Hugging Face's logo

Kiran2004
/
Roberta_QCA_Squad

Question Answering
Transformers
TensorFlow
roberta
generated_from_keras_callback
Inference Endpoints
Model card Files Community
Roberta_QCA_Squad / README.md
Kiran2004's picture
Kiran2004
Update README.md
a7fd238 verified
preview code
|
raw
history blame contribute delete
2.84 kB
metadata
license: cc-by-4.0
base_model: deepset/roberta-base-squad2
tags:
  - generated_from_keras_callback
model-index:
  - name: Kiran2004/Roberta_QCA_Squad
    results: []
datasets:
  - rajpurkar/squad
metrics:
  - accuracy
  - precision
  - recall
  - f1

Model description

This model is a fine-tuned version of deepset/roberta-base-squad2 on an squad dataset.It's been trained on question-answer pairs, including unanswerable questions, for the task of Question Answering for 6 Epochs. It achieves the following results after training:

  • Train Loss: 0.1434
  • Validation Loss: 0.4821

Model Training

Evaluation

The model's performance can be evaluated using various metrics such as Accuracy, Recall, Precision, F1 score.

  • Accuracy: 0.9100
  • Precision: 0.9099
  • Recall: 0.9099
  • F1 Score: 0.9603

Example Usage 

from transformers import pipeline

model_name = "Kiran2004/Roberta_QCA_Squad"

question_answerer = pipeline("question-answering", model = model_name)

question = "How many programming languages does BLOOM support?"
context = "BLOOM has 176 billion parameters and can generate text in 46 languages natural languages and 13 programming languages."

question_answerer(question=question, context=context)

Training hyperparameters

The following hyperparameters were used during training:

  • optimizer: {'name': 'Adam', 'weight_decay': None, 'clipnorm': None, 'global_clipnorm': None, 'clipvalue': None, 'use_ema': False, 'ema_momentum': 0.99, 'ema_overwrite_frequency': None, 'jit_compile': True, 'is_legacy_optimizer': False, 'learning_rate': {'module': 'keras.optimizers.schedules', 'class_name': 'PolynomialDecay', 'config': {'initial_learning_rate': 2e-05, 'decay_steps': 250, 'end_learning_rate': 0.0, 'power': 1.0, 'cycle': False, 'name': None}, 'registered_name': None}, 'beta_1': 0.9, 'beta_2': 0.999, 'epsilon': 1e-08, 'amsgrad': False}
  • training_precision: float32

Training results

Train Loss Validation Loss Epoch
0.5774 0.4305 0
0.3089 0.4597 1
0.2268 0.4541 2
0.1852 0.4718 3
0.1618 0.4821 4
0.1434 0.4821 5

Framework versions

  • Transformers 4.38.2
  • TensorFlow 2.15.0
  • Datasets 2.18.0
  • Tokenizers 0.15.2