QOL improvement: allow integer values for Optimizer.set_learning_rate()

[maxfrei750]

Sometimes, I forget to specify the learning rate with a decimal point and get

Exception has occurred: Exception
Optimizer.set_learning_rate(): value should be a float or a dict!
  File "/src/mitsuba3/build/python/mitsuba/python/ad/optimizers.py", line 103, in set_learning_rate
    raise Exception('Optimizer.set_learning_rate(): value should be a float or a dict!')
  File "/src/mitsuba3/build/python/mitsuba/python/ad/optimizers.py", line 21, in __init__
    self.set_learning_rate(lr)
  File "/src/mitsuba3/build/python/mitsuba/python/ad/optimizers.py", line 261, in __init__
    super().__init__(lr, params)

Would it result in problems, if we added

if isinstance(lr, int):
    lr = float(lr)

to

mitsuba3/src/python/python/ad/optimizers.py

Lines 83 to 103 in 6902109

	def set_learning_rate(self, lr) -> None:
	"""
	Set the learning rate.
	Parameter ``lr`` (``float``, ``dict``):
	The new learning rate. A ``dict`` can be provided instead to
	specify the learning rate for specific parameters.
	"""
	# We use `dr.opaque` so the that the JIT compiler does not include
	# the learning rate as a scalar literal into generated code, which
	# would defeat kernel caching when updating learning rates.
	if isinstance(lr, float):
	self.lr_default = lr
	self.lr_default_v = dr.opaque(dr.detached_t(mi.Float), lr, shape=1)
	elif isinstance(lr, dict):
	for k, v in lr.items():
	self.lr[k] = v
	self.lr_v[k] = dr.opaque(dr.detached_t(mi.Float), v, shape=1)
	else:
	raise Exception('Optimizer.set_learning_rate(): value should be a float or a dict!')

?

[Speierers]

Thanks @maxfrei750 for reporting this. It will be fixed in the next release: 53143db