|
--- |
|
license: wtfpl |
|
--- |
|
Trained for 500 steps with a lr of 0.003 and 4 steps gradient accumulation. |
|
|
|
 |
|
|
|
 |
|
|
|
 |
|
|
|
 |
|
|
|
 |
|
|
|
 |
|
|