Update README.md
Browse files
README.md
CHANGED
|
@@ -13,8 +13,26 @@ datasets:
|
|
| 13 |
- test μ€νλ¦Ώ μμ 1,000κ°μ μνλ§ μΈ‘μ
|
| 14 |
|
| 15 |
|
|
|
|
| 16 |
|
| 17 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 18 |
|
| 19 |
TrainOutput(
|
| 20 |
global_step=300, training_loss=2.666887741088867,
|
|
@@ -26,7 +44,7 @@ metrics={'train_runtime': 961.226,
|
|
| 26 |
|
| 27 |
|
| 28 |
|
| 29 |
-
|
| 30 |
|
| 31 |
Midm: μ νλ 0.88
|
| 32 |
|
|
@@ -37,6 +55,6 @@ Midm: μ νλ 0.88
|
|
| 37 |
|
| 38 |
|
| 39 |
|
| 40 |
-
|
| 41 |
|
| 42 |
cxoijve
|
|
|
|
| 13 |
- test μ€νλ¦Ώ μμ 1,000κ°μ μνλ§ μΈ‘μ
|
| 14 |
|
| 15 |
|
| 16 |
+
## Training procedure
|
| 17 |
|
| 18 |
+
### Training hyperparameters
|
| 19 |
+
|
| 20 |
+
The following hyperparameters were used during training:
|
| 21 |
+
- learning_rate: 0.0001
|
| 22 |
+
- train_batch_size: 1
|
| 23 |
+
- eval_batch_size: 1
|
| 24 |
+
- seed: 42
|
| 25 |
+
- gradient_accumulation_steps: 2
|
| 26 |
+
- total_train_batch_size: 2
|
| 27 |
+
- optimizer: adam_beta1=0.9, adam_beta2=0.999, adam_epsilon=1e-08,
|
| 28 |
+
- lr_scheduler_type: cosine
|
| 29 |
+
- lr_scheduler_warmup_ratio: 0.03
|
| 30 |
+
- training_args.logging_steps: 50
|
| 31 |
+
- training_args.max_steps : 300
|
| 32 |
+
- trainable params: 16,744,448 || all params: 7,034,347,520 || trainable%: 0.23803839591934178
|
| 33 |
+
|
| 34 |
+
|
| 35 |
+
### Training Results
|
| 36 |
|
| 37 |
TrainOutput(
|
| 38 |
global_step=300, training_loss=2.666887741088867,
|
|
|
|
| 44 |
|
| 45 |
|
| 46 |
|
| 47 |
+
### Accuracy
|
| 48 |
|
| 49 |
Midm: μ νλ 0.88
|
| 50 |
|
|
|
|
| 55 |
|
| 56 |
|
| 57 |
|
| 58 |
+
### Model Card Authors
|
| 59 |
|
| 60 |
cxoijve
|