what is ''wget https://bapengstorage.blob.core.windows.net/fileshare/scgpt.tar.gz''
Thanks for your contribution and works.
May I ask you what is ''wget https://bapengstorage.blob.core.windows.net/fileshare/scgpt.tar.gz''?
what dataset were used for this pre-trained scgpt model?
Thanks in advance!
Thanks for your contribution and works.
May I ask you what is ''wget https://bapengstorage.blob.core.windows.net/fileshare/scgpt.tar.gz''?
what dataset were used for this pre-trained scgpt model?
Thanks in advance!
It is the pre-trained "scgpt" model download link that will allow you to further fine-tune for your domain data (which you won't need a lot).
Thanks for your contribution and works. May I ask you what is ''wget https://bapengstorage.blob.core.windows.net/fileshare/scgpt.tar.gz''? what dataset were used for this pre-trained scgpt model? Thanks in advance!
It is the pre-trained "scgpt" model download link that will allow you to further fine-tune for your domain data (which you won't need a lot).
Thanks for your information. But, what my concern is what dataset have been used for this scgpt model? MultiWOZ or DSTC8 or other task-oriented dialogue dataset?
From the response in below link it appears the available pre-trained model is not fine-tuned on MULTIWOZ, so must be on DSTC and facebook data (I don't know which one is this). https://github.com/pengbaolin/SC-GPT/issues/3#issuecomment-620904479