---
license: afl-3.0
---
A MacBERTh model fine-tuned on SQuAD_v2. Hopefully, this will allow the model to perform well on QA tasks on historical texts.
Finetune parameters:

```
training_args = TrainingArguments(
        output_dir="./results",
        evaluation_strategy="epoch",
        learning_rate=3e-5,
        per_device_train_batch_size=64,
        per_device_eval_batch_size=64,
        num_train_epochs=2,
        weight_decay=0.01,
        lr_scheduler_type=SchedulerType.LINEAR,
        warmup_ratio=0.2
    )
```