Error exporting T5 model to ONNX with optimum-cli

🤗Optimum
1

I have tried exporting the google-t5/t5-small model to ONNX format using optimum.

I have tried exporting this model 2 ways.

The first was I used the ORTModelForSeq2SeqLM class as defined in optimum and then used save_pretrained method. This seemed to work swimmingly but only outputs:

  1. encoder_model.onnx
  2. decoder_model.onnx
  3. decoder_with_past_model.onnx

And I need the decoder_model_merged.onnx. I then tried with the optimum-cli using:

optimum-cli export onnx --model google-t5/t5-small onnx_model

This throws the error:

Exception: An error occured during validation, but the model was saved nonetheless at onnx_model. Detailed error: [ONNXRuntimeError] : 1 : FAIL : Load model from onnx_model/decoder_model_merged.onnx failed:/Users/runner/work/1/s/onnxruntime/core/graph/model.cc:180 onnxruntime::Model::Model(ModelProto &&, const PathString &, const IOnnxRuntimeOpSchemaRegistryList *, const logging::Logger &, const ModelOptions &) Unsupported model IR version: 10, max supported IR version: 9

My package versions are:

optimum==1.18.0
onnxruntime==1.17.3
onnx==1.16.0
transformers==4.39.3

Anyone have any ideas?

2

@JamesXanda Can you update Optimum with pip install -U optimum and try again please?

optimum-cli export onnx --model google-t5/t5-small onnx_model

runs well on my laptop with latest stable release of Optimum.

3

This worked, thanks.

4

This topic was automatically closed 12 hours after the last reply. New replies are no longer allowed.