ONNX version Silero VAD model fails consistently with tensor shape issues feeding LSTM node

[chipchump]

have built both 1.15.2 and 1.21.1 versions of onnxruntime for arm64-v8a (android) but consistently receive this error using the official snakers4 onnx model (and after conversion to ORT format) with either onnxruntime build. Anybody seen this error?

ONNX inference failed: Non-zero status code returned while running If node. Name:'If_0' Status Message: Non-zero status code returned while running If node. Name:'If_0_then_branch__Inline_0__/decoder/If_1' Status Message: Non-zero status code returned while running LSTM node. Name:'If_0_then_branch__Inline_0__/decoder/rnn/LSTM' Status Message: Input X must have 3 dimensions only. Actual:{1,1,1,128,2}
ONNX shape error detected - please check model input requirements
LSTM requires exactly 3 dimensions for input tensor. Please reshape to [batch_size, sequence_length, input_size]

I dump out the model input/outputs and believe I am using the correct input shapes:

Model expects 3 inputs:
Input 0: input
Shape: [-1, -1] (rank: 2)
Input 1: state
Shape: [2, -1, 128] (rank: 3)
Input 2: sr
Shape: [] (rank: 0)
Model provides 2 outputs:
Output 0: output
Shape: [-1, 1] (rank: 2)
Output 1: stateN
Shape: [-1, -1, -1] (rank: 3)
Resetting VAD state
State reset: size=256, shape=[2, 1, 128]
State tensor will have shape: [2, 1, 128]
State vector data address: 0xb4000070edbb10c0, capacity: 256

Resetting VAD...
Resetting VAD state
State reset: size=256, shape=[2, 1, 128]
State tensor will have shape: [2, 1, 128]
State vector data address: 0xb4000070edbb10c0, capacity: 256
Processing optimal chunk of size: 1024
Creating input tensor with shape: [1, 1024]
Created input tensor with actual shape: [1, 1024]
Creating state tensor with shape: [2, 1, 128]
Creating sample rate tensor with shape: [1], value: 16000
Created state tensor with shape: [2, 1, 128]
Created sample rate tensor with shape: [1]
Running ONNX inference...
ONNX inference failed: Non-zero status code returned while running If node. Name:'If_0' Status Message: Non-zero status code returned while running If node. Name:'If_0_then_branch__Inline_0__/decoder/If_1' Status Message: Non-zero status code returned while running LSTM node. Name:'If_0_then_branch__Inline_0__/decoder/rnn/LSTM' Status Message: Input X must have 3 dimensions only. Actual:{1,1,1,128,2}
ONNX shape error detected - please check model input requirements
LSTM requires exactly 3 dimensions for input tensor. Please reshape to [batch_size, sequence_length, input_size]

[chipchump]

Never mind - I was stupid - PS, follow the silero VAD examples in the snaker4 repo