baalajimaestro/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

docs: Add chat examples to openapi ui

Browse source
This commit is contained in:
Andrei Betlen 2024-03-19 10:52:53 -04:00
parent 60d8498f21
commit f7decc9562
1 changed files with 59 additions and 9 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

68
llama_cpp/server/app.py
View file

@ -12,14 +12,7 @@ import llama_cpp
import anyio import anyio
from anyio.streams.memory import MemoryObjectSendStream from anyio.streams.memory import MemoryObjectSendStream
from starlette.concurrency import run_in_threadpool, iterate_in_threadpool from starlette.concurrency import run_in_threadpool, iterate_in_threadpool
from fastapi import ( from fastapi import Depends, FastAPI, APIRouter, Request, HTTPException, status, Body
Depends,
FastAPI,
APIRouter,
Request,
HTTPException,
status,
)
from fastapi.middleware import Middleware from fastapi.middleware import Middleware
from fastapi.middleware.cors import CORSMiddleware from fastapi.middleware.cors import CORSMiddleware
from fastapi.security import HTTPBearer from fastapi.security import HTTPBearer
@ -356,7 +349,64 @@ async def create_embedding(
) )
async def create_chat_completion( async def create_chat_completion(
request: Request, request: Request,
body: CreateChatCompletionRequest, body: CreateChatCompletionRequest = Body(
openapi_examples={
"normal": {
"summary": "Chat Completion",
"value": {
"model": "gpt-3.5-turbo",
"messages": [
{"role": "system", "content": "You are a helpful assistant."},
{"role": "user", "content": "What is the capital of France?"},
],
},
},
"json_mode": {
"summary": "JSON Mode",
"value": {
"model": "gpt-3.5-turbo",
"messages": [
{"role": "system", "content": "You are a helpful assistant."},
{"role": "user", "content": "Who won the world series in 2020"},
],
"response_format": { "type": "json_object" }
},
},
"tool_calling": {
"summary": "Tool Calling",
"value": {
"model": "gpt-3.5-turbo",
"messages": [
{"role": "system", "content": "You are a helpful assistant."},
{"role": "user", "content": "Extract Jason is 30 years old."},
],
"tools": [
{
"type": "function",
"function": {
"name": "User",
"description": "User record",
"parameters": {
"type": "object",
"properties": {
"name": {"type": "string"},
"age": {"type": "number"},
},
"required": ["name", "age"],
},
}
}
],
"tool_choice": {
"type": "function",
"function": {
"name": "User",
}
}
},
},
}
),
llama_proxy: LlamaProxy = Depends(get_llama_proxy), llama_proxy: LlamaProxy = Depends(get_llama_proxy),
) -> llama_cpp.ChatCompletion: ) -> llama_cpp.ChatCompletion:
exclude = { exclude = {