Closed takase closed 3 years ago
I found that eos_indices of a dictionary and training data are different from each other. Do I need to modify bert_dictionary.py?
I'm so sorry, I ignored the task option during pre-processing. I attached the option, and then solved the above problem.
Hi, thank you for distributing your code! I tried to fine-tune the pre-trained ProphetNet (160G) on English Gigaword summarization dataset. I conducted pre-processing described in README and then tried fine-tuning but faced the following Assertion Error:
pytorch version == 1.4.0
fairseq version == 0.9.0
In addition, when I tried to train the original Transformer (--arch transformer_wmt_en_de) with label_smoothed_cross_entropy, I succeeded training.
Do you have any idea to solve the above error?