RUNTIME_EXCEPTION : Non-zero status code returned while running Reshape node. about onnx HOT 6 CLOSED

Hi, I don't understand the issue. The onnxruntime error message seems a valid error: you cannot reshape a zero-sized tensor into a tensor of size 1. So, it looks like an error in the model (or model-conversion).

from onnx.

@gramalingam I have edited the issue of input shape and requested shape.

When keeping the batch size as 32, am getting below failure.
onnxruntime.capi.onnxruntime_pybind11_state.RuntimeException: [ONNXRuntimeError] : 6 : RUNTIME_EXCEPTION : Non-zero status code returned while running Reshape node. Name:'/Reshape_16_output_0_ReduceMax_Reshape' Status Message: /onnxruntime_src/onnxruntime/core/providers/cpu/tensor/reshape_helper.h:44 onnxruntime::ReshapeHelper::ReshapeHelper(const onnxruntime::TensorShape&, onnxruntime::TensorShapeVector&, bool) input_shape_size == size was false. The input tensor cannot be reshaped to the requested shape. Input shape:{0,1}, requested shape:{1}
When keeping the batch size as 1, am getting below failure.
[ONNXRuntimeError] : 6 : RUNTIME_EXCEPTION : Non-zero status code returned while running Reshape node. Name:'/Tile_output_0_ReduceMax_Reshape' Status Message: /onnxruntime_src/onnxruntime/core/providers/cpu/tensor/reshape_helper.h:44 onnxruntime::ReshapeHelper::ReshapeHelper(const onnxruntime::TensorShape&, onnxruntime::TensorShapeVector&, bool) input_shape_size == size was false. The input tensor cannot be reshaped to the requested shape. Input shape:{1,0,1}, requested shape:{1}

from onnx.

@Priyanshu88 : You could ask in the onnxruntime repo (https://github.com/microsoft/onnxruntime/issues). It would help if you provide the onnx model and input to reproduce the error. I still think the error looks like a valid error message, so it could be an error in the model/exporter ... and maybe asking in mmlab would be more appropriate.

from onnx.

But this problem is coming up for the conversion of model from .pth to .onnx through the onnx_config i mentioned earlier. I am not yet on training or inference generation, i hope this to be the right place to ask.

from onnx.

For open-mmlab generated models, it may be a good idea to consult with developers of https://github.com/open-mmlab.

from onnx.

But this problem is coming up for the conversion of model from .pth to .onnx through the onnx_config i mentioned earlier

The onnx config you mention seems to be a feature of open-mmlab, not ONNX itself.

from onnx.