Does the inference computed at once?

[idan-tankel]

SemOOD/SEED-Bench/lavis/models/blip2_models/modeling_t5.py

Lines 1845 to 1858 in 1377966

	decoder_outputs = self.decoder(
	input_ids=decoder_input_ids,
	attention_mask=decoder_attention_mask,
	inputs_embeds=decoder_inputs_embeds,
	past_key_values=past_key_values,
	encoder_hidden_states=hidden_states,
	encoder_attention_mask=attention_mask,
	head_mask=decoder_head_mask,
	cross_attn_head_mask=cross_attn_head_mask,
	use_cache=use_cache,
	output_attentions=output_attentions,
	output_hidden_states=output_hidden_states,
	return_dict=return_dict,
	)

[idan-tankel]

right_shift is being done