Update README.md
Browse files
README.md
CHANGED
|
@@ -65,7 +65,7 @@ The following techniques were used to shorten training time:
|
|
| 65 |
#### Phase 1
|
| 66 |
- **Hardware:** 6 x 8 x H100 (80GB)
|
| 67 |
- **Optimizer:** LAMB
|
| 68 |
-
- **Batch:**
|
| 69 |
- **Learning rate:** 5e-03
|
| 70 |
|
| 71 |
#### Phases 2-4
|
|
|
|
| 65 |
#### Phase 1
|
| 66 |
- **Hardware:** 6 x 8 x H100 (80GB)
|
| 67 |
- **Optimizer:** LAMB
|
| 68 |
+
- **Batch:** 18432
|
| 69 |
- **Learning rate:** 5e-03
|
| 70 |
|
| 71 |
#### Phases 2-4
|