pretrain core 0
Browse files
scripts/pretrain-core-model-0.yaml
CHANGED
|
@@ -61,9 +61,11 @@ train:
|
|
| 61 |
global_batch_size: 512
|
| 62 |
# global_batch_size: 256
|
| 63 |
# global_batch_size: 128
|
|
|
|
| 64 |
|
| 65 |
# Number of samples per data-parallel rank (type: int, default: 4)
|
| 66 |
-
micro_batch_size:
|
|
|
|
| 67 |
# micro_batch_size: 2
|
| 68 |
# micro_batch_size: 1
|
| 69 |
|
|
|
|
| 61 |
global_batch_size: 512
|
| 62 |
# global_batch_size: 256
|
| 63 |
# global_batch_size: 128
|
| 64 |
+
# global_batch_size: 64
|
| 65 |
|
| 66 |
# Number of samples per data-parallel rank (type: int, default: 4)
|
| 67 |
+
micro_batch_size: 6
|
| 68 |
+
# micro_batch_size: 4
|
| 69 |
# micro_batch_size: 2
|
| 70 |
# micro_batch_size: 1
|
| 71 |
|