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

(feat) proxy - use acces groups

Browse source
This commit is contained in:
ishaan-jaff 2024-02-14 17:23:25 -08:00
parent d4ffc98a39
commit f2d6aab8f5
1 changed files with 5 additions and 10 deletions
Download patch file Download diff file Expand all files Collapse all files

15
litellm/proxy/proxy_config.yaml
View file

@ -9,14 +9,19 @@ model_list:
mode: chat mode: chat
max_tokens: 4096 max_tokens: 4096
base_model: azure/gpt-4-1106-preview base_model: azure/gpt-4-1106-preview
access_groups: ["public"]
- model_name: openai-gpt-3.5 - model_name: openai-gpt-3.5
litellm_params: litellm_params:
model: gpt-3.5-turbo model: gpt-3.5-turbo
api_key: os.environ/OPENAI_API_KEY api_key: os.environ/OPENAI_API_KEY
model_info:
access_groups: ["public"]
- model_name: anthropic-claude-v2.1 - model_name: anthropic-claude-v2.1
litellm_params: litellm_params:
model: bedrock/anthropic.claude-v2:1 model: bedrock/anthropic.claude-v2:1
timeout: 300 # sets a 5 minute timeout timeout: 300 # sets a 5 minute timeout
model_info:
access_groups: ["private"]
- model_name: anthropic-claude-v2 - model_name: anthropic-claude-v2
litellm_params: litellm_params:
model: bedrock/anthropic.claude-v2 model: bedrock/anthropic.claude-v2
@ -39,16 +44,6 @@ model_list:
litellm_settings: litellm_settings:
fallbacks: [{"openai-gpt-3.5": ["azure-gpt-3.5"]}] fallbacks: [{"openai-gpt-3.5": ["azure-gpt-3.5"]}]
success_callback: ['langfuse'] success_callback: ['langfuse']
max_budget: 50 # global budget for proxy
max_user_budget: 0.0001
budget_duration: 30d # global budget duration, will reset after 30d
default_key_generate_params:
max_budget: 1.5000
models: ["azure-gpt-3.5"]
duration: None
upperbound_key_generate_params:
max_budget: 100
duration: "30d"
# setting callback class # setting callback class
# callbacks: custom_callbacks.proxy_handler_instance # sets litellm.callbacks = [proxy_handler_instance] # callbacks: custom_callbacks.proxy_handler_instance # sets litellm.callbacks = [proxy_handler_instance]