diff --git a/README.md b/README.md
index be7bf07d3..39239a1a2 100644
--- a/README.md
+++ b/README.md
@@ -392,7 +392,7 @@
         <td class="xl69" x:str>NLP</td>
     <td class="xl69" x:str>f32/f16</td>
     <td class="xl69" x:str>W32A16</td>
-    <td class="xl69" x:str>Incoming</td>
+    <td class="xl69" x:str>f16</td>
         <td class="xl69" x:str>N/A</td>
    </tr>
     <tr height="16.80" style='height:16.80pt;'>
diff --git a/training/iluvatar/t5_small-pytorch/README.md b/training/iluvatar/t5_small-pytorch/README.md
index 53d90ee58..a82bd8531 100644
--- a/training/iluvatar/t5_small-pytorch/README.md
+++ b/training/iluvatar/t5_small-pytorch/README.md
@@ -50,9 +50,3 @@
 | BI-V100单机单卡(1x1) | fp32      | /      | /   | /     | /     | /    | 41.2064  | 18.9082  | 29.1922  | 38.4298     | 27.378 /32.0 |
 | BI-V100两机8卡(2x8) | fp32      |       |         |         |         |        |       |       |        |        |            |
 
-注意: T5模型MFU数值较低, 为11.8%
-1x8训练的MFU计算过程如下:
-`MFU = 400.26068691305795 * 1024 * (60 * 10^6) * 6 / (156 * 1000^4) / 8 = 11.8%`
-
-其中, 1024为seq_len, 60 millions为参数量, (156 * 1000^4)为A100 tf32算力
-
diff --git a/training/run_benchmarks/config/test_conf.py b/training/run_benchmarks/config/test_conf.py
index e04f10854..633f0c78d 100644
--- a/training/run_benchmarks/config/test_conf.py
+++ b/training/run_benchmarks/config/test_conf.py
@@ -130,7 +130,6 @@
     # "transformer:pytorch:R300:1:8:1": "/raid/dataset/transformer/wmt14_en_de_joined_dict",
     # "bigtransfer:pytorch:R300:1:8:1": "/raid/dataset/ImageNet_1k_2012/",
     # "efficientnet:pytorch:R300:1:8:1": "/raid/dataset/ImageNet_1k_2012/",
-
   
     # iluvatar cases
     # "bigtransfer:pytorch:BI-V100:1:8:1": "/raid/dataset/ImageNet_1k_2012/",