Update llama.py

This commit is contained in:
Daniel Han-Chen 2024-02-03 23:22:21 +11:00
commit 257cd7d531

View file

@ -680,7 +680,7 @@ def LlamaModel_fast_forward_inference(
decoder_layer.self_attn,
hidden_states,
past_key_values[idx],
dp_prefill = False,
do_prefill = False,
temp_QA = temp_QA,
temp_KV = temp_KV,
RH_Q = RH_Q,