|
01_baseline_r16_a16_lr5e5_adamw
|
changes
|
2026-09-30 21:22:53 +02:00 |
|
02_optimizer_adamw8bit_r16_a16_lr5e5
|
changes
|
2026-09-30 21:22:53 +02:00 |
|
03_rank8_a16_lr5e5
|
changes
|
2026-09-30 21:22:53 +02:00 |
|
04_rank32_a32_lr5e5
|
changes
|
2026-09-30 21:22:53 +02:00 |
|
05_r16_alpha32_lr5e5
|
changes
|
2026-09-30 21:22:53 +02:00 |
|
06_lr1e4_r16_a16
|
changes
|
2026-09-30 21:22:53 +02:00 |
|
07_lr2e5_r16_a16
|
changes
|
2026-09-30 21:22:53 +02:00 |
|
08_attention_only_r16_a16_lr5e5
|
changes
|
2026-09-30 21:22:53 +02:00 |
|
09_batch2_accum4_r16_a16_lr5e5
|
changes
|
2026-09-30 21:22:53 +02:00 |
|
10_r16_a32_lr1e4
|
changes
|
2026-09-30 21:22:53 +02:00 |