eval
Browse files
README.md
CHANGED
@@ -108,6 +108,19 @@ Epoch 1 | iter 512 step 8 | loss train: 11.970, val: n/a | iter time: 280.74 ms
|
|
108 |
Epoch 1 | iter 576 step 9 | loss train: 11.970, val: n/a | iter time: 279.90 ms (step) remaining time: 4 days, 19:44:24
|
109 |
Epoch 1 | iter 640 step 10 | loss train: 11.971, val: n/a | iter time: 279.74 ms (step) remaining time: 4 days, 17:59:44
|
110 |
# ...
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
111 |
```
|
112 |
|
113 |
Backup `wandb`:
|
|
|
108 |
Epoch 1 | iter 576 step 9 | loss train: 11.970, val: n/a | iter time: 279.90 ms (step) remaining time: 4 days, 19:44:24
|
109 |
Epoch 1 | iter 640 step 10 | loss train: 11.971, val: n/a | iter time: 279.74 ms (step) remaining time: 4 days, 17:59:44
|
110 |
# ...
|
111 |
+
Epoch 2 | iter 1364224 step 21316 | loss train: 3.433, val: 3.336 | iter time: 279.98 ms (step) remaining time: 0:00:04
|
112 |
+
Validating ...
|
113 |
+
Final evaluation | val loss: 3.336 | val ppl: 28.097
|
114 |
+
Saving checkpoint to '../out/pretrain-core-0/final/lit_model.pth'
|
115 |
+
----------------------------------------
|
116 |
+
| Performance
|
117 |
+
| - Total tokens : 11,186,768,000
|
118 |
+
| - Training Time : 209021.90 s
|
119 |
+
| - Tok/sec : 5430.54 tok/s
|
120 |
+
| ----------------------------------------
|
121 |
+
| Memory Usage
|
122 |
+
| - Memory Used : 19.86 GB
|
123 |
+
----------------------------------------
|
124 |
```
|
125 |
|
126 |
Backup `wandb`:
|