tinynn/converter/operators/optimize.py (7 lines):
	- line 1357: # TODO: Support this case
	- line 1366: # TODO: Support multi-multi mappings
	- line 2178: # TODO: Figure out the rev index
	- line 2251: # TODO: Figure out the rev index
	- line 4327: # TODO: support multi indices
	- line 4340: # TODO: support multi indices
	- line 4355: # TODO: support multi indices


tinynn/graph/tracer.py (5 lines):
	- line 948: # TODO: Fix the case if user calls `numel` on `torch.Size`
	- line 1071: # TODO: Fix the case if user calls `numel` on `torch.Size`
	- line 2389: # TODO: Selective module importing
	- line 2980: # TODO: Support the following case
	- line 3131: # TODO: Implement this codepath


tinynn/graph/quantization/quantizer.py (2 lines):
	- line 577: # TODO: Enable QAT analysis for TensorRT
	- line 2821: # TODO: Support complex cases


tinynn/graph/modifier.py (2 lines):
	- line 3031: # TODO: Is it possible to skip when center_group has only one element?
	- line 3297: # TODO: It is more reasonable to arrange according to the forward order


tinynn/prune/netadapt_pruner.py (1 line):
	- line 96: # TODO: Implement budget type: weights, latency


tinynn/prune/oneshot_pruner.py (1 line):
	- line 73: # TODO: 为了保持剪枝精度，目前暂时先将所有的常量参与的剪枝子图取消


tinynn/converter/operators/torch/base.py (1 line):
	- line 116: # TODO: Better selection for multiple schemas


tinynn/graph/configs/gen_creation_funcs_yml.py (1 line):
	- line 7: # TODO: Better detection


tinynn/converter/operators/torch/aten.py (1 line):
	- line 3859: # TODO: +/- inf check for variable tensors


tinynn/graph/configs/gen_funcs_yml.py (1 line):
	- line 10: # TODO: Better detection