I am looking for a similar feature as in model.generate() which takes a parameter num_return_sequences. It decides how many generations should be returned for each sample. It is especially useful when using beam search and analyzing the effect of beam search on the metrics. Trainer.predict() does n…

Can trainer.predict() return multiple generations for each sample?

berkayberabi January 18, 2022, 9:39pm 3

Hi Silvia,

No unfortunately, I could not solve it

Topic		Replies	Views
Generating multiple sequences with `Trainer.predict()` Beginners	0	344	February 17, 2023
Difference in trainer.predict() and model.generate() for LM 🤗Transformers	0	1803	July 5, 2023
Model.generate() is extremely slow while using beam search 🤗Transformers	2	5431	July 24, 2022
[Urgent] trainer.predict() and model.generate creates totally different predictions 🤗Transformers	4	6926	February 1, 2021
Batch_decode does not give the correct output as generate 🤗Transformers	0	301	March 17, 2022