[Logger Refactor] Add tb summary writer like functions

[rahul-tuli]

This PR adds a log method with similar signature to https://github.com/pytorch/pytorch/blob/92998693a9455af6259cae468265f01cfff8810e/torch/utils/tensorboard/writer.py#L353 to our LoggerManager

Test:

#local.py
from logging import INFO
from sparseml.core import LoggerManager

def check_tb_type_logging_functions():
    logger_manager = LoggerManager()
    logger_manager.metric.add_scalar(value="Hello World!", level=INFO)
    logger_manager.metric.add_scalars(values={1: "Hello World!"}, level=INFO, tag="my_scalar")

if __name__ == "__main__":
    check_tb_type_logging_functions()

Output:

$ python local.py                                                                                                         (add-tb-like-logging-function|✚1…2⚑7)
2024-01-02 09:09:47 sparseml.core.logger.logger INFO     Logging all SparseML modifier-level logs to sparse_logs/02-01-2024_09.09.47.log
python defaul_tag step None: Hello World!
python my_scalar step None: {1: 'Hello World!'}

Based off of #1932

---

• To see the specific tasks where the Asana app for GitHub is being used, see below:
  • https://app.asana.com/0/0/1205848553025320