Qwen2 Possibility

[dikyridhlo]

i have tried to train qwen with my own dataset. i have tested it and working properly. i did the training with "instruction-response" template.
Next, i want to use qwen to create chatbot engine like chatgpt.
now i have 100 page paragraph of knowledge and will train it to qwen.

my question is

1. is posible to deal it with qwen?
2. is the training still use "instruction-response" template? if yes, should i define all posible question and answer for all 100 paragraph? is qwen can be trained with context like gpt ?