Update README.md
Browse files
README.md
CHANGED
|
@@ -24,7 +24,21 @@ coefficients only. This approach allows us to decouple the number of trainable p
|
|
| 24 |
|
| 25 |
## Evaluation Results
|
| 26 |
|
| 27 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 28 |
|
| 29 |
## Usage
|
| 30 |
|
|
|
|
| 24 |
|
| 25 |
## Evaluation Results
|
| 26 |
|
| 27 |
+
***** train metrics *****
|
| 28 |
+
epoch = 0.0907
|
| 29 |
+
total_flos = 4091386GF
|
| 30 |
+
train_loss = 1.618
|
| 31 |
+
train_runtime = 2:02:24.94
|
| 32 |
+
train_samples_per_second = 0.109
|
| 33 |
+
train_steps_per_second = 0.014
|
| 34 |
+
|
| 35 |
+
|
| 36 |
+
***** eval metrics *****
|
| 37 |
+
epoch = 0.0907
|
| 38 |
+
eval_loss = 1.5115
|
| 39 |
+
eval_runtime = 0:11:33.00
|
| 40 |
+
eval_samples_per_second = 0.144
|
| 41 |
+
eval_steps_per_second = 0.144
|
| 42 |
|
| 43 |
## Usage
|
| 44 |
|