Update README.md
Browse files
README.md
CHANGED
|
@@ -101,7 +101,7 @@ The data was processed into fixed-size chunks of 512 tokens, respecting document
|
|
| 101 |
|
| 102 |
### Pre-training
|
| 103 |
|
| 104 |
-
The model was pre-trained from scratch for **150,000 steps** on 8x NVIDIA A100 40GB GPUs, using BFloat16 (`bf16`) mixed-precision for stability and speed. The training took approximately **
|
| 105 |
|
| 106 |
The key hyperparameters used were:
|
| 107 |
|
|
|
|
| 101 |
|
| 102 |
### Pre-training
|
| 103 |
|
| 104 |
+
The model was pre-trained from scratch for **150,000 steps** on 8x NVIDIA A100 40GB GPUs, using BFloat16 (`bf16`) mixed-precision for stability and speed. The training took approximately **81 hours and 39 minutes** to complete.
|
| 105 |
|
| 106 |
The key hyperparameters used were:
|
| 107 |
|