From 9a5ebef14849371216c68e03e53730c93b8bb353 Mon Sep 17 00:00:00 2001 From: Daniel Han-Chen Date: Sun, 4 Feb 2024 03:23:26 +1100 Subject: [PATCH] Update llama.py --- unsloth/models/llama.py | 1 + 1 file changed, 1 insertion(+) diff --git a/unsloth/models/llama.py b/unsloth/models/llama.py index 40e5e56e19..56d43a51a8 100644 --- a/unsloth/models/llama.py +++ b/unsloth/models/llama.py @@ -309,6 +309,7 @@ def LlamaAttention_fast_forward( V = V.transpose(1, 2) A = flash_attn_func(Q, K, V, causal = True) else: + print(2) # Grouped query attention if n_groups != 1: K = K[:, :, None, :, :].expand(bsz, n_kv_heads, n_groups, kv_seq_len, head_dim)