hosseinfarahi/squad-roberta-large

RoBERTa-large Model for SQuAD

This is a RoBERTa-large model fine-tuned on a combined dataset of SQuAD 1.1 and SQuAD 2.0. It achieves the following performance metrics:

• Exact Match (EM): 0.7003
• F1-score: 0.8368

The model was trained for 24 epochs using the following settings:

• Batch size: 2
• Learning rate: 3e-6
• Gradient accumulation steps: 4

Usage

You can load this model using the HuggingFace Transformers library:

from transformers import AutoModelForQuestionAnswering, AutoTokenizer

model_name = "hosseinfarahi/squad-roberta-large"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForQuestionAnswering.from_pretrained(model_name)