Chong Chen
Results
5
issues of
Chong Chen
E.g., which dataset it is (Movielens, EachMovie, or Netflix?), and how to split it into training set and test set as in your sigir paper?
Excuse me, I am really interested in your work, and when could you kindly release your code? Thanks a lot!
Do you have plans to support token_type_ids? This is important for QA and search model inference.
Great work, I have some questions below: class RetroMAECollator(DataCollatorForWholeWordMask): max_seq_length: int = 512 encoder_mlm_probability: float = 0.15 decoder_mlm_probability: float = 0.15 def __call__(self, examples): input_ids_batch = [] attention_mask_batch = []...
It seems that there is only loss_ab in the released code, right?