Export to onnx

[petropetropetro]

Hi, I try to export your model to onnx.
Did you try to do something like this?
Now I`m stuck with "RuntimeError: Expected a sequence type, but received a non-iterable type in graph output index 0". Is this because flow_preds is a dic, or can it be related to something else?

[petropetropetro]

Added more context
torch.onnx.export(model_without_ddp, # model being run
args=(right, left, 'self_swin2d_cross_swin1d',
[2, 8], [-1, 4],
[-1, 1], 3,
False, 'stereo', None, None,
1. / 0.5, 1. / 10, 64,
False, False), # model input (or a tuple for multiple inputs)
f="super_resolution.onnx", # where to save the model (can be a file or file-like object)
verbose=True,
export_params=True, # store the trained parameter weights inside the model file
opset_version=16, # the ONNX version to export the model to
do_constant_folding=True, # whether to execute constant folding for optimization
input_names = ['img0', 'img1', 'attn_type',
'attn_splits_list', 'corr_radius_list',
'prop_radius_list', 'num_reg_refine',
'pred_bidir_flow' , 'task', 'intrinsics', 'pose',
'min_depth', 'max_depth', 'num_depth_candidates',
'depth_from_argmax', 'pred_bidir_depth'], # the model's input names
output_names = ['flow_preds'])

RuntimeError Traceback (most recent call last)
Cell In[21], line 1
----> 1 torch.onnx.export(model_without_ddp, # model being run
2 args=(right, left, 'self_swin2d_cross_swin1d',
3 [2, 8], [-1, 4],
4 [-1, 1], 3,
5 False, 'stereo', None, None,
6 1. / 0.5, 1. / 10, 64,
7 False, False), # model input (or a tuple for multiple inputs)
8 f="super_resolution.onnx", # where to save the model (can be a file or file-like object)
9 verbose=True,
10 export_params=True, # store the trained parameter weights inside the model file
11 opset_version=16, # the ONNX version to export the model to
12 do_constant_folding=True, # whether to execute constant folding for optimization
13 input_names = ['img0', 'img1', 'attn_type',
14 'attn_splits_list', 'corr_radius_list',
15 'prop_radius_list', 'num_reg_refine',
16 'pred_bidir_flow' , 'task', 'intrinsics', 'pose',
17 'min_depth', 'max_depth', 'num_depth_candidates',
18 'depth_from_argmax', 'pred_bidir_depth'], # the model's input names
19 output_names = ['flow_preds'])

File c:\ProgramData\Anaconda3\lib\site-packages\torch\onnx\utils.py:506, in export(model, args, f, export_params, verbose, training, input_names, output_names, operator_export_type, opset_version, do_constant_folding, dynamic_axes, keep_initializers_as_inputs, custom_opsets, export_modules_as_functions)
188 @_beartype.beartype
189 def export(
190 model: Union[torch.nn.Module, torch.jit.ScriptModule, torch.jit.ScriptFunction],
(...)
206 export_modules_as_functions: Union[bool, Collection[Type[torch.nn.Module]]] = False,
207 ) -> None:
208 r"""Exports a model into ONNX format.
209
210 If model is not a :class:torch.jit.ScriptModule nor a
(...)
503 All errors are subclasses of :class:errors.OnnxExporterError.
504 """
--> 506 _export(
507 model,
508 args,
509 f,
510 export_params,
511 verbose,
512 training,
513 input_names,
514 output_names,
515 operator_export_type=operator_export_type,
516 opset_version=opset_version,
517 do_constant_folding=do_constant_folding,
518 dynamic_axes=dynamic_axes,
519 keep_initializers_as_inputs=keep_initializers_as_inputs,
520 custom_opsets=custom_opsets,
521 export_modules_as_functions=export_modules_as_functions,
522 )

File c:\ProgramData\Anaconda3\lib\site-packages\torch\onnx\utils.py:1548, in _export(model, args, f, export_params, verbose, training, input_names, output_names, operator_export_type, export_type, opset_version, do_constant_folding, dynamic_axes, keep_initializers_as_inputs, fixed_batch_size, custom_opsets, add_node_names, onnx_shape_inference, export_modules_as_functions)
1545 dynamic_axes = {}
1546 _validate_dynamic_axes(dynamic_axes, model, input_names, output_names)
-> 1548 graph, params_dict, torch_out = _model_to_graph(
1549 model,
1550 args,
1551 verbose,
1552 input_names,
1553 output_names,
1554 operator_export_type,
1555 val_do_constant_folding,
1556 fixed_batch_size=fixed_batch_size,
1557 training=training,
1558 dynamic_axes=dynamic_axes,
1559 )
1561 # TODO: Don't allocate a in-memory string for the protobuf
1562 defer_weight_export = (
1563 export_type is not _exporter_states.ExportTypes.PROTOBUF_FILE
1564 )

File c:\ProgramData\Anaconda3\lib\site-packages\torch\onnx\utils.py:1160, in _model_to_graph(model, args, verbose, input_names, output_names, operator_export_type, do_constant_folding, _disable_torch_constant_prop, fixed_batch_size, training, dynamic_axes)
1156 # assign_output_shape pass is not compatible with quantized outputs.
1157 # Quantized outputs are flattened to 3 values in ONNX, while packed as
1158 # single value in PyTorch.
1159 if not any(getattr(out, "is_quantized", False) for out in output_tensors):
-> 1160 _C._jit_pass_onnx_assign_output_shape(
1161 graph,
1162 output_tensors,
1163 out_desc,
1164 GLOBALS.onnx_shape_inference,
1165 is_script,
1166 GLOBALS.export_onnx_opset_version,
1167 )
1169 _set_input_and_output_names(graph, input_names, output_names)
1170 params_dict = _get_named_param_dict(graph, params)

RuntimeError: Expected a sequence type, but received a non-iterable type in graph output index 0

[juandavid212]

Do you have any updates about this?

[petropetropetro]

@juandavid212 No, I had found the https://github.com/fateshelled/unimatch_onnx, It was okay for me as PoV design