Task-Centric Memory Code Samples Guide
_(EXPERIMENTAL, RESEARCH IN PROGRESS)_ <p align="right"> <img src="../../packages/autogen-ext/imgs/task_centric_memory.png" alt="Description" width="300" align="right" style="margin-left: 10px;"> </p> This directory contains code samples that illustrate the following forms of fast, memory-based learning: * Direct memory storage and retrieval * Learning from user advice and corrections * Learning from user demonstrations * Learning from the agent's own experience Each sample connects task-centric memory to a selectable agent with no changes to that agent's code. See the block diagram to the right for an overview of the components and their interactions. Each sample is contained in a separate python script, using data and configs stored in yaml files for easy modification. Note that since agent behavior is non-deterministic, results will vary between runs. To watch operations live in a browser and see how task-centric memory works, open the HTML page at the location specified at the top of the config file, such as: ./pagelogs/teachability/0 Call Tree.html To turn off logging entirely, set logging level to NONE in the config file. The config files specify an _AssistantAgent_ by default, which uses a fixed, multi-step system prompt. To use _MagenticOneGroupChat_ instead, specify that in the yaml file where indicated.
When to use Task-Centric Memory Code Samples
_(EXPERIMENTAL, RESEARCH IN PROGRESS)_ <p align="right"> <img src="../../packages/autogen-ext/imgs/task_centric_memory.png" alt="Description" width="300" align="right" style="margin-left: 10px;"> </p> This directory contains code samples that illustrate the following forms of fast, memory-based learning: * Direct memory storage and retrieval * Learning from user advice and corrections * Learning from user demonstrations * Learning from the agent's own experience Each sample connects task-centric memory to a selectable agent with no changes to that agent's code. See the block diagram to the right for an overview of the components and their interactions. Each sample is contained in a separate python script, using data and configs stored in yaml files for easy modification. Note that since agent behavior is non-deterministic, results will vary between runs. To watch operations live in a browser and see how task-centric memory works, open the HTML page at the location specified at the top of the config file, such as: ./pagelogs/teachability/0 Call Tree.html To turn off logging entirely, set logging level to NONE in the config file. The config files specify an _AssistantAgent_ by default, which uses a fixed, multi-step system prompt. To use _MagenticOneGroupChat_ instead, specify that in the yaml file where indicated.
How to use Task-Centric Memory Code Samples
Task-Centric Memory Code Samples is a single agent agent built on the AutoGen framework. Set it up from the source repository, configure your model credentials, and invoke it for tasks that match its description. Review the safety profile below before running it against production data or systems.
Safety profile
Autonomy
Semi-autonomous
Sandbox-aware
Yes
Network access
Unspecified
Filesystem access
Unspecified