Open pgsrv opened 3 years ago
Hi! According to the results in 'Appendix: Full results' (https://www.fast.ai/2018/07/02/adam-weight-decay/) is it correct that Adam is better for fine-tuning, while AdamW is better for training from scratch? Thanks!
Hi! According to the results in 'Appendix: Full results' (https://www.fast.ai/2018/07/02/adam-weight-decay/) is it correct that Adam is better for fine-tuning, while AdamW is better for training from scratch? Thanks!