Update README.md
Browse files
README.md
CHANGED
|
@@ -26,7 +26,7 @@ Training was done using Huggingface and Deepspeed with ZeRO stage 2.
|
|
| 26 |
* per device batch size: 20
|
| 27 |
* GPUs: 8 x NVIDIA A100 40GB
|
| 28 |
* total batch size: 160
|
| 29 |
-
* steps:
|
| 30 |
* lowercase: currently yes, about to change
|
| 31 |
* fp16
|
| 32 |
* entire decoder was frozen
|
|
|
|
| 26 |
* per device batch size: 20
|
| 27 |
* GPUs: 8 x NVIDIA A100 40GB
|
| 28 |
* total batch size: 160
|
| 29 |
+
* steps: 12000
|
| 30 |
* lowercase: currently yes, about to change
|
| 31 |
* fp16
|
| 32 |
* entire decoder was frozen
|