S
Soham
@sohampal
Correct flow
preprocess → pretrain → SFT → DPO → evaluate. (2 votes)
preprocess → SFT → pretrain→ DPO → evaluate. (0 votes)
pretrain→ preprocess→ SFT → DPO → evaluate. (1 votes)
pretrain → preprocess → DPO → SFT → evaluate. (0 votes)
❤️ 3 likes
💬 0 comments
Replies
No replies yet.