Spaces:
Sleeping
Sleeping
Update app.py
#4
by
CSB261
- opened
app.py
CHANGED
@@ -1,6 +1,10 @@
|
|
1 |
import gradio as gr
|
2 |
from huggingface_hub import InferenceClient
|
3 |
import os
|
|
|
|
|
|
|
|
|
4 |
|
5 |
MODELS = {
|
6 |
"Zephyr 7B Beta": "HuggingFaceH4/zephyr-7b-beta",
|
@@ -11,16 +15,32 @@ MODELS = {
|
|
11 |
"Mixtral 8x7B": "mistralai/Mistral-7B-Instruct-v0.3",
|
12 |
"Mixtral Nous-Hermes": "NousResearch/Nous-Hermes-2-Mixtral-8x7B-DPO",
|
13 |
"Cohere Command R+": "CohereForAI/c4ai-command-r-plus",
|
14 |
-
"Cohere Aya-23-35B": "CohereForAI/aya-23-35B"
|
|
|
15 |
}
|
16 |
|
17 |
def get_client(model_name):
|
|
|
|
|
18 |
model_id = MODELS[model_name]
|
19 |
hf_token = os.getenv("HF_TOKEN")
|
20 |
if not hf_token:
|
21 |
raise ValueError("HF_TOKEN environment variable is required")
|
22 |
return InferenceClient(model_id, token=hf_token)
|
23 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
24 |
def respond(
|
25 |
message,
|
26 |
chat_history,
|
@@ -31,7 +51,14 @@ def respond(
|
|
31 |
system_message,
|
32 |
):
|
33 |
try:
|
34 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
35 |
except ValueError as e:
|
36 |
chat_history.append((message, str(e)))
|
37 |
return chat_history
|
@@ -44,7 +71,6 @@ def respond(
|
|
44 |
|
45 |
try:
|
46 |
if "Cohere" in model_name:
|
47 |
-
# Cohere 모델을 위한 비스트리밍 처리
|
48 |
response = client.chat_completion(
|
49 |
messages,
|
50 |
max_tokens=max_tokens,
|
@@ -55,7 +81,6 @@ def respond(
|
|
55 |
chat_history.append((message, assistant_message))
|
56 |
yield chat_history
|
57 |
else:
|
58 |
-
# 다른 모델들을 위한 스트리밍 처리
|
59 |
stream = client.chat_completion(
|
60 |
messages,
|
61 |
max_tokens=max_tokens,
|
@@ -115,4 +140,4 @@ with gr.Blocks() as demo:
|
|
115 |
clear_button.click(clear_conversation, outputs=chatbot, queue=False)
|
116 |
|
117 |
if __name__ == "__main__":
|
118 |
-
demo.launch()
|
|
|
1 |
import gradio as gr
|
2 |
from huggingface_hub import InferenceClient
|
3 |
import os
|
4 |
+
import openai # OpenAI API 클라이언트 추가
|
5 |
+
|
6 |
+
# OpenAI API 키 설정
|
7 |
+
openai.api_key = os.getenv("OPENAI_API_KEY")
|
8 |
|
9 |
MODELS = {
|
10 |
"Zephyr 7B Beta": "HuggingFaceH4/zephyr-7b-beta",
|
|
|
15 |
"Mixtral 8x7B": "mistralai/Mistral-7B-Instruct-v0.3",
|
16 |
"Mixtral Nous-Hermes": "NousResearch/Nous-Hermes-2-Mixtral-8x7B-DPO",
|
17 |
"Cohere Command R+": "CohereForAI/c4ai-command-r-plus",
|
18 |
+
"Cohere Aya-23-35B": "CohereForAI/aya-23-35B",
|
19 |
+
"ChatGPT-4o-mini": "gpt-4o-mini" # ChatGPT-4o-mini 모델 추가
|
20 |
}
|
21 |
|
22 |
def get_client(model_name):
|
23 |
+
if model_name == "ChatGPT-4o-mini":
|
24 |
+
return None # OpenAI API는 따로 클라이언트를 생성할 필요 없음
|
25 |
model_id = MODELS[model_name]
|
26 |
hf_token = os.getenv("HF_TOKEN")
|
27 |
if not hf_token:
|
28 |
raise ValueError("HF_TOKEN environment variable is required")
|
29 |
return InferenceClient(model_id, token=hf_token)
|
30 |
|
31 |
+
def call_openai_api(content, system_message, max_tokens, temperature, top_p):
|
32 |
+
response = openai.ChatCompletion.create(
|
33 |
+
model="gpt-4o-mini",
|
34 |
+
messages=[
|
35 |
+
{"role": "system", "content": system_message},
|
36 |
+
{"role": "user", "content": content},
|
37 |
+
],
|
38 |
+
max_tokens=max_tokens,
|
39 |
+
temperature=temperature,
|
40 |
+
top_p=top_p,
|
41 |
+
)
|
42 |
+
return response.choices[0].message['content']
|
43 |
+
|
44 |
def respond(
|
45 |
message,
|
46 |
chat_history,
|
|
|
51 |
system_message,
|
52 |
):
|
53 |
try:
|
54 |
+
if model_name == "ChatGPT-4o-mini":
|
55 |
+
assistant_message = call_openai_api(message, system_message, max_tokens, temperature, top_p)
|
56 |
+
chat_history.append((message, assistant_message))
|
57 |
+
yield chat_history
|
58 |
+
else:
|
59 |
+
client = get_client(model_name)
|
60 |
+
if client is None:
|
61 |
+
raise ValueError(f"No client available for model: {model_name}")
|
62 |
except ValueError as e:
|
63 |
chat_history.append((message, str(e)))
|
64 |
return chat_history
|
|
|
71 |
|
72 |
try:
|
73 |
if "Cohere" in model_name:
|
|
|
74 |
response = client.chat_completion(
|
75 |
messages,
|
76 |
max_tokens=max_tokens,
|
|
|
81 |
chat_history.append((message, assistant_message))
|
82 |
yield chat_history
|
83 |
else:
|
|
|
84 |
stream = client.chat_completion(
|
85 |
messages,
|
86 |
max_tokens=max_tokens,
|
|
|
140 |
clear_button.click(clear_conversation, outputs=chatbot, queue=False)
|
141 |
|
142 |
if __name__ == "__main__":
|
143 |
+
demo.launch()
|