ValueError: too many values to unpack (expected 2) in step 3 of Customize task-specific AGBT-FPs

[luk47]

Hi, after I run train.py in step 3 of Customize task-specific AGBT-FPs, I get ValueError: too many values to unpack (expected 2). since it wraps part of fairseq, but there is no mention of what version, I would like to ask how to solve this problem. I am using python version 3.8 ; pytorch version torch-1.7.1%2Bcu110-cp38-cp38-win_amd64.whl and cuda version cuda_11.0.2_451.48_win10. The attached image is a screenshot of the run.

[ChenDdon]

Hi, I repeated all steps and didn't produce your error. I am using python==3.7.10 and pytorch==1.4.0. As for the fairseq, I modified the source code in 2019, but I forgot the exact date. I hope this information can help you.