From 5a82297a4c8b97e8801270ffa5086417847f33f8 Mon Sep 17 00:00:00 2001 From: Patrick von Platen Date: Tue, 10 Oct 2023 21:28:00 +0000 Subject: [PATCH] improve bart --- src/transformers/models/bart/modeling_bart.py | 1 - 1 file changed, 1 deletion(-) diff --git a/src/transformers/models/bart/modeling_bart.py b/src/transformers/models/bart/modeling_bart.py index e3649357832ec1..dd4223a47f4735 100755 --- a/src/transformers/models/bart/modeling_bart.py +++ b/src/transformers/models/bart/modeling_bart.py @@ -420,7 +420,6 @@ def forward( return attn_output, attn_weights, past_key_value - # Copied from transformers.models.llama.modeling_llama.LlamaFlashAttention2._flash_attention_forward def _flash_attention_forward( self, query_states, key_states, value_states, padding_mask, query_length, causal=True, dropout=0.0, softmax_scale=None ):