README.md

metadata

language:
  - en
base_model:
  - google-bert/bert-base-uncased

This modelcard aims to be a base template for new models. It has been generated using this raw template.

Model Details

Model Description

• Finetuned from model : google-bert/bert-base-uncased

Uses

This model classifies food recipe or ingredient into three disctinct categories: Vegan, Vegetarian and Non-vegetarian

Bias, Risks, and Limitations

This model is trained on controlled dataset.

Recommendations

Model should be fine-tuned on huge book corpus and large synthetic dataset.

How to Get Started with the Model

1. Download model.
2. Run testing script.

Training Details

Training Data

1. https://huggingface.co/datasets/rajputnavya/food-classification-mlm-clean
2. https://huggingface.co/datasets/rajputnavya/food-classification-nsp-format
3. https://huggingface.co/datasets/rajputnavya/food-classification-recipe-classification-data/blob/main/fine_tune_format.jsonl

Training Procedure

1. Training on mlm and nsp dataset combined
2. Fine-tuning on synthetic dataset for recipe classification

Results

{'accuracy': 0.9166666666666666, 'precision': 0.9333333333333332, 'recall': 0.9166666666666666, 'f1_score': 0.9153439153439153}