Upload 3 files
Browse files- app.py +65 -0
- requirements.txt +3 -0
- test.py +19 -0
app.py
ADDED
@@ -0,0 +1,65 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
# encoding = "utf-8"
|
2 |
+
|
3 |
+
|
4 |
+
'''
|
5 |
+
This is a mediator: a gradio server for OpenAI APIs
|
6 |
+
'''
|
7 |
+
|
8 |
+
import os
|
9 |
+
import json
|
10 |
+
import argparse
|
11 |
+
import gradio as gr
|
12 |
+
import requests
|
13 |
+
from openai import OpenAI
|
14 |
+
|
15 |
+
def http_bot(messages, argsbox):
|
16 |
+
|
17 |
+
args = eval(argsbox)
|
18 |
+
messages = eval(messages)
|
19 |
+
|
20 |
+
print(messages)
|
21 |
+
print(argsbox)
|
22 |
+
|
23 |
+
# api_key = args["api_key"]
|
24 |
+
# base_url = args["base_url"]
|
25 |
+
# model = args["model"]
|
26 |
+
# temperature = args["temperature"]
|
27 |
+
# max_tokens = args["max_tokens"]
|
28 |
+
#
|
29 |
+
# headers = {
|
30 |
+
# "Content-Type": "application/json",
|
31 |
+
# "Authorization": f"Bearer {}" # Users will provide their own OPENAI_API_KEY
|
32 |
+
# }
|
33 |
+
|
34 |
+
|
35 |
+
client = OpenAI(api_key=args["api_key"], base_url = args["base_url"])
|
36 |
+
# n = 0
|
37 |
+
# while True:
|
38 |
+
# try:
|
39 |
+
chat_completion = client.chat.completions.create(
|
40 |
+
messages=messages,
|
41 |
+
model=args["model"], #"gpt-3.5-turbo-16k", # "gpt-3.5-turbo", # gpt-4-1106-preview
|
42 |
+
temperature=float(args["temperature"]),
|
43 |
+
max_tokens=int(args["max_tokens"])
|
44 |
+
)
|
45 |
+
# break
|
46 |
+
# except Exception as e:
|
47 |
+
# continue
|
48 |
+
print(chat_completion)
|
49 |
+
return chat_completion.choices[0].message.content
|
50 |
+
|
51 |
+
|
52 |
+
|
53 |
+
with gr.Blocks() as demo:
|
54 |
+
gr.Markdown("# vLLM text completion demo\n")
|
55 |
+
inputbox = gr.Textbox(label="Input",
|
56 |
+
placeholder="Enter text and press ENTER")
|
57 |
+
argsbox = gr.Textbox(label="Args", placeholder="a dict of {api_key, base_url, model, temperature, max_tokens}")
|
58 |
+
outputbox = gr.Textbox(label="Output",
|
59 |
+
placeholder="Generated result from the model")
|
60 |
+
submit = gr.Button("Submit")
|
61 |
+
|
62 |
+
submit.click(http_bot, [inputbox, argsbox], [outputbox], api_name="submit")
|
63 |
+
|
64 |
+
demo.launch(share=True)
|
65 |
+
|
requirements.txt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
gradio==4.11.0
|
2 |
+
openai==1.6.0
|
3 |
+
Requests==2.31.0
|
test.py
ADDED
@@ -0,0 +1,19 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
from gradio_client import Client
|
2 |
+
import json
|
3 |
+
|
4 |
+
messages = json.dumps([{"role":"user","content":"hi(from macbook)"}])
|
5 |
+
#
|
6 |
+
args = json.dumps({"api_key":"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",
|
7 |
+
"base_url":"https://api.openai.com/v1",
|
8 |
+
"model":"gpt-3.5-turbo",
|
9 |
+
"temperature":"1.0",
|
10 |
+
"max_tokens":"1024" })
|
11 |
+
|
12 |
+
|
13 |
+
client = Client("https://f2569cc8b84a4386fa.gradio.live/")
|
14 |
+
result = client.predict(
|
15 |
+
messages, # str in 'Input' Textbox component
|
16 |
+
args, # str in 'Args' Textbox component
|
17 |
+
api_name="/submit"
|
18 |
+
)
|
19 |
+
print(result)
|