Docs add var len obs to readme and overview #4972
Conversation
| @@ -26,17 +26,14 @@ developer communities. | |||
|
|
|||
| ## Features | |||
|
|
|||
| - 15+ [example Unity environments](docs/Learning-Environment-Examples.md) | |||
| - Support for multiple environment configurations and training scenarios | |||
There was a problem hiding this comment.
Because I did not like them anymore :)
I can put them back
| @@ -684,6 +684,23 @@ three network architectures: | |||
| The choice of the architecture depends on the visual complexity of the scene and | |||
| the available computational resources. | |||
|
|
|||
| ### Learning from Variable Length Observations using Attention | |||
|
|
|||
| Using the ML-Agents Toolkit, it is possible to have agents learn from a | |||
There was a problem hiding this comment.
I think this section should start with an example of where you may have different inputs to help a non-RL expert understand when and why they want to use it. (See subsection below for instance.)
| about variable length observations and the BufferSensor | ||
| [here](Learning-Environment-Design-Agents.md#variable-length-observations). | ||
| When variable length observations are utilized, the ML-Agents Toolkit | ||
| leverages attention networks to learn from a varying number of entities. |
There was a problem hiding this comment.
add a reference to attention networks.
There was a problem hiding this comment.
There is a reference already in the section Learning-Environment-Design-Agents.md#variable-length-observations I think we should minimize the number of places with external references and avoid duplicate references.
| @@ -684,6 +684,23 @@ three network architectures: | |||
| The choice of the architecture depends on the visual complexity of the scene and | |||
| the available computational resources. | |||
|
|
|||
| ### Learning from Variable Length Observations using Attention | |||
There was a problem hiding this comment.
- need to update the Toc above with a link to this section.
- any visuals we can add?
There was a problem hiding this comment.
I can try to make another comic book
| @@ -684,6 +684,23 @@ three network architectures: | |||
| The choice of the architecture depends on the visual complexity of the scene and | |||
| the available computational resources. | |||
|
|
|||
| ### Learning from Variable Length Observations using Attention | |||
| about variable length observations and the BufferSensor | ||
| [here](Learning-Environment-Design-Agents.md#variable-length-observations). | ||
| When variable length observations are utilized, the ML-Agents Toolkit | ||
| leverages attention networks to learn from a varying number of entities. |
Proposed change(s)
This PR adds a section to the ML-Agents overview describing variable length observation and removes some old bullets from the list of feature in the main ML-Agents readme.
Useful links (Github issues, JIRA tickets, ML-Agents forum threads etc.)
This slack message
This JIRA ticket
Types of change(s)