model_20m_16B card
Browse files- model_20m_16B/README.md +0 -4
model_20m_16B/README.md
CHANGED
|
@@ -12,12 +12,8 @@ The pretraining base and SFT init for this model are
|
|
| 12 |
|
| 13 |
| | |
|
| 14 |
|---|---|
|
| 15 |
-
| run id | `C6p5e18_20m_alpha0.300_beta0.008` |
|
| 16 |
| parameters | 20m |
|
| 17 |
| pretraining tokens | 15,824,042,105 (15.8B) |
|
| 18 |
-
| compute class | 6p5e18 |
|
| 19 |
-
| alpha (pretrain fraction) | 0.3 |
|
| 20 |
-
| beta (SFT fraction) | 0.008 |
|
| 21 |
| checkpoints here | 30 (steps 100–3000) |
|
| 22 |
| checkpoints saved by the run | 60 |
|
| 23 |
|
|
|
|
| 12 |
|
| 13 |
| | |
|
| 14 |
|---|---|
|
|
|
|
| 15 |
| parameters | 20m |
|
| 16 |
| pretraining tokens | 15,824,042,105 (15.8B) |
|
|
|
|
|
|
|
|
|
|
| 17 |
| checkpoints here | 30 (steps 100–3000) |
|
| 18 |
| checkpoints saved by the run | 60 |
|
| 19 |
|