lvkaokao
commited on
Commit
•
7cc4991
1
Parent(s):
830e8d3
update hyper-parameters
Browse files
README.md
CHANGED
|
@@ -34,8 +34,8 @@ The following hyperparameters were used during training:
|
|
| 34 |
- total_train_batch_size: 64
|
| 35 |
- total_eval_batch_size:
|
| 36 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 37 |
-
- lr_scheduler_type:
|
| 38 |
-
- lr_scheduler_warmup_ratio: 0.
|
| 39 |
- num_epochs: 2.0
|
| 40 |
|
| 41 |
## FP32 Inference with transformers
|
|
|
|
| 34 |
- total_train_batch_size: 64
|
| 35 |
- total_eval_batch_size:
|
| 36 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 37 |
+
- lr_scheduler_type: cosine
|
| 38 |
+
- lr_scheduler_warmup_ratio: 0.03
|
| 39 |
- num_epochs: 2.0
|
| 40 |
|
| 41 |
## FP32 Inference with transformers
|