Output 0 is out of bounds

[justinchuby]

PyTorch ONNX Conversion Error Report

✅ Obtain model graph with `torch.export.export`
✅ Translate the graph into ONNX
❌ Run `onnx.checker` on the ONNX model
⚪ Execute the model with ONNX Runtime
⚪ Validate model output accuracy

Error message:

Traceback (most recent call last):

  File "/Users/justinc/Documents/GitHub/torch-onnx/src/torch_onnx/_core.py", line 911, in export
    onnx.checker.check_model(onnx_program.model_proto, full_check=True)

  File "/Users/justinc/Documents/GitHub/torch-onnx/venv/lib/python3.11/site-packages/onnx/checker.py", line 179, in check_model
    C.check_model(

onnx.onnx_cpp2py_export.shape_inference.InferenceError: [ShapeInferenceError] (op_type:aten_zeros, node name: node_aten_zeros_1): Output 0 is out of bounds.

Analysis

PyTorch ONNX Conversion Analysis

Model Information

The model has 2176 parameters and 0 buffers (non-trainable parameters). Number of parameters per dtype:

defaultdict(<class 'int'>, {torch.float32: 2176})

Number of buffers per dtype:

defaultdict(<class 'int'>, {})

Inputs:

• arg4_1: TensorMetadata(shape=torch.Size([10, 16, 16]), dtype=torch.float32, requires_grad=False, stride=(256, 16, 1), memory_format=torch.contiguous_format, is_quantized=False, qparams={})

Outputs:

• cat: TensorMetadata(shape=torch.Size([10, 16, 16]), dtype=torch.float32, requires_grad=False, stride=(256, 16, 1), memory_format=torch.contiguous_format, is_quantized=False, qparams={})
• stack: TensorMetadata(shape=torch.Size([1, 1, 16, 16]), dtype=torch.float32, requires_grad=False, stride=(256, 256, 16, 1), memory_format=torch.channels_last, is_quantized=False, qparams={})
• stack_1: TensorMetadata(shape=torch.Size([1, 1, 16, 16]), dtype=torch.float32, requires_grad=False, stride=(256, 256, 16, 1), memory_format=torch.channels_last, is_quantized=False, qparams={})

The FX graph has 224 nodes in total. Number of FX nodes per op:

• placeholder: 5
• call_function: 218
• output: 1

Of the call_function nodes, the counts of operators used are:

• <built-in function getitem>: 52
• aten.sigmoid.default: 30
• aten.mul.Tensor: 30
• aten.view.default: 22
• aten.add.Tensor: 20
• aten.tanh.default: 20
• aten.t.default: 11
• aten.addmm.default: 11
• aten.split.Tensor: 10
• aten.unbind.int: 3
• aten.zeros.default: 2
• aten.unsqueeze.default: 2
• aten.squeeze.dim: 2
• aten.stack.default: 2
• aten.cat.default: 1

ONNX Conversion Information

All operators in the model have registered ONNX decompositions.

[justinchuby]

This is because the torchlib functions do not have return types annotated correctly.