updated app
Browse files
app.py
CHANGED
|
@@ -9,9 +9,10 @@ from sentence_transformers import CrossEncoder
|
|
| 9 |
|
| 10 |
QWEN_MODEL="qwen/qwen3-235b-a22b:free"
|
| 11 |
DEEPSEEK_MODEL="deepseek/deepseek-chat-v3.1:free"
|
|
|
|
| 12 |
api_key = os.getenv("OPENROUTER_API_KEY")
|
| 13 |
deepseek_key = os.getenv("DEEPSEEK_API_KEY")
|
| 14 |
-
client = OpenAI(base_url="https://openrouter.ai/api/v1", api_key=
|
| 15 |
|
| 16 |
# open ai method
|
| 17 |
#api_key = os.getenv("OPENAI_API_KEY")
|
|
@@ -24,7 +25,7 @@ collection = load_collection()
|
|
| 24 |
encoder = load_encoder()
|
| 25 |
reranker = CrossEncoder("BAAI/bge-reranker-large")
|
| 26 |
|
| 27 |
-
def reformulate_query(user_question, model_name=
|
| 28 |
prompt = f"""你是一个BangDream知识检索助手。请把用户的问题扩写或转写为适合知识库语义检索的检索语句,涵盖所有可能的提问方式或同义关键词。
|
| 29 |
用户问题:{user_question}
|
| 30 |
"""
|
|
@@ -101,7 +102,7 @@ def respond(
|
|
| 101 |
|
| 102 |
response = ""
|
| 103 |
stream = client.chat.completions.create(
|
| 104 |
-
model=
|
| 105 |
messages=messages,
|
| 106 |
temperature=temperature,
|
| 107 |
max_tokens=max_tokens,
|
|
|
|
| 9 |
|
| 10 |
QWEN_MODEL="qwen/qwen3-235b-a22b:free"
|
| 11 |
DEEPSEEK_MODEL="deepseek/deepseek-chat-v3.1:free"
|
| 12 |
+
GPT_OSS_MODEL="openai/gpt-oss-20b:free"
|
| 13 |
api_key = os.getenv("OPENROUTER_API_KEY")
|
| 14 |
deepseek_key = os.getenv("DEEPSEEK_API_KEY")
|
| 15 |
+
client = OpenAI(base_url="https://openrouter.ai/api/v1", api_key=api_key)
|
| 16 |
|
| 17 |
# open ai method
|
| 18 |
#api_key = os.getenv("OPENAI_API_KEY")
|
|
|
|
| 25 |
encoder = load_encoder()
|
| 26 |
reranker = CrossEncoder("BAAI/bge-reranker-large")
|
| 27 |
|
| 28 |
+
def reformulate_query(user_question, model_name=GPT_OSS_MODEL):
|
| 29 |
prompt = f"""你是一个BangDream知识检索助手。请把用户的问题扩写或转写为适合知识库语义检索的检索语句,涵盖所有可能的提问方式或同义关键词。
|
| 30 |
用户问题:{user_question}
|
| 31 |
"""
|
|
|
|
| 102 |
|
| 103 |
response = ""
|
| 104 |
stream = client.chat.completions.create(
|
| 105 |
+
model=GPT_OSS_MODEL,
|
| 106 |
messages=messages,
|
| 107 |
temperature=temperature,
|
| 108 |
max_tokens=max_tokens,
|