JustinIA V1.0 2023 Revisión FINAL
Browse files
README.md
CHANGED
|
@@ -28,7 +28,7 @@ JustinIA V1.0 2023
|
|
| 28 |
## Model Details
|
| 29 |
|
| 30 |
Modelo Base: roBERTa
|
| 31 |
-
Dataset: Propietario
|
| 32 |
Publicado: 31 de marzo del 2023
|
| 33 |
|
| 34 |
### Model Description
|
|
@@ -53,14 +53,14 @@ Clasificación de acuerdo al tema de una sentencia en materia de justicia admini
|
|
| 53 |
|
| 54 |
5 ciclos de entrenamiento, con una precisión máxima de 28.52%
|
| 55 |
|
| 56 |
-
Epoch Training Loss Validation Loss Accuracy
|
| 57 |
-
1 1.982000 2.498654 0.274072
|
| 58 |
-
2 1.730200 2.486177 0.278226
|
| 59 |
-
3 1.551800 2.573319 0.273923
|
| 60 |
-
|
| 61 |
-
5 1.208200 2.623226 0.283863
|
| 62 |
|
| 63 |
-
TrainOutput(global_step=9930, training_loss=1.6167555717783273, metrics={'train_runtime': 1732.3659, 'train_samples_per_second': 57.32, 'train_steps_per_second': 5.732, 'total_flos': 2877096870824520.0, 'train_loss': 1.6167555717783273, 'epoch': 5.0})
|
| 64 |
|
| 65 |
#### Metrics
|
| 66 |
|
|
|
|
| 28 |
## Model Details
|
| 29 |
|
| 30 |
Modelo Base: roBERTa
|
| 31 |
+
Dataset: Propietario
|
| 32 |
Publicado: 31 de marzo del 2023
|
| 33 |
|
| 34 |
### Model Description
|
|
|
|
| 53 |
|
| 54 |
5 ciclos de entrenamiento, con una precisión máxima de 28.52%
|
| 55 |
|
| 56 |
+
-Epoch Training Loss Validation Loss Accuracy
|
| 57 |
+
-1 1.982000 2.498654 0.274072
|
| 58 |
+
-2 1.730200 2.486177 0.278226
|
| 59 |
+
-3 1.551800 2.573319 0.273923
|
| 60 |
+
-**4 1.359700 2.571429 0.285199**
|
| 61 |
+
-5 1.208200 2.623226 0.283863
|
| 62 |
|
| 63 |
+
-TrainOutput(global_step=9930, training_loss=1.6167555717783273, metrics={'train_runtime': 1732.3659, 'train_samples_per_second': 57.32, 'train_steps_per_second': 5.732, 'total_flos': 2877096870824520.0, 'train_loss': 1.6167555717783273, 'epoch': 5.0})
|
| 64 |
|
| 65 |
#### Metrics
|
| 66 |
|