From 8dba46becfbcc4669867db4487aa9b91c51e4afa Mon Sep 17 00:00:00 2001
From: sigoden <sigoden@gmail.com>
Date: Tue, 30 Apr 2024 12:52:58 +0800
Subject: feat: openai-compatible platforms share the same client (#469)

---
 config.example.yaml | 42 ++++++++++++++++++++++++------------------
 1 file changed, 24 insertions(+), 18 deletions(-)

(limited to 'config.example.yaml')

diff --git a/config.example.yaml b/config.example.yaml
index ebf59ae..85630ac 100644
--- a/config.example.yaml
+++ b/config.example.yaml
@@ -47,6 +47,16 @@ clients:
     api_base: https://api.openai.com/v1               # ENV: {client_name}_API_BASE
     organization_id: org-xxx                          # Optional
 
+  # For any platform compatible with OpenAI's API
+  - type: openai-compatible
+    name: localai
+    api_base: http://localhost:8080/v1                # ENV: {client_name}_API_BASE
+    api_key: xxx                                      # ENV: {client_name}_API_KEY
+    chat_endpoint: /chat/completions                  # Optional
+    models:
+      - name: llama3
+        max_input_tokens: 8192
+
   # See https://ai.google.dev/docs
   - type: gemini
     api_key: xxx                                      # ENV: {client_name}_API_KEY
@@ -58,7 +68,8 @@ clients:
     api_key: sk-ant-xxx                               # ENV: {client_name}_API_KEY
 
   # See https://docs.mistral.ai/
-  - type: mistral
+  - type: openai-compatible
+    name: mistral
     api_key: xxx                                      # ENV: {client_name}_API_KEY
 
   # See https://docs.cohere.com/docs/the-cohere-platform
@@ -129,20 +140,9 @@ clients:
   - type: moonshot
     api_key: sk-xxx                                  # ENV: {client_name}_API_KEY
 
-  # For any platform compatible with OpenAI's API
-  - type: openai-compatible
-    name: localai
-    api_base: http://localhost:8080/v1                # ENV: {client_name}_API_BASE
-    api_key: sk-xxx                                   # ENV: {client_name}_API_KEY
-    chat_endpoint: /chat/completions                  # Optional
-    models:                                           # Required
-      - name: llama3
-        max_input_tokens: 8192
-
   # See https://docs.endpoints.anyscale.com/
   - type: openai-compatible
     name: anyscale
-    api_base: https://api.endpoints.anyscale.com/v1
     api_key: xxx
     models:
       # https://docs.endpoints.anyscale.com/text-generation/query-a-model#select-a-model
@@ -154,7 +154,6 @@ clients:
   # See https://deepinfra.com/docs
   - type: openai-compatible
     name: deepinfra
-    api_base: https://api.deepinfra.com/v1/openai
     api_key: xxx
     models:
       # https://deepinfra.com/models
@@ -166,7 +165,6 @@ clients:
   # See https://readme.fireworks.ai/docs/quickstart
   - type: openai-compatible
     name: fireworks
-    api_base: https://api.fireworks.ai/inference/v1
     api_key: xxx
     models:
       # https://fireworks.ai/models
@@ -175,12 +173,21 @@ clients:
         input_price: 0.9
         output_price: 0.9
 
+  # See https://openrouter.ai/docs#quick-start
+  - type: openai-compatible
+    name: openrouter
+    api_key: xxx                                      # ENV: {client_name}_API_KEY
+    models:
+      # https://openrouter.ai/docs#models
+      - name: meta-llama/llama-3-70b-instruct
+        max_input_tokens: 8192
+        input_price: 0.81
+        output_price: 0.81
 
   # See https://octo.ai/docs/getting-started/quickstart
   - type: openai-compatible
     name: octoai
-    api_base: https://text.octoai.run/v1
-    api_key: xxx
+    api_key: xxx                                      # ENV: {client_name}_API_KEY
     models:
       # https://octo.ai/docs/getting-started/inference-models
       - name: meta-llama-3-70b-instruct
@@ -191,8 +198,7 @@ clients:
   # See https://docs.together.ai/docs/quickstart
   - type: openai-compatible
     name: together
-    api_base: https://api.together.xyz/v1
-    api_key: xxx
+    api_key: xxx                                      # ENV: {client_name}_API_KEY
     models:
       # https://docs.together.ai/docs/inference-models
       - name: meta-llama/Llama-3-70b-chat-hf
-- 
cgit v1.2.3