metadata
license: apache-2.0
base_model: facebook/bart-large
tags:
- generated_from_trainer
model-index:
- name: question-generator
results: []
datasets:
- rajpurkar/squad
question-generator
This model is a fine-tuned version of facebook/bart-large on the SQUAD dataset. It achieves the following results on the evaluation set:
- Loss: 3.1315
Model description
More information needed
Intended uses & limitations
Sample Usage
def generate_question(context): inputs = tokenizer(context, return_tensors="pt") output = model.generate(**inputs) question = tokenizer.decode(output[0], skip_special_tokens=True) return question print(generate_question("Paris is the capital city of France"))
Training and evaluation data
This model is trained and evaluated on the SQUAD dataset
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-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
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 2.9322 | 1.0 | 10950 | 3.1334 |
| 2.6046 | 2.0 | 21900 | 3.1102 |
| 2.3742 | 3.0 | 32850 | 3.1315 |
Framework versions
- Transformers 4.42.3
- Pytorch 2.3.0+cu121
- Datasets 2.20.0
- Tokenizers 0.19.1