KingTechnician/bert-osmosis-coverage-v2

Osmosis Coverage Classifier

This model classifies educational content coverage into three categories: low, fair, and satisfactory.

Model Description

• Base Model: distilbert-base-uncased
• Task: Multi-class text classification
• Classes: low (0), fair (1), satisfactory (2)
• Training Data: Custom osmosis dataset

Performance

• Overall Test Accuracy: 0.7600
• Macro F1 Score: 0.7584
• Macro Precision: 0.7571
• Macro Recall: 0.7600

Per-Class Performance

• Low: Precision: 0.8897, Recall: 0.9240, F1: 0.9065
• Fair: Precision: 0.6577, Recall: 0.6365, F1: 0.6469
• Satisfactory: Precision: 0.7239, Recall: 0.7193, F1: 0.7216

Usage

from transformers import AutoTokenizer, AutoModelForSequenceClassification
import torch

# Load model and tokenizer
tokenizer = AutoTokenizer.from_pretrained("KingTechnician/bert-osmosis-coverage-v2")
model = AutoModelForSequenceClassification.from_pretrained("KingTechnician/bert-osmosis-coverage-v2")

# Example usage
objective = "Your learning objective here"
response = "The response text to classify"

inputs = tokenizer(objective, response, return_tensors="pt", padding=True, truncation=True)
with torch.no_grad():
    outputs = model(**inputs)
    predictions = torch.nn.functional.softmax(outputs.logits, dim=-1)
    predicted_class = torch.argmax(predictions, dim=-1).item()
    
# Map prediction to label
id2label = {0: "low", 1: "fair", 2: "satisfactory"}
predicted_label = id2label[predicted_class]
print(f"Predicted coverage: {predicted_label}")

Training Details

• Learning Rate: 2e-5
• Batch Size: 64
• Epochs: 3
• Weight Decay: 0.01