Sparse Tensor Onnx support

[vinayakkunder97]

I have a Minkowski Engine trained model, when i try to convert this model into onnx format it gives the following error of sparse tensor unsupported.

Input shape - torch.Size([69, 3])

/home/aiml/miniconda3/envs/s2b_cpu/lib/python3.7/site-packages/torch/onnx/utils.py:355: UserWarning: Skipping _decide_input_format 'SparseTensor' object is not subscriptable

File "/home/aiml/miniconda3/envs/s2b_cpu/lib/python3.7/site-packages/torch/onnx/__init__.py", line 276, in export custom_opsets, enable_onnx_checker, use_external_data_format) File "/home/aiml/miniconda3/envs/s2b_cpu/lib/python3.7/site-packages/torch/onnx/utils.py", line 94, in export use_external_data_format=use_external_data_format) File "/home/aiml/miniconda3/envs/s2b_cpu/lib/python3.7/site-packages/torch/onnx/utils.py", line 701, in _export dynamic_axes=dynamic_axes) File "/home/aiml/miniconda3/envs/s2b_cpu/lib/python3.7/site-packages/torch/onnx/utils.py", line 459, in _model_to_graph use_new_jit_passes) File "/home/aiml/miniconda3/envs/s2b_cpu/lib/python3.7/site-packages/torch/onnx/utils.py", line 420, in _create_jit_graph graph, torch_out = _trace_and_get_graph_from_model(model, args) File "/home/aiml/miniconda3/envs/s2b_cpu/lib/python3.7/site-packages/torch/onnx/utils.py", line 380, in _trace_and_get_graph_from_model torch.jit._get_trace_graph(model, args, strict=False, _force_outplace=False, _return_inputs_states=True) File "/home/aiml/miniconda3/envs/s2b_cpu/lib/python3.7/site-packages/torch/jit/_trace.py", line 1139, in _get_trace_graph outs = ONNXTracedModule(f, strict, _force_outplace, return_inputs, _return_inputs_states)(*args, **kwargs) File "/home/aiml/miniconda3/envs/s2b_cpu/lib/python3.7/site-packages/torch/nn/modules/module.py", line 889, in _call_impl result = self.forward(*input, **kwargs) File "/home/aiml/miniconda3/envs/s2b_cpu/lib/python3.7/site-packages/torch/jit/_trace.py", line 93, in forward in_vars, in_desc = _flatten(args) RuntimeError: Only tuples, lists and Variables are supported as JIT inputs/outputs. Dictionaries and strings are also accepted, but their usage is not recommended. Here, received an input of unsupported type: SparseTensor

Can anyone suggest any solution for this above issue ?

[cama36]

Have you solved it?Minkowski to onnx problem.