From ec3a700c632dba97c01529c0049c383d68636aa8 Mon Sep 17 00:00:00 2001
From: felix-wang <35718120+numb3r3@users.noreply.github.com>
Date: Mon, 18 Apr 2022 15:15:21 +0800
Subject: [PATCH] chore: update docs (#684)

* chore: update docs

* chore: update changelog
---
 docs/changelog/index.md    | 12 ++++++++++++
 docs/user-guides/server.md |  6 +++---
 2 files changed, 15 insertions(+), 3 deletions(-)

diff --git a/docs/changelog/index.md b/docs/changelog/index.md
index 32520555a..c35d18b81 100644
--- a/docs/changelog/index.md
+++ b/docs/changelog/index.md
@@ -4,3 +4,15 @@ CLIP-as-service follows semantic versioning. However, before the project reach 1
 
 This chapter only tracks the most important breaking changes and explain the rationale behind them.
 
+## 0.2.0: improve the service scalability with replicas
+
+This change is mainly intended to improve the inference performance with replicas.
+
+Here is the short benchmark summary of the improvement (`replicas=4`):
+
+| batch_size  | before | after   |
+|-------------|--------|---------|
+| 1           | 23.74  | 18.89   |
+| 8           | 58.88  | 30.38   |
+| 16          | 14.96  | 91.86   |
+| 32          | 14.78  | 101.75  |
diff --git a/docs/user-guides/server.md b/docs/user-guides/server.md
index 2b8c9e8a6..49220f346 100644
--- a/docs/user-guides/server.md
+++ b/docs/user-guides/server.md
@@ -184,9 +184,9 @@ There are also runtime-specific parameters listed below:
 
 ````{tab} ONNX
 
-| Parameter | Description                                                                                       |
-|-----------|---------------------------------------------------------------------------------------------------|
-| `providers`  | [ONNX runtime provides](https://onnxruntime.ai/docs/execution-providers/), default is auto-detect |
+| Parameter | Description                                                                                                                    |
+|-----------|--------------------------------------------------------------------------------------------------------------------------------|
+| `device`  | `cuda` or `cpu`. Default is `None` means auto-detect.
 
 ````