Summarization and Classificaiton fine-tune

[savasy]

Hi @stefan-it , is there any training code in transformer repo or else where such as run_glue.py or run_ner.py to fine-tune Turkish-bert model for the summarization and classificaiton problem ?

[stefan-it]

Hi @savasy ,

for text summarization you would need a sequence-to-sequence model architecture, e.g. like the recently introduced BART model: https://arxiv.org/abs/1910.13461

For classification you could modify the run_glue.py script. However, there are simpler ways, like using FARM or simpletransformers :)

[savasy]

Thank you @stefan-it for quick reply.