Add parameter count to technical report
Browse files
technical_report_capacitor_decoder.md
CHANGED
|
@@ -45,6 +45,7 @@ configuration uses:
|
|
| 45 |
| LR schedule | constant after warmup |
|
| 46 |
| Warmup steps | `2,000` |
|
| 47 |
| Batch size | `128` |
|
|
|
|
| 48 |
| Gradient accumulation | `1` |
|
| 49 |
| Gradient clip | `1.0` max norm |
|
| 50 |
| Precision | AMP bfloat16 |
|
|
|
|
| 45 |
| LR schedule | constant after warmup |
|
| 46 |
| Warmup steps | `2,000` |
|
| 47 |
| Batch size | `128` |
|
| 48 |
+
| Parameters | `61.93M` |
|
| 49 |
| Gradient accumulation | `1` |
|
| 50 |
| Gradient clip | `1.0` max norm |
|
| 51 |
| Precision | AMP bfloat16 |
|