metadata
language: en
tags:
- protein-structure-prediction
- bioinformatics
- deep-learning
- ensemble-learning
- BiLSTM
- transformer
license: mit
datasets:
- IITM-NPPE2-Protein-Structure
metrics:
- f1
model-index:
- name: NPPE2-Protein-Structure-Ensemble
results:
- task:
type: protein-secondary-structure-prediction
name: Protein Secondary Structure Prediction
dataset:
name: NPPE-II
type: protein-sequences
metrics:
- type: f1_harmonic
value: 0.469
name: Test F1 Score
- type: f1_harmonic
value: 0.6287
name: Validation F1 Score
NPPE2 Protein Secondary Structure Prediction - Ensemble Model
Model Description
This is an ensemble of three deep learning models for protein secondary structure prediction, achieving 0.469 F1 score on the NPPE-II test set.
Key Achievement 🎯
Fixed Critical Bug: Q3 class 'C' (representing 42% of training data) was incorrectly mapped to index 0, causing it to be ignored by the loss function.
Citation
@misc{nppe2-protein-ensemble-2025, title={NPPE2 Protein Secondary Structure Prediction Ensemble}, author={IIT Madras BS Programme Student}, year={2025}, howpublished={\url{https://huggingface.co/YOUR_USERNAME/nppe2-protein-ensemble}}, note={Assignment submission for NPPE-II course} }
License
MIT License