【Fix bug] w4afp8 的nblock固定为256，并且fa3的append attn 增加mask参数 (#3771)

* fix w4afp8

* 增加集中式配置

* codestyle

* fix fa3 append attn

fastdeploy/model_executor/layers/attention/flash_attn_backend.py

@@ -359,6 +359,7 @@ class FlashAttentionBackend(AttentionBackend):
             getattr(layer, "cache_v_zp", None),
             layer.linear_shift,
             layer.linear_smooth,
+            forward_meta.attn_mask_offsets,
             metadata.kv_signal_data_list[layer.layer_id],
             getattr(layer, "q_norm_weight", None),
             getattr(layer, "k_norm_weight", None),