Dirk
Dirk
Hi @AmoabaKelvin, thanks a lot for your suggestion! Looks like a great approach. One note is that the temperature=0.9 might be a bit too high and will not produce consistent...
Note: I actually think your approach could work, but might need some adjusting. Currently your approach fetches documents based on a similarity score, I think potentially instead of a simple...
Sounds great! Thanks for your contributions, Kelvin! Spending the next few days finishing the implementation with the HTML pre-processing approach. Once I am done, I will implement your approach. I...
Plus 1 to database other than (default)