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

docs(proxy_server): cleanup docs

Browse source
This commit is contained in:
Krrish Dholakia 2023-10-16 15:38:50 -07:00
parent d8aa7d1aaf
commit 1df487a45d
1 changed files with 1 additions and 40 deletions
Download patch file Download diff file Expand all files Collapse all files

41
docs/my-website/docs/proxy_server.md
View file

@ -520,46 +520,7 @@ $ litellm --config -f ./litellm_config.toml
LiteLLM will save a copy of this file in it's package, so it can persist these settings across restarts.
**Complete Config File**
```shell
### API KEYS ###
[keys]
# HUGGINGFACE_API_KEY="" # Uncomment to save your Hugging Face API key
# OPENAI_API_KEY="" # Uncomment to save your OpenAI API Key
# TOGETHERAI_API_KEY="" # Uncomment to save your TogetherAI API key
# NLP_CLOUD_API_KEY="" # Uncomment to save your NLP Cloud API key
# ANTHROPIC_API_KEY="" # Uncomment to save your Anthropic API key
# REPLICATE_API_KEY="" # Uncomment to save your Replicate API key
# AWS_ACCESS_KEY_ID = "" # Uncomment to save your Bedrock/Sagemaker access keys
# AWS_SECRET_ACCESS_KEY = "" # Uncomment to save your Bedrock/Sagemaker access keys
### LITELLM PARAMS ###
[general]
# add_function_to_prompt = True # e.g: Ollama doesn't support functions, so add it to the prompt instead
# drop_params = True # drop any params not supported by the provider (e.g. Ollama)
# default_model = "gpt-4" # route all requests to this model
# fallbacks = ["gpt-3.5-turbo", "gpt-3.5-turbo-16k"] # models you want to fallback to in case completion call fails (remember: add relevant keys)
### MODEL PARAMS ###
[model."ollama/llama2"] # run via `litellm --model ollama/llama2`
# max_tokens = "" # set max tokens for the model
# temperature = "" # set temperature for the model
# api_base = "" # set a custom api base for the model
[model."ollama/llama2".prompt_template] # [OPTIONAL] LiteLLM can automatically formats the prompt - docs: https://docs.litellm.ai/docs/completion/prompt_formatting
# MODEL_SYSTEM_MESSAGE_START_TOKEN = "[INST] <<SYS>>\n" # This does not need to be a token, can be any string
# MODEL_SYSTEM_MESSAGE_END_TOKEN = "\n<</SYS>>\n [/INST]\n" # This does not need to be a token, can be any string
# MODEL_USER_MESSAGE_START_TOKEN = "[INST] " # This does not need to be a token, can be any string
# MODEL_USER_MESSAGE_END_TOKEN = " [/INST]\n" # Applies only to user messages. Can be any string.
# MODEL_ASSISTANT_MESSAGE_START_TOKEN = "" # Applies only to assistant messages. Can be any string.
# MODEL_ASSISTANT_MESSAGE_END_TOKEN = "\n" # Applies only to system messages. Can be any string.
# MODEL_PRE_PROMPT = "You are a good bot" # Applied at the start of the prompt
# MODEL_POST_PROMPT = "Now answer as best as you can" # Applied at the end of the prompt
```
[**Complete Config File**](https://github.com/BerriAI/litellm/blob/main/secrets_template.toml)
[**🔥 [Tutorial] modify a model prompt on the proxy**](./tutorials/model_config_proxy.md)