Update llama.py

This commit is contained in:
Daniel Han-Chen 2024-02-02 20:35:00 +11:00
commit 1b274212e1

View file

@ -699,7 +699,7 @@ def LlamaModel_fast_forward_inference(
decoder_layer.self_attn,
hidden_states,
past_key_values[idx],
position_ids,
None,
)
hidden_states += residual