when I use fastt5 library to convert t5-base fine-tune models like "pszemraj/t5-v1_1-base-ft-jflAUG", "pszemraj/grammar-synthesis-base", fastt5 library Image of error show tuple index out of range error in this file "/usr/local/lib/python3.7/dist-packages/transformers/utils/generic.py". The second error is coming that "encoder = model.encoder" is not defined and I was working in colab notebook, So how can we solve these issues and convert these t5-base-grammar fine-tuned model as I mentioned earlier one model is -> pszemraj/t5-v1_1-base-ft-jflAUG, ... to onnx and quantize them.
