metadata
license: mit
tags:
- generated_from_trainer
datasets:
- squad
base_model: xlm-roberta-base
model-index:
- name: xlm-roberta-base_squad
results: []
xlm-roberta-base_squad
This model is a fine-tuned version of xlm-roberta-base on the squad dataset.
- "eval_exact_match": 82.69631031220435
- "eval_f1": 89.4562841806503
- "eval_samples": 10918
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: 3e-05
- train_batch_size: 32
- eval_batch_size: 32
- 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 4.14.1
- Pytorch 1.9.0
- Datasets 1.16.1
- Tokenizers 0.10.3