Update app.py
Browse files
app.py
CHANGED
|
@@ -6,11 +6,7 @@ from openai import OpenAI
|
|
| 6 |
|
| 7 |
subprocess.Popen("bash /home/user/app/start.sh", shell=True)
|
| 8 |
|
| 9 |
-
client = OpenAI(
|
| 10 |
-
base_url="http://0.0.0.0:8000/v1",
|
| 11 |
-
api_key="sk-local",
|
| 12 |
-
timeout=600
|
| 13 |
-
)
|
| 14 |
|
| 15 |
|
| 16 |
def respond(
|
|
@@ -24,7 +20,7 @@ def respond(
|
|
| 24 |
messages = []
|
| 25 |
if system_message:
|
| 26 |
messages = [{"role": "system", "content": system_message}]
|
| 27 |
-
|
| 28 |
for user, assistant in history:
|
| 29 |
if user:
|
| 30 |
messages.append({"role": "user", "content": user})
|
|
@@ -41,14 +37,43 @@ def respond(
|
|
| 41 |
temperature=temperature,
|
| 42 |
top_p=top_p,
|
| 43 |
stream=True,
|
| 44 |
-
tools=[
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 45 |
)
|
| 46 |
|
| 47 |
print("messages", messages)
|
| 48 |
output = ""
|
| 49 |
for chunk in stream:
|
| 50 |
delta = chunk.choices[0].delta
|
| 51 |
-
|
| 52 |
try:
|
| 53 |
output += delta.reasoning_content
|
| 54 |
except:
|
|
@@ -60,9 +85,8 @@ def respond(
|
|
| 60 |
print(f"[Error] {e}")
|
| 61 |
yield "⚠️ Llama.cpp server error"
|
| 62 |
|
| 63 |
-
|
| 64 |
-
|
| 65 |
-
)
|
| 66 |
|
| 67 |
if __name__ == "__main__":
|
| 68 |
demo.launch(show_api=False)
|
|
|
|
| 6 |
|
| 7 |
subprocess.Popen("bash /home/user/app/start.sh", shell=True)
|
| 8 |
|
| 9 |
+
client = OpenAI(base_url="http://0.0.0.0:8000/v1", api_key="sk-local", timeout=600)
|
|
|
|
|
|
|
|
|
|
|
|
|
| 10 |
|
| 11 |
|
| 12 |
def respond(
|
|
|
|
| 20 |
messages = []
|
| 21 |
if system_message:
|
| 22 |
messages = [{"role": "system", "content": system_message}]
|
| 23 |
+
|
| 24 |
for user, assistant in history:
|
| 25 |
if user:
|
| 26 |
messages.append({"role": "user", "content": user})
|
|
|
|
| 37 |
temperature=temperature,
|
| 38 |
top_p=top_p,
|
| 39 |
stream=True,
|
| 40 |
+
tools=[
|
| 41 |
+
{
|
| 42 |
+
"type": "function",
|
| 43 |
+
"function": {
|
| 44 |
+
"name": "browser_search",
|
| 45 |
+
"description": (
|
| 46 |
+
"Search the web for a given query and return the most relevant results."
|
| 47 |
+
),
|
| 48 |
+
"parameters": {
|
| 49 |
+
"type": "object",
|
| 50 |
+
"properties": {
|
| 51 |
+
"query": {
|
| 52 |
+
"type": "string",
|
| 53 |
+
"description": "The search query string.",
|
| 54 |
+
},
|
| 55 |
+
"max_results": {
|
| 56 |
+
"type": "integer",
|
| 57 |
+
"description": (
|
| 58 |
+
"Maximum number of search results to return. "
|
| 59 |
+
"If omitted the service will use its default."
|
| 60 |
+
),
|
| 61 |
+
"default": 5,
|
| 62 |
+
},
|
| 63 |
+
},
|
| 64 |
+
"required": ["query"],
|
| 65 |
+
},
|
| 66 |
+
},
|
| 67 |
+
},
|
| 68 |
+
{"type": "code_interpreter"},
|
| 69 |
+
],
|
| 70 |
)
|
| 71 |
|
| 72 |
print("messages", messages)
|
| 73 |
output = ""
|
| 74 |
for chunk in stream:
|
| 75 |
delta = chunk.choices[0].delta
|
| 76 |
+
|
| 77 |
try:
|
| 78 |
output += delta.reasoning_content
|
| 79 |
except:
|
|
|
|
| 85 |
print(f"[Error] {e}")
|
| 86 |
yield "⚠️ Llama.cpp server error"
|
| 87 |
|
| 88 |
+
|
| 89 |
+
demo = gr.ChatInterface(respond)
|
|
|
|
| 90 |
|
| 91 |
if __name__ == "__main__":
|
| 92 |
demo.launch(show_api=False)
|