This project is a demo achieving entity extraction in dialogues, which can be used in AI assistant by combining with other actions.
The LLM is using OpenAI API.
Using booking flight as example.
inspiring by ujhrkzy/llm-slot-filling, this project uses the ConversationChain and ConversationBufferMemory from Langchain.
SlotMemory is a module to do entity extraction, store slot values and check information completeness, which based on ConversationBufferMemory.
Prompt and slot key in the SlotMemory can be modified according to task in other various scenarios.
The current_slot box displays the value of each slot, provides a more intuitive way to observe the performance.