<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->
test_trainer
This model is a fine-tuned version of cardiffnlp/twitter-roberta-base-sentiment on an unknown dataset. It achieves the following results on the evaluation set:
- eval_loss: 0.9501
- eval_accuracy: 0.491
- eval_runtime: 64.4098
- eval_samples_per_second: 31.051
- eval_steps_per_second: 3.881
- step: 0
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: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3
Framework versions
- Transformers 4.25.1
- Pytorch 1.13.0+cu116
- Datasets 2.8.0
- Tokenizers 0.13.2