Sagemaker SDK on lambda function invoking serverless endpoint

[KaramRazooq]

I'm getting the following message while running the Sagemaker SDK on my lambda function.
I have installed the correct libraries that are comparable with python 3.11 and the lambda.

The log data from lambda:

sagemaker.config INFO - Not applying SDK defaults from location: /etc/xdg/sagemaker/config.yaml
--
sagemaker.config INFO - Not applying SDK defaults from location: /home/sbx_user1051/.config/sagemaker/config.yaml

The log from the endpoint

2023-09-16T21:55:46,174 [INFO ] W-9000-model_1.0-stdout MODEL_LOG - Executing input_fn from inference.py ...
--
2023-09-16T21:55:46,174 [INFO ] W-9000-model_1.0 org.pytorch.serve.wlm.WorkerThread - Backend response time: 3557
2023-09-16T21:55:46,174 [INFO ] W-9000-model_1.0-stdout MODEL_LOG - Executing predict_fn from inference.py ...
2023-09-16T21:55:46,174 [INFO ] W-9000-model_1.0-stdout MODEL_LOG - Executing output_fn from inference.py ...
2023-09-16T21:55:46,174 [INFO ] W-9000-model_1.0 ACCESS_LOG - /127.0.0.1:35074 "POST /invocations HTTP/1.1" 200 3558
2023-09-16T21:55:46,174 [INFO ] W-9000-model_1.0 TS_METRICS - Requests2XX.Count:1\|#Level:Host\|#hostname:169.254.29.125,timestamp:1694901018

My code is

from sagemaker.pytorch.model import PyTorchModel, PyTorchPredictor 
from sagemaker.deserializers import JSONDeserializer
    
predictor = PyTorchPredictor(endpoint_name=ENDPOINT_NAME, deserializer=JSONDeserializer())
result = predictor.predict(decode)

The predictor function is what's breaking the run and sending the INFO level log

any thoughts on how to fix this?
I tried to look into documentation on what could fix this but did not find anything

[dongreenberg]

I'm getting this as well, running locally on a Mac.

INFO | 2023-09-21 20:29:41.200081 | Found credentials in shared credentials file: ~/.aws/credentials
sagemaker.config INFO - Not applying SDK defaults from location: /Library/Application Support/sagemaker/config.yaml
sagemaker.config INFO - Not applying SDK defaults from location: /Users/donny/Library/Application Support/sagemaker/config.yaml

[RistovaIvona]

I am also experiencing this issue while running the SageMaker SDK. Has anyone found a solution to this problem?

sagemaker.config INFO - Not applying SDK defaults from location: /etc/xdg/sagemaker/config.yaml
sagemaker.config INFO - Not applying SDK defaults from location: /home/ec2-user/.config/sagemaker/config.yaml
sagemaker.config INFO - Not applying SDK defaults from location: /etc/xdg/sagemaker/config.yaml
sagemaker.config INFO - Not applying SDK defaults from location: /home/ec2-user/.config/sagemaker/config.yaml
sagemaker.config INFO - Not applying SDK defaults from location: /etc/xdg/sagemaker/config.yaml
sagemaker.config INFO - Not applying SDK defaults from location: /home/ec2-user/.config/sagemaker/config.yaml

[djwbamboo]

From what I'm observing it seems to be a be associated with versions >=2.183.0. I solved by downgrading:

pip install sagemaker==2.182.0

[ConsciousCoder07]

Thanks @djwbamboo, that worked for me

[brifordwylie]

My logs were full of these messages....IMO they should make this a debug message instead of info :)

For now just pop this into your code and it should suppress the messages

# Sagemaker continuously complains about config, so we'll suppress it
logging.getLogger("sagemaker.config").setLevel(logging.WARNING)

It's extremely annoying. Needed to disable this at various places to avoid cluttering logs.

[bouachalazhar]

From what I'm observing it seems to be a be associated with versions >=2.183.0. I solved by downgrading:

pip install sagemaker==2.182.0

If I can't do it what can I do ?

[cldixon]

My logs were full of these messages....IMO they should make this a debug message instead of info :)

For now just pop this into your code and it should suppress the messages

# Sagemaker continuously complains about config, so we'll suppress it
logging.getLogger("sagemaker.config").setLevel(logging.WARNING)

This worked for me, but I had to place the line before import sagemaker.