Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Training Iteration Efficiency on Qwen3-14B base (train)
Loading...
150.1
Per-step Latency (ms)
FORGE
107.44
395.395
683.35
971.305
Jun 22, 2026
Per-step Latency (ms)
Peak GPU Memory (GiB)
Updated 1mo ago
Evaluation Results
Method
Method
Links
Per-step Latency (ms)
Peak GPU Memory (GiB)
FORGE
Hardware=H200, BS=1, S...
2026.06
150.1
63.2
GaLore r=128
Hardware=H200, BS=1, S...
2026.06
247.2
65.9
FlashOptim
Hardware=H200, BS=1, S...
2026.06
257.8
76.6
APOLLO-Mini r=256
Hardware=H200, BS=1, S...
2026.06
315
67.6
optimi.gradient_release
Hardware=H200, BS=1, S...
2026.06
343.6
117.3
bnb 8-bit AdamW
Hardware=H200, BS=1, S...
2026.06
564.1
83.7
AdaLomo
Hardware=H200, BS=1, S...
2026.06
1,216.6
40.7
Feedback
Search any
task
Search any
task