Spaces:
Paused
Paused
Update app.py
Browse files
app.py
CHANGED
|
@@ -38,13 +38,19 @@ retriever = vectorstore.as_retriever(search_type="similarity", search_kwargs={"k
|
|
| 38 |
# 2. Setup Robust LLM Interface via OpenAI-Compatible Endpoint
|
| 39 |
# -------------------------------------------------------------------------
|
| 40 |
print("Configuring LLM Interface...")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 41 |
chatModel = ChatOpenAI(
|
| 42 |
model="Qwen/Qwen2.5-1.5B-Instruct",
|
| 43 |
base_url="https://api-inference.huggingface.co/models/Qwen/Qwen2.5-1.5B-Instruct/v1/",
|
| 44 |
api_key=hf_token,
|
| 45 |
max_tokens=400,
|
| 46 |
temperature=0.3,
|
| 47 |
-
max_retries=2
|
|
|
|
| 48 |
)
|
| 49 |
|
| 50 |
system_prompt = (
|
|
|
|
| 38 |
# 2. Setup Robust LLM Interface via OpenAI-Compatible Endpoint
|
| 39 |
# -------------------------------------------------------------------------
|
| 40 |
print("Configuring LLM Interface...")
|
| 41 |
+
|
| 42 |
+
# Safety check to ensure the token is actually being loaded
|
| 43 |
+
if not hf_token:
|
| 44 |
+
print("WARNING: HF_TOKEN is missing. Please check your environment variables.")
|
| 45 |
+
|
| 46 |
chatModel = ChatOpenAI(
|
| 47 |
model="Qwen/Qwen2.5-1.5B-Instruct",
|
| 48 |
base_url="https://api-inference.huggingface.co/models/Qwen/Qwen2.5-1.5B-Instruct/v1/",
|
| 49 |
api_key=hf_token,
|
| 50 |
max_tokens=400,
|
| 51 |
temperature=0.3,
|
| 52 |
+
max_retries=4, # Increased from 2 to give the API more chances to connect
|
| 53 |
+
timeout=120.0 # FIX: Extended timeout to 2 minutes to survive cold starts
|
| 54 |
)
|
| 55 |
|
| 56 |
system_prompt = (
|