support for initial_epoch in keras TqdmCallback

[fabiocarrara]

• I have marked all applicable categories:
  • exception-raising bug
  • visual output bug
  • documentation request (i.e. "X is missing from the documentation." If instead I want to ask "how to use X?" I understand StackOverflow#tqdm is more appropriate)
  • new feature request
• I have visited the source website, and in particular read the known issues
• I have searched through the issue tracker for duplicates
• I have mentioned version numbers, operating system and evnironment, where applicable:

version 4.59.0, Google Colab

Would love the keras TqdmCallback to support initial_epochs in keras .fit() call to correctly initialize the starting point of the progress bar when resuming training.

Passing the initial epoch using initial kwarg does not seem to work, as it is not considered in on_train_begin().

tqdm/tqdm/keras.py

Lines 69 to 73 in bcce20f

	def on_train_begin(self, *_, **__):
	params = self.params.get
	auto_total = params('epochs', params('nb_epoch', None))
	if auto_total is not None:
	self.epoch_bar.reset(total=auto_total)

Colab Gist reproducing the problem:
https://colab.research.google.com/gist/fabiocarrara/3f81bed4bc735291205b1b42fd5c1567/untitled1.ipynb

[javoweb]

Hi @casperdcl, are you working on this? I want to contribute to this fix.

[casperdcl]

Oh sure go ahead; I wasn't going to look at this for another couple of weeks