwargs: Any,
     ) -> RerankResponse:
         """
         Async method to generate completions based on a given prompt using a specified model.
@@ -108,6 +111,7 @@ class AsyncRerank:
             top_n=top_n,
             return_documents=return_documents,
             rank_fields=rank_fields,
+            **kwargs,
         ).model_dump(exclude_none=True)
 
         response, _, _ = await requestor.arequest(

{together-1.2.10 → together-1.2.12}/src/together/types/chat_completions.py

@@ -96,6 +96,7 @@ class ChatCompletionRequest(BaseModel):
     frequency_penalty: float | None = None
     min_p: float | None = None
     logit_bias: Dict[str, float] | None = None
+    seed: int | None = None
     # stream SSE token chunks
     stream: bool = False
     # return logprobs
@@ -126,6 +127,7 @@ class ChatCompletionRequest(BaseModel):
 class ChatCompletionChoicesData(BaseModel):
     index: int | None = None
     logprobs: LogprobsPart | None = None
+    seed: int | None = None
     finish_reason: FinishReason | None = None
     message: ChatCompletionMessage | None = None
 
@@ -150,6 +152,7 @@ class ChatCompletionResponse(BaseModel):
 class ChatCompletionChoicesChunk(BaseModel):
     index: int | None = None
     logprobs: float | None = None
+    seed: int | None = None
     finish_reason: FinishReason | None = None
     delta: DeltaContent | None = None
 

{together-1.2.10 → together-1.2.12}/src/together/types/completions.py

@@ -35,6 +35,7 @@ class CompletionRequest(BaseModel):
     frequency_penalty: float | None = None
     min_p: float | None = None
     logit_bias: Dict[str, float] | None = None
+    seed: int | None = None
     # stream SSE token chunks
     stream: bool = False
     # return logprobs
@@ -61,13 +62,15 @@ class CompletionRequest(BaseModel):
 class CompletionChoicesData(BaseModel):
     index: int
     logprobs: LogprobsPart | None = None
-    finish_reason: FinishReason | None = None
+    seed: int | None = None
+    finish_reason: FinishReason
     text: str
 
 
 class CompletionChoicesChunk(BaseModel):
     index: int | None = None
     logprobs: float | None = None
+    seed: int | None = None
     finish_reason: FinishReason | None = None
     delta: DeltaContent | None = None