covid_qa_analysis_albert_base_squad_v2

This model is a fine-tuned version of abhilash1910/albert-squad-v2 on the covid_qa_deepset dataset.

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: 2e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• distributed_type: tpu
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 3

Training results

Framework versions

• Transformers 4.16.2
• Pytorch 1.9.0+cu102
• Datasets 1.18.3
• Tokenizers 0.11.6