From 48da34a886c4850eb0b8bd5d126520adbd4c1266 Mon Sep 17 00:00:00 2001
From: zhouzj <41366441+zzjjay@users.noreply.github.com>
Date: Thu, 12 Aug 2021 13:11:47 +0800
Subject: [PATCH] add description about the batch_size (#870)

---
 demo/dygraph/quant/README.md | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/demo/dygraph/quant/README.md b/demo/dygraph/quant/README.md
index 17693014..05c9f804 100755
--- a/demo/dygraph/quant/README.md
+++ b/demo/dygraph/quant/README.md
@@ -93,3 +93,5 @@ quanter.save_quantized_model(net, 'save_dir', input_spec=[paddle.static.InputSpe
 | ----------- | --------------------------- | ------------ | --------------------------- |
 | MobileNetV1 | 70.99/89.65                 | 普通在线量化 | 70.63/89.65                 |
 | MobileNetV3 | 78.96/94.48                 | PACT在线量化 | 77.52/93.77                 |
+
+注：在batch_size=256时能达到上表中理想的量化精度，但显存需求较大，因此在train.py中batch_size设置为128。
-- 
GitLab