test-electra-small-yelp
This model is a fine-tuned version of google/electra-small-discriminator on the yelp_review_full yelp_review_full dataset. It achieves the following results on the evaluation set:
- Loss: 2.2601
- Accuracy: 0.5677
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: 3.0
Training results
Framework versions
- Transformers 4.18.0.dev0
- Pytorch 1.10.0
- Datasets 1.18.3
- Tokenizers 0.11.0