Did your method include any automated evaluation?
Thanks for publishing Jackrong.
How did you evaluate results (if at all)? These highly compact models can be very sensitive to perturbance.
I can inform you of an aspect of the English language that is relevant:
Definitions:
Method: How you did it; The process you used
Methodology: How you selected the method to do it.
It's a regrettably widespread error to use 'methodology' when 'method' is correct.
The reason is that those unaware of the correct usage always choose 'methodology' because it looks more sciencey and impressive. Sadly the misuse of the term is a badge of ignorance in the scientific field.
Cheers,
Thank you for your suggestions. The current model uses an automated Skills‑Agent pipeline for hyperparameter selection, fine‑tuning, and post‑training model‑card generation and publishing. I wasn’t able to review everything thoroughly, and due to limited resources I also haven’t conducted full evaluations. This work is intended purely for education and demonstration purposes. As a student, my resources are a bit limited. I also apologize for any misunderstanding caused by my misuse of English terminology, and I appreciate your support and understanding.