Commit
•
b23fe7d
1
Parent(s):
92b6946
Update README.md
Browse files
README.md
CHANGED
|
@@ -41,9 +41,12 @@ Trained with LoRA targetting `["q_proj", "v_proj"]` in 4 bit and merged before u
|
|
| 41 |
|
| 42 |
The adapters are in the `adapters` branch.
|
| 43 |
|
|
|
|
|
|
|
|
|
|
| 44 |
### Training hyperparameters
|
| 45 |
|
| 46 |
-
The following hyperparameters were used during training:
|
| 47 |
- learning_rate: 0.0003
|
| 48 |
- train_batch_size: 12
|
| 49 |
- eval_batch_size: 12
|
|
|
|
| 41 |
|
| 42 |
The adapters are in the `adapters` branch.
|
| 43 |
|
| 44 |
+
Initial training investigation on the Tier-1 HPC of [Vlaams Supercomputer Centrum (VSC)](https://www.vscentrum.be/) and training on our own research cluster of 4x 3090s.
|
| 45 |
+
|
| 46 |
+
|
| 47 |
### Training hyperparameters
|
| 48 |
|
| 49 |
+
The following hyperparameters were used during training in the HPC investigation:
|
| 50 |
- learning_rate: 0.0003
|
| 51 |
- train_batch_size: 12
|
| 52 |
- eval_batch_size: 12
|