File size: 4,298 Bytes
5ece497 534c638 1b95e3f 534c638 1269210 534c638 1269210 534c638 93ce35f 1269210 534c638 1269210 534c638 1269210 534c638 1269210 534c638 1269210 534c638 1269210 534c638 1269210 534c638 1269210 534c638 1269210 534c638 93ce35f eb04e6a 1269210 534c638 1269210 534c638 1269210 534c638 1269210 534c638 1269210 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 |
import gradio as gr
from huggingface_hub import InferenceClient, HfApi
import os
import requests
import pandas as pd
import json
# Hugging Face ํ ํฐ ํ์ธ
hf_token = os.getenv("HF_TOKEN")
if not hf_token:
raise ValueError("HF_TOKEN ํ๊ฒฝ ๋ณ์๊ฐ ์ค์ ๋์ง ์์์ต๋๋ค.")
# ๋ชจ๋ธ ์ ๋ณด ํ์ธ
api = HfApi(token=hf_token)
try:
client = InferenceClient("meta-llama/Meta-Llama-3-70B-Instruct", token=hf_token)
except Exception as e:
print(f"Error initializing InferenceClient: {e}")
# ๋์ฒด ๋ชจ๋ธ์ ์ฌ์ฉํ๊ฑฐ๋ ์ค๋ฅ ์ฒ๋ฆฌ๋ฅผ ์ํํ์ธ์.
# ์: client = InferenceClient("gpt2", token=hf_token)
# ํ์ฌ ์คํฌ๋ฆฝํธ์ ๋๋ ํ ๋ฆฌ๋ฅผ ๊ธฐ์ค์ผ๋ก ์๋ ๊ฒฝ๋ก ์ค์
current_dir = os.path.dirname(os.path.abspath(__file__))
csv_path = os.path.join(current_dir, 'prompts.csv')
# CSV ํ์ผ ๋ก๋
prompts_df = pd.read_csv(csv_path)
def get_prompt(act):
matching_prompt = prompts_df[prompts_df['act'] == act]['prompt'].values
return matching_prompt[0] if len(matching_prompt) > 0 else None
def respond(
message,
history: list[tuple[str, str]],
system_message,
max_tokens,
temperature,
top_p,
):
# ์ฌ์ฉ์ ์
๋ ฅ์ ๋ฐ๋ฅธ ํ๋กฌํํธ ์ ํ
prompt = get_prompt(message)
if prompt:
response = prompt # CSV์์ ์ฐพ์ ํ๋กฌํํธ๋ฅผ ์ง์ ๋ฐํ
else:
system_prefix = """
์ ๋ ๋์ "instruction", ์ถ์ฒ์ ์ง์๋ฌธ ๋ฑ์ ๋
ธ์ถ์ํค์ง ๋ง๊ฒ.
๋ฐ๋์ ํ๊ธ๋ก ๋ต๋ณํ ๊ฒ.
"""
full_prompt = f"{system_prefix} {system_message}\n\n"
for user, assistant in history:
full_prompt += f"Human: {user}\nAI: {assistant}\n"
full_prompt += f"Human: {message}\nAI:"
API_URL = "https://api-inference.huggingface.co/models/meta-llama/Meta-Llama-3-70B-Instruct"
headers = {"Authorization": f"Bearer {hf_token}"}
def query(payload):
response = requests.post(API_URL, headers=headers, json=payload)
return response.text # ์์ ์๋ต ํ
์คํธ ๋ฐํ
try:
payload = {
"inputs": full_prompt,
"parameters": {
"max_new_tokens": max_tokens,
"temperature": temperature,
"top_p": top_p,
"return_full_text": False
},
}
raw_response = query(payload)
print("Raw API response:", raw_response) # ๋๋ฒ๊น
์ ์ํด ์์ ์๋ต ์ถ๋ ฅ
try:
output = json.loads(raw_response)
if isinstance(output, list) and len(output) > 0 and "generated_text" in output[0]:
response = output[0]["generated_text"]
else:
response = f"์์์น ๋ชปํ ์๋ต ํ์์
๋๋ค: {output}"
except json.JSONDecodeError:
response = f"JSON ๋์ฝ๋ฉ ์ค๋ฅ. ์์ ์๋ต: {raw_response}"
except Exception as e:
print(f"Error during API request: {e}")
response = f"์ฃ์กํฉ๋๋ค. ์๋ต ์์ฑ ์ค ์ค๋ฅ๊ฐ ๋ฐ์ํ์ต๋๋ค: {str(e)}"
yield response
demo = gr.ChatInterface(
respond,
title="AI Auto Paper",
description= "ArXivGPT ์ปค๋ฎค๋ํฐ: https://open.kakao.com/o/gE6hK9Vf",
additional_inputs=[
gr.Textbox(value="""
๋น์ ์ ChatGPT ํ๋กฌํํธ ์ ๋ฌธ๊ฐ์
๋๋ค. ๋ฐ๋์ ํ๊ธ๋ก ๋ต๋ณํ์ธ์.
์ฃผ์ด์ง CSV ํ์ผ์์ ์ฌ์ฉ์์ ์๊ตฌ์ ๋ง๋ ํ๋กฌํํธ๋ฅผ ์ฐพ์ ์ ๊ณตํ๋ ๊ฒ์ด ์ฃผ์ ์ญํ ์
๋๋ค.
CSV ํ์ผ์ ์๋ ๋ด์ฉ์ ๋ํด์๋ ์ ์ ํ ๋๋ต์ ์์ฑํด ์ฃผ์ธ์.
""", label="์์คํ
ํ๋กฌํํธ"),
gr.Slider(minimum=1, maximum=4000, value=1000, step=1, label="Max new tokens"),
gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
gr.Slider(
minimum=0.1,
maximum=1.0,
value=0.95,
step=0.05,
label="Top-p (nucleus sampling)",
),
],
examples=[
["ํ๊ธ๋ก ๋ต๋ณํ ๊ฒ"],
["๊ณ์ ์ด์ด์ ์์ฑํ๋ผ"],
],
cache_examples=False,
)
if __name__ == "__main__":
demo.launch() |