is_causal fix for evals?

src/axolotl/monkeypatch/llama_attn_hijack_flash.py

@@ -155,6 +155,8 @@ def flashattn_forward(
         # during training q,k,v always have same seqlen
         assert key_states.shape == query_states.shape
         is_causal = True
+    elif past_key_value is None:
+        is_causal = True
     else:
         # turn off FA causal mask after first inference autoregressive iteration
         # only on first autoregressive step q,k,v have same seqlen