From 1b59b57b57010e6119282f3dbf37f8c7c6d6313e Mon Sep 17 00:00:00 2001 From: thomwolf Date: Wed, 8 Jan 2020 09:52:10 +0100 Subject: [PATCH] ignore_index equal -100 in T5 model --- src/transformers/modeling_t5.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/transformers/modeling_t5.py b/src/transformers/modeling_t5.py index 0f9bd679fc9..81906e86eab 100644 --- a/src/transformers/modeling_t5.py +++ b/src/transformers/modeling_t5.py @@ -905,7 +905,7 @@ class T5WithLMHeadModel(T5PreTrainedModel): if lm_labels is not None: shift_logits = lm_logits[..., :-1, :].contiguous() shift_labels = lm_labels[..., 1:].contiguous() - loss_fct = CrossEntropyLoss(ignore_index=-1) + loss_fct = CrossEntropyLoss(ignore_index=-100) loss = loss_fct(shift_logits.view(-1, shift_logits.size(-1)), shift_labels.view(-1)) decoder_outputs = ( loss,