I want to finetune the bison-32k model for sequential conversation for a fitness session. How do I format my dataset which has the entire fitness session as 1 example. Each conversation has multiple dialogues back and forth. And I have multiple such conversations to fine the model for. How can I add sequential dialogues within each example and multiple such examples.
In the above case I can only add "InputOutputTextPair"s but how do I feed multiple conversation examples. Because each conversation has within itself multiple dialogues.
User | Count |
---|---|
13 | |
2 | |
1 | |
1 | |
1 |