mountain-ner-model

Model Card for Model ID

<!-- Provide a quick summary of what the model is/does. -->

This project involves fine-tuning a BERT-based model (dslim/bert-large-NER) to perform Named Entity Recognition (NER) on mountain names in text.

The model has been trained to identify mentions of mountain names and differentiate them from other geographic entities or non-entities.

Features:

Fine-tuned on a custom dataset that includes sentences both with and without mountain names.

Uses focal loss to handle class imbalance, which ensures the model focuses on correctly classifying rare mountain names.

Token-level classification for identifying the B-MOUNTAIN, I-MOUNTAIN, and O (non-entity) labels.

Balances training between sentences with mountains (80%) and without mountains (20%).

Model Description

<!-- Provide a longer summary of what this model is. -->

This is the model card of a 🤗 transformers model that has been pushed on the Hub.

• Developed by: Oleksandr Kharytonov
• Model type: BERT
• Language(s) (NLP): Python
• License: MIT
• Finetuned from model [optional]: https://huggingface.co/dslim/bert-large-NER

Model Sources

<!-- Provide the basic links for the model. -->

• Repository: https://github.com/Shah1st/mountain-ner

Direct Use

<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->

from transformers import AutoTokenizer, AutoModelForTokenClassification

tokenizer = AutoTokenizer.from_pretrained('./saved_model')
model = AutoModelForTokenClassification.from_pretrained('./saved_model')

Recommendations

<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->

Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model.

How to Get Started with the Model

Use the github below to get started with the model.

https://github.com/Shah1st/mountain-ner

Training Details

Training Data

<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->

"DFKI-SLT/few-nerd", "supervised"

Filter for sentences with 'fine_ner_tags' == 24 (mountains)

Evaluation

<!-- This section describes the evaluation protocols and provides the results. -->

'eval_loss': 0.009154710918664932, 'eval_macro_f1': 0.8952192988290304, 'eval_accuracy': 0.9746226793108054

Testing Data, Factors & Metrics

Metrics

<!-- These are the evaluation metrics being used, ideally with a description of why. -->

macro F1: 0.895 Accuracy: 0.974

Summary

This project involves fine-tuning a BERT-based model (dslim/bert-large-NER) to perform Named Entity Recognition (NER) on mountain names in text. The model has been trained to identify mentions of mountain names and differentiate them from other geographic entities or non-entities.