Update VisualDL documentation structure on PPO #416
Conversation
| @@ -1,4 +1,4 @@ | |||
| # VisualDL Backend / SDK structure and Develop Guide | |||
| # Backend Architecture | |||
There was a problem hiding this comment.
Maybe just call Backend Guide? We have a Frontend Guide already. Also this doc talks about how to dev and architecture is a sub session.
There was a problem hiding this comment.
Backend Guide works for me too.
| @@ -0,0 +1,124 @@ | |||
| # 后台架构 | |||
There was a problem hiding this comment.
oh...i can translate this to chinese later
| @@ -1,4 +1,4 @@ | |||
| # VisualDL Web app Develop | |||
| # Frontend Guide | |||
There was a problem hiding this comment.
hmmm even its in How to develop section, but maybe change to Frontend Develop Guide/ Backend Develop Guide seems clear than just Frontend Guide / Backend Guide
There was a problem hiding this comment.
Have develop appears again is repetitive. I tried Backend Develop Guide and it doesn't look so good
| If you still have issues installing VisualDL, try installing from sources as in following section. | ||
|
|
||
|
|
||
| ### Install from source |
There was a problem hiding this comment.
I made a mistake before, it should be "##" instead of "###"
There was a problem hiding this comment.
I will update them.
| @@ -0,0 +1,100 @@ | |||
| # Installation | |||
|
|
|||
| ### How to install from pypi | |||
There was a problem hiding this comment.
Most issues are not coming from pip install, they show up when user launch visualdl, so it should include running the demo script and launch visualdl to verify installation is good.
I think "Install with native pip" would be a better title, coz either using Virtualenv or Anaconda or none of them, we are also still installing from pypi
I think just copied the "Quick Start" section from what we had and change title to "Install with native pip" should be good as following:
Install with native pip
# Install the VisualDL. Preferably under a virtual environment or anaconda.
pip install --upgrade visualdl
# run a demo, vdl_create_scratch_log will create logs for testing.
vdl_create_scratch_log
visualdl --logdir=scratch_log --port=8080
# visit http://127.0.0.1:8080
There was a problem hiding this comment.
There is no native pip. pip is the method.
I will add the vdl_create_scratch_log part. to help the users to test. Good suggestion.
|
|
||
| VisualDL has three components. | ||
| 1. The Python/C++ SDK that logs the data during training. | ||
| 1. The single page client side app that visualized training data. |
| VisualDL has three components. | ||
| 1. The Python/C++ SDK that logs the data during training. | ||
| 1. The single page client side app that visualized training data. | ||
| 1. The server (powered on Flask) that reads the logs data and delivers it to the client side app for displaying graphs (scalars/histograms) and embeddings. |
There was a problem hiding this comment.
powered by Flask, reads the log data
|
|
||
| ## Tests | ||
|
|
||
| Modification in SDK API and any major function changes should be adding test cases. |
There was a problem hiding this comment.
Modification in SDK API or any major function
| - defines server arguments and config | ||
| - provides API and router for front end | ||
| - ```lib.py``` : delegate wrapper for function calls | ||
| - read and caches log data |
| - ```Storage``` : main set of data with different modes such as "train" or "test", contain multiple ```Tablet``` | ||
| - ```Tablet``` : group a single type of data with ```tag```. One tablet will only represent one data type such as scalar, histogram, image and contains multiple ```Record```. For example one tablet represents a line in scalar chart. | ||
| - ```Record``` : represent any data structure for any component | ||
| - Handles the actual logic to serialized to disk and deserialized from disk |
There was a problem hiding this comment.
serialize & deserialize
| @@ -0,0 +1,78 @@ | |||
| # 前端指南 | |||
There was a problem hiding this comment.
I can translate this front-end intro into Chinese
|
|
||
| 如果以上步骤出现问题，很可能是因为python或pip不同版本或不同位置所致，以下安装方法能解决。 | ||
|
|
||
| ## 使用 virtualenv 安装 |
There was a problem hiding this comment.
Should we move away from 'virtualenv' towards Anaconda's 'conda create' ? Because we tried 'virtualenv' but got some issues, right?
There was a problem hiding this comment.
That's valid point. In general we need to understand what exactly causes the issue.
I don't think this PR is suitable to address this. Let's sit down together and research more on this topic.
This one completely updates the documentation structure on the PPO.
We now change the structure to the following.