Update: ModelAnalysis.from_onnx(...) to work with ModelProto

[rahul-tuli]

Before this PR, sparezoo.analysis.ModelAnalysis.from_onnx(...) method
worked only with filepaths to onnx files, this can be inconvenient especially
if we need to analyze an already loaded onnx model.

import onnx
from sparsezoo.analysis import ModelAnalysis
onnx_model = onnx.load("./model.onnx")
ModelAnalysis.from_onnx(onnx_model)

Results in the following Error:

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/home/rahul/projects/sparsezoo/src/sparsezoo/analysis/analysis.py", line 435, in from_onnx
    model_onnx = onnx.load(onnx_file_path)
  File "/home/rahul/virtual_environments/sparsezoo3.9/lib/python3.9/site-packages/onnx/__init__.py", line 120, in load_model
    s = _load_bytes(f)
  File "/home/rahul/virtual_environments/sparsezoo3.9/lib/python3.9/site-packages/onnx/__init__.py", line 34, in _load_bytes
    with open(cast(Text, f), 'rb') as readable:
TypeError: expected str, bytes or os.PathLike object, not ModelProto

After this PR ModelAnalysis will work with both loaded onnx ModelProto and
local filepaths

The same test code(after this PR) now returns a ModelAnalysis object

ModelAnalysis(node_counts={'Conv': 53, 'Gemm': 1, 'BatchNormalization': ...

This will allow easier testing where we will not have to reload onnx files,
and, will be useful for upcoming PerformanceAnalysis api

Note: The changes added should not affect existing flows.