Update README.md
Browse files
README.md
CHANGED
|
@@ -44,7 +44,7 @@ seed: 8
|
|
| 44 |
|
| 45 |
### dataset
|
| 46 |
dataset: code_290k
|
| 47 |
-
template:
|
| 48 |
cutoff_len: 8192
|
| 49 |
overwrite_cache: false
|
| 50 |
preprocessing_num_workers: 12
|
|
@@ -76,3 +76,22 @@ eval_steps: 1000
|
|
| 76 |
|
| 77 |
### BAdam training results
|
| 78 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 44 |
|
| 45 |
### dataset
|
| 46 |
dataset: code_290k
|
| 47 |
+
template: hermes_chatml
|
| 48 |
cutoff_len: 8192
|
| 49 |
overwrite_cache: false
|
| 50 |
preprocessing_num_workers: 12
|
|
|
|
| 76 |
|
| 77 |
### BAdam training results
|
| 78 |
|
| 79 |
+
| Training Loss | Epoch | Step | Validation Loss |
|
| 80 |
+
|:-------------:|:------:|:-----:|:---------------:|
|
| 81 |
+
| 0.7404 | 0.0559 | 1000 | 0.7784 |
|
| 82 |
+
| 0.7858 | 0.1118 | 2000 | 0.7702 |
|
| 83 |
+
| 0.7274 | 0.1677 | 3000 | 0.7604 |
|
| 84 |
+
| 0.6956 | 0.2236 | 4000 | 0.7570 |
|
| 85 |
+
| 0.7711 | 0.2795 | 5000 | 0.7541 |
|
| 86 |
+
| 0.7643 | 0.3354 | 6000 | 0.7518 |
|
| 87 |
+
| 0.8255 | 0.3913 | 7000 | 0.7496 |
|
| 88 |
+
| 0.7456 | 0.4472 | 8000 | 0.7483 |
|
| 89 |
+
| 0.7718 | 0.5031 | 9000 | 0.7447 |
|
| 90 |
+
| 0.6693 | 0.5590 | 10000 | 0.7445 |
|
| 91 |
+
| 0.7409 | 0.6149 | 11000 | 0.7433 |
|
| 92 |
+
| 0.7319 | 0.6709 | 12000 | 0.7424 |
|
| 93 |
+
| 0.7636 | 0.7268 | 13000 | 0.7415 |
|
| 94 |
+
| 0.7504 | 0.7827 | 14000 | 0.7414 |
|
| 95 |
+
| 0.7735 | 0.8386 | 15000 | 0.7374 |
|
| 96 |
+
| 0.7438 | 0.8945 | 16000 | 0.7375 |
|
| 97 |
+
| 0.839 | 0.9504 | 17000 | 0.7373 |
|