Spaces:
Sleeping
Sleeping
remove callbacks
Browse files- src/llm/adapters.py +4 -5
src/llm/adapters.py
CHANGED
|
@@ -11,8 +11,8 @@ from langchain_openai.chat_models import ChatOpenAI
|
|
| 11 |
# Legacy dependencies
|
| 12 |
from huggingface_hub import InferenceClient
|
| 13 |
from langchain_community.llms import HuggingFaceEndpoint
|
| 14 |
-
from langchain.callbacks.streaming_stdout import StreamingStdOutCallbackHandler
|
| 15 |
from langchain_community.chat_models.huggingface import ChatHuggingFace
|
|
|
|
| 16 |
|
| 17 |
|
| 18 |
from ..config.loader import load_config
|
|
@@ -71,7 +71,7 @@ def _create_dedicated_endpoint_client():
|
|
| 71 |
max_tokens = dedicated_config.get("max_tokens", 768)
|
| 72 |
|
| 73 |
# Set up the streaming callback handler
|
| 74 |
-
callback = StreamingStdOutCallbackHandler()
|
| 75 |
|
| 76 |
# Initialize the HuggingFaceEndpoint with streaming enabled
|
| 77 |
llm_qa = HuggingFaceEndpoint(
|
|
@@ -80,11 +80,10 @@ def _create_dedicated_endpoint_client():
|
|
| 80 |
repetition_penalty=1.03,
|
| 81 |
timeout=70,
|
| 82 |
huggingfacehub_api_token=api_key,
|
| 83 |
-
streaming=True
|
| 84 |
-
callbacks=[callback]
|
| 85 |
)
|
| 86 |
|
| 87 |
-
# Create a ChatHuggingFace instance with the streaming-enabled endpoint
|
| 88 |
return ChatHuggingFace(llm=llm_qa)
|
| 89 |
|
| 90 |
|
|
|
|
| 11 |
# Legacy dependencies
|
| 12 |
from huggingface_hub import InferenceClient
|
| 13 |
from langchain_community.llms import HuggingFaceEndpoint
|
|
|
|
| 14 |
from langchain_community.chat_models.huggingface import ChatHuggingFace
|
| 15 |
+
# from langchain.callbacks.streaming_stdout import StreamingStdOutCallbackHandler
|
| 16 |
|
| 17 |
|
| 18 |
from ..config.loader import load_config
|
|
|
|
| 71 |
max_tokens = dedicated_config.get("max_tokens", 768)
|
| 72 |
|
| 73 |
# Set up the streaming callback handler
|
| 74 |
+
# callback = StreamingStdOutCallbackHandler()
|
| 75 |
|
| 76 |
# Initialize the HuggingFaceEndpoint with streaming enabled
|
| 77 |
llm_qa = HuggingFaceEndpoint(
|
|
|
|
| 80 |
repetition_penalty=1.03,
|
| 81 |
timeout=70,
|
| 82 |
huggingfacehub_api_token=api_key,
|
| 83 |
+
streaming=True
|
| 84 |
+
# callbacks=[callback]
|
| 85 |
)
|
| 86 |
|
|
|
|
| 87 |
return ChatHuggingFace(llm=llm_qa)
|
| 88 |
|
| 89 |
|