learn_hf_food_not_food_text_classifier-distilbert-base-uncased

This model is a fine-tuned version of distilbert/distilbert-base-uncased on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.0005
• Accuracy: 1.0

Model description

This model is a fine-tuned version of distilbert-base-uncased for binary text classification: it predicts whether a piece of text is about food or not. It was built by Noorullah Zamindar, AI Engineer & Data Scientist, as a learning project to practice fine-tuning a transformer classifier with Hugging Face Transformers. On the held-out evaluation set it reports a loss of 0.0005 and an accuracy of 1.0.

Intended uses & limitations

Intended use: classifying short English text as food-related or not — for example, sorting food mentions in reviews, messages, or lists. This is a small learning project, not a production system. Limitations: it was trained on a small dataset (about 224 examples per epoch), so it may not generalize to new phrasings or domains; the reported 1.0 evaluation accuracy reflects the small held-out set and should not be read as perfect performance on unseen text. It handles English text only (distilbert-base-uncased).

Training and evaluation data

The training data details were not recorded when this model was published. The model was trained to distinguish food-related text from non-food text over 10 epochs (see the training hyperparameters and results below). Dataset details will be added here once confirmed.

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 0.0001
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 10

Training results

Framework versions

• Transformers 5.13.1
• Pytorch 2.11.0+cu128
• Datasets 5.0.0
• Tokenizers 0.22.2