Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
@@ -1,48 +1,61 @@
|
|
1 |
import gradio as gr
|
2 |
from huggingface_hub import InferenceClient
|
|
|
|
|
|
|
|
|
|
|
3 |
|
4 |
def respond(message, history, system_message, max_tokens, temperature, top_p, selected_model):
|
5 |
-
|
6 |
-
|
7 |
-
|
8 |
-
|
9 |
-
|
10 |
-
|
11 |
-
|
12 |
-
|
13 |
-
|
14 |
-
|
15 |
-
|
16 |
-
|
17 |
-
|
18 |
-
|
19 |
-
|
20 |
-
|
21 |
-
|
22 |
-
|
23 |
-
|
24 |
-
|
25 |
-
|
|
|
|
|
|
|
|
|
|
|
26 |
|
|
|
27 |
models = {
|
28 |
"deepseek-ai/DeepSeek-Coder-V2-Instruct": "DeepSeek-Coder-V2-Instruct",
|
29 |
"CohereForAI/c4ai-command-r-plus": "Cohere Command-R Plus",
|
30 |
"meta-llama/Meta-Llama-3.1-8B-Instruct": "Meta-Llama-3.1-8B-Instruct",
|
31 |
"bartowski/DeepSeek-V2-Chat-0628-GGUF": "DeepSeek-V2-Chat-0628-GGUF",
|
32 |
"google/gemma-7b": "Gemma-7b",
|
33 |
-
"openai-community/gpt2": "
|
34 |
}
|
35 |
|
36 |
demo = gr.ChatInterface(
|
37 |
respond,
|
38 |
additional_inputs=[
|
39 |
-
gr.Textbox(value="
|
40 |
gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="최대 새 토큰 수"),
|
41 |
-
gr.Slider(minimum=0.1, maximum=
|
42 |
gr.Slider(minimum=0.1, maximum=1.0, value=0.95, step=0.05, label="Top-p (핵 샘플링)"),
|
43 |
gr.Radio(list(models.keys()), value=list(models.keys())[0], label="언어 모델 선택", info="사용할 언어 모델을 선택하세요")
|
44 |
],
|
45 |
)
|
46 |
|
47 |
if __name__ == "__main__":
|
48 |
-
|
|
|
|
|
|
1 |
import gradio as gr
|
2 |
from huggingface_hub import InferenceClient
|
3 |
+
import traceback
|
4 |
+
import os
|
5 |
+
|
6 |
+
# Hugging Face API 토큰을 환경 변수에서 가져옵니다.
|
7 |
+
hf_token = os.getenv("HF_TOKEN")
|
8 |
|
9 |
def respond(message, history, system_message, max_tokens, temperature, top_p, selected_model):
|
10 |
+
try:
|
11 |
+
# API 토큰을 사용하여 InferenceClient를 초기화합니다.
|
12 |
+
client = InferenceClient(model=selected_model, token=hf_token)
|
13 |
+
|
14 |
+
messages = [{"role": "system", "content": system_message}]
|
15 |
+
for val in history:
|
16 |
+
if val[0]:
|
17 |
+
messages.append({"role": "user", "content": val[0]})
|
18 |
+
if val[1]:
|
19 |
+
messages.append({"role": "assistant", "content": val[1]})
|
20 |
+
messages.append({"role": "user", "content": message})
|
21 |
+
|
22 |
+
response = ""
|
23 |
+
for message in client.chat_completion(
|
24 |
+
messages,
|
25 |
+
max_tokens=max_tokens,
|
26 |
+
stream=True,
|
27 |
+
temperature=temperature,
|
28 |
+
top_p=top_p,
|
29 |
+
):
|
30 |
+
token = message.choices[0].delta.content
|
31 |
+
response += token
|
32 |
+
yield response
|
33 |
+
except Exception as e:
|
34 |
+
error_msg = f"오류 발생: {str(e)}\n\n상세 오류:\n{traceback.format_exc()}"
|
35 |
+
yield error_msg
|
36 |
|
37 |
+
# 모델 목록 (무료 및 유료 모델 포함)
|
38 |
models = {
|
39 |
"deepseek-ai/DeepSeek-Coder-V2-Instruct": "DeepSeek-Coder-V2-Instruct",
|
40 |
"CohereForAI/c4ai-command-r-plus": "Cohere Command-R Plus",
|
41 |
"meta-llama/Meta-Llama-3.1-8B-Instruct": "Meta-Llama-3.1-8B-Instruct",
|
42 |
"bartowski/DeepSeek-V2-Chat-0628-GGUF": "DeepSeek-V2-Chat-0628-GGUF",
|
43 |
"google/gemma-7b": "Gemma-7b",
|
44 |
+
"openai-community/gpt2": "GPT-2"
|
45 |
}
|
46 |
|
47 |
demo = gr.ChatInterface(
|
48 |
respond,
|
49 |
additional_inputs=[
|
50 |
+
gr.Textbox(value="당신은 친절한 챗봇입니다.", label="시스템 메시지"),
|
51 |
gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="최대 새 토큰 수"),
|
52 |
+
gr.Slider(minimum=0.1, maximum=2.0, value=0.7, step=0.1, label="온도"),
|
53 |
gr.Slider(minimum=0.1, maximum=1.0, value=0.95, step=0.05, label="Top-p (핵 샘플링)"),
|
54 |
gr.Radio(list(models.keys()), value=list(models.keys())[0], label="언어 모델 선택", info="사용할 언어 모델을 선택하세요")
|
55 |
],
|
56 |
)
|
57 |
|
58 |
if __name__ == "__main__":
|
59 |
+
if not hf_token:
|
60 |
+
print("경고: HF_TOKEN 환경 변수가 설정되지 않았습니다. 일부 모델에 접근할 수 없을 수 있습니다.")
|
61 |
+
demo.launch(share=True) # public link 생성
|