phoenix/litellm
1
0
Fork 0
forked from phoenix/litellm-mirror
Code Pull requests Activity

bump version

Browse source
This commit is contained in:
ishaan-jaff 2023-09-15 14:14:47 -07:00
parent 9e808c0c1c
commit 5ec1fc5048
3 changed files with 27 additions and 3 deletions
Download patch file Download diff file Expand all files Collapse all files

15
litellm/main.py
View file

@ -659,6 +659,14 @@ def completion(
chat_model = CodeChatModel.from_pretrained(model) chat_model = CodeChatModel.from_pretrained(model)
chat = chat_model.start_chat() chat = chat_model.start_chat()
if stream:
model_response = chat.send_message_streaming(prompt, **optional_params)
response = CustomStreamWrapper(
model_response, model, custom_llm_provider="vertexai", logging_obj=logging
)
return response
completion_response = chat.send_message(prompt, **optional_params) completion_response = chat.send_message(prompt, **optional_params)
## LOGGING ## LOGGING
@ -692,6 +700,13 @@ def completion(
else: else:
vertex_model = CodeGenerationModel.from_pretrained(model) vertex_model = CodeGenerationModel.from_pretrained(model)
if stream:
model_response = vertex_model.predict_streaming(prompt, **optional_params)
response = CustomStreamWrapper(
model_response, model, custom_llm_provider="vertexai", logging_obj=logging
)
return response
completion_response = vertex_model.predict(prompt, **optional_params) completion_response = vertex_model.predict(prompt, **optional_params)
## LOGGING ## LOGGING

13
litellm/utils.py
View file

@ -885,8 +885,8 @@ def get_optional_params( # use the openai defaults
if stop != None: if stop != None:
optional_params["stop"] = stop #TG AI expects a list, example ["\n\n\n\n","<|endoftext|>"] optional_params["stop"] = stop #TG AI expects a list, example ["\n\n\n\n","<|endoftext|>"]
elif ( elif (
model == "chat-bison" model in litellm.vertex_chat_models or model in litellm.vertex_code_chat_models
): # chat-bison has diff args from chat-bison@001 ty Google ): # chat-bison has diff args from chat-bison@001, ty Google :)
if temperature != 1: if temperature != 1:
optional_params["temperature"] = temperature optional_params["temperature"] = temperature
if top_p != 1: if top_p != 1:
@ -900,6 +900,12 @@ def get_optional_params( # use the openai defaults
optional_params["temperature"] = temperature optional_params["temperature"] = temperature
optional_params["top_p"] = top_p optional_params["top_p"] = top_p
optional_params["top_k"] = top_k optional_params["top_k"] = top_k
if max_tokens != float("inf"):
optional_params["max_output_tokens"] = max_tokens
elif model in model in litellm.vertex_code_text_models:
optional_params["temperature"] = temperature
if max_tokens != float("inf"):
optional_params["max_output_tokens"] = max_tokens
elif custom_llm_provider == "baseten": elif custom_llm_provider == "baseten":
optional_params["temperature"] = temperature optional_params["temperature"] = temperature
optional_params["stream"] = stream optional_params["stream"] = stream
@ -2482,6 +2488,9 @@ class CustomStreamWrapper:
elif self.model in litellm.nlp_cloud_models or self.custom_llm_provider == "nlp_cloud": elif self.model in litellm.nlp_cloud_models or self.custom_llm_provider == "nlp_cloud":
chunk = next(self.completion_stream) chunk = next(self.completion_stream)
completion_obj["content"] = self.handle_nlp_cloud_chunk(chunk) completion_obj["content"] = self.handle_nlp_cloud_chunk(chunk)
elif self.model in (litellm.vertex_chat_models + litellm.vertex_code_chat_models + litellm.vertex_text_models + litellm.vertex_code_text_models):
chunk = next(self.completion_stream)
completion_obj["content"] = str(chunk)
elif self.model in litellm.cohere_models or self.custom_llm_provider == "cohere": elif self.model in litellm.cohere_models or self.custom_llm_provider == "cohere":
chunk = next(self.completion_stream) chunk = next(self.completion_stream)
completion_obj["content"] = self.handle_cohere_chunk(chunk) completion_obj["content"] = self.handle_cohere_chunk(chunk)

2
pyproject.toml
View file

@ -1,6 +1,6 @@
[tool.poetry] [tool.poetry]
name = "litellm" name = "litellm"
version = "0.1.639" version = "0.1.640"
description = "Library to easily interface with LLM API providers" description = "Library to easily interface with LLM API providers"
authors = ["BerriAI"] authors = ["BerriAI"]
license = "MIT License" license = "MIT License"