Closed Ber666 closed 3 days ago
It seems each sample in the deita dataset consists of a lot of turns and is super long (>10k tokens). Your paper mentioned the max length of input is 2048 for SFT. Does that mean most text of each training sample is truncated and discarded?
Hi. Yes, for those samples whose length are longer than model_max_length, we simply truncate the length of those samples to 2048.
It seems each sample in the deita dataset consists of a lot of turns and is super long (>10k tokens). Your paper mentioned the max length of input is 2048 for SFT. Does that mean most text of each training sample is truncated and discarded?