You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Currently, KServe application logs are not unified.
some places use import logging and some places use from kserve.logging import logger.
there is default log config defined but not applied by default for the application logger.
example format for huggingface server:
python -m huggingfaceserver --model_id=bert-base-uncased --model_name=bert
INFO:kserve:successfully loaded tokenizer for task: 5
Some weights of the model checkpoint at bert-base-uncased were not used when initializing BertForMaskedLM: ['cls.seq_relationship.weight', 'bert.pooler.dense.bias', 'cls.seq_relationship.bias', 'bert.pooler.dense.weight']
- This IS expected if you are initializing BertForMaskedLM from the checkpoint of a model trained on another task or with another architecture (e.g. initializing a BertForSequenceClassification model from a BertForPreTraining model).
- This IS NOT expected if you are initializing BertForMaskedLM from the checkpoint of a model that you expect to be exactly identical (initializing a BertForSequenceClassification model from a BertForSequenceClassification model).
INFO:kserve:successfully loaded huggingface model from path bert-base-uncased
INFO:kserve:Registering model: model
INFO:kserve:Setting max asyncio worker threads as 16
INFO:kserve:Starting uvicorn with 1 workers
2024-01-08 06:32:08.801 uvicorn.error INFO: Started server process [75012]
2024-01-08 06:32:08.801 uvicorn.error INFO: Waiting for application startup.
2024-01-08 06:32:08.804 75012 kserve INFO [start():62] Starting gRPC server on [::]:8081
2024-01-08 06:32:08.804 uvicorn.error INFO: Application startup complete.
2024-01-08 06:32:08.805 uvicorn.error INFO: Uvicorn running on http://0.0.0.0:8080 (Press CTRL+C to quit)
I would expect:
KServe application logs have default format
Storage initializer logs have default format
[optional] The user can overwrite the default format if necessary.
Information like timestamp will be helpful to know how long it take for model being downloaded in storage initializer and how long it take to load the model.
Anything else you would like to add:
My naive approach is to add logging.config.dictConfig(KSERVE_LOG_CONFIG) in kserve.logging and use kserve.logging import logger.
Concern on whether it will overwrite the user's application log format mentioned here.
Links to the design documents:
[Optional, start with the short-form RFC template to outline your ideas and get early feedback.]
[Required, use the longer-form design doc template to specify and discuss your design in more detail]
The text was updated successfully, but these errors were encountered:
/kind feature
Describe the solution you'd like
Currently, KServe application logs are not unified.
import logging
and some places usefrom kserve.logging import logger
.example format for huggingface server:
I would expect:
Information like
timestamp
will be helpful to know how long it take for model being downloaded in storage initializer and how long it take to load the model.Anything else you would like to add:
My naive approach is to add
logging.config.dictConfig(KSERVE_LOG_CONFIG)
in kserve.logging and usekserve.logging import logger
.Concern on whether it will overwrite the user's application log format mentioned here.
Links to the design documents:
[Optional, start with the short-form RFC template to outline your ideas and get early feedback.]
[Required, use the longer-form design doc template to specify and discuss your design in more detail]
The text was updated successfully, but these errors were encountered: