metadata
license: apache-2.0
base_model: distilbert-base-uncased
tags:
- generated_from_trainer
model-index:
- name: Distilbert-uncased
results: []
Distilbert-uncased-AS
Este es un modelo de finetuning de distilbert-base-uncased sobre un dataset propios de tweets. Se logra una error cuadrático bajo lo cual quiere decir que los valores predichos son muy cercanos a los observables o gold.
- Loss: 0.3510
- Rmse: 0.2543
Model description
More information needed
Intended uses & limitations
More information needed
Training and evaluation data
More information needed
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 5
Training results
| Training Loss | Epoch | Step | Validation Loss | Rmse |
|---|---|---|---|---|
| 0.2091 | 1.0 | 642 | 0.1933 | 0.3052 |
| 0.1334 | 2.0 | 1284 | 0.1909 | 0.2481 |
| 0.0684 | 3.0 | 1926 | 0.2617 | 0.2466 |
| 0.0355 | 4.0 | 2568 | 0.3113 | 0.2513 |
| 0.0116 | 5.0 | 3210 | 0.3510 | 0.2543 |
Framework versions
- Transformers 4.40.2
- Pytorch 2.0.1+cu117
- Datasets 2.18.0
- Tokenizers 0.19.1