Hi @Z3K3, let’s move this discussion to Optimize an ONNX Seq2Seq model as you are describing the same problem there. Please don’t cross-post the same question in multiple topics in the future as it makes things difficult to track.
1 Like