Hello, and thank you for your excellent work! I really enjoyed reading your paper, and it inspired a couple of questions:
-
Could you share what the Hammer 1.0 series models were initialized from? I imagine it’s unlikely that they were trained entirely from scratch.
-
Do you have any plans or possibilities to release technical papers covering Hammer 2.0 and 2.1?
I truly appreciate your time and effort, and I look forward to learning more about your work!
Hello, and thank you for your excellent work! I really enjoyed reading your paper, and it inspired a couple of questions:
Could you share what the Hammer 1.0 series models were initialized from? I imagine it’s unlikely that they were trained entirely from scratch.
Do you have any plans or possibilities to release technical papers covering Hammer 2.0 and 2.1?
I truly appreciate your time and effort, and I look forward to learning more about your work!