[QNN] Op Dequantize does not support per-channel quant tensor

[vybhav-ibr]

🐛 Describe the bug

When i try to compile my custom model for QNN htp. I get the error

Visiting: quantized_decomposed_dequantize_per_tensor_tensor_3, quantized_decomposed.dequantize_per_tensor.tensor
[ERROR] [Qnn ExecuTorch]:  <E> Op Dequantize does not support per-channel quant tensor

[ERROR] [Qnn ExecuTorch]:  <E> Failed to construct common node for graph 256

[ERROR] [Qnn ExecuTorch]:  <E> Failed to add node with err 1000

[ERROR] [Qnn ExecuTorch]: Failed to add node to Qnn Graph with error: 1000

This issue persist even if i add this op to the skip nodes is set like this

qnn_config.skip_delegate_node_ids.add("quantized_decomposed_dequantize_per_tensor_tensor_3")`

what exactly is causing this error ?
would i need to modify my model or the training recipe to sort this ? how so ?

My training recipe is based on the QNN deeplab v3 example

Versions

executorch: v1.3.0
QNN-SDK: 2.37