arxiv:2502.17055
Haotian Hu
cspikachu
ยท
AI & ML interests
Efficient training for LLMs
Recent Activity
authored
a paper
about 17 hours ago
Stable-SPAM: How to Train in 4-Bit More Stably than 16-Bit Adam
upvoted
a
paper
1 day ago
Stable-SPAM: How to Train in 4-Bit More Stably than 16-Bit Adam
upvoted
a
paper
3 months ago
APOLLO: SGD-like Memory, AdamW-level Performance
Organizations
None yet
Papers
1
models
None public yet
datasets
None public yet