GautamR/agri_bert_classifier

Agri-flow Classification Model

This model classifies grievances into five distinct buckets:

• agricultural_scheme
• agriculture
• pest
• seed
• weather
• price
• non_agri

Description of the Buckets

1. agricultural_scheme: The farmer query is about schemes in Odisha

2. agriculture: General agri queries

3. pest: The farmer query is about pests

4. seed: The farmer query is about seed varieties

5. weather : The farmer query is asking about the weather for a district /place e.g. : 'What's the weather forecast for Sundargarh?'

6. price : The farmer query is asking about the price of some crop e.g. 'Price for paddy'

7. non_agri : The farmer query is just some salutation or unrelated to agri

Training Metrics

Epoch 1/1000 - Loss: 0.8210 - Accuracy: 0.7443 - F1 Score: 0.7360 Validation Accuracy: 0.9037 Validation F1 Score: 0.9022 Epoch 2/1000 - Loss: 0.2868 - Accuracy: 0.9199 - F1 Score: 0.9197 Validation Accuracy: 0.9241 Validation F1 Score: 0.9236 Epoch 3/1000 - Loss: 0.1620 - Accuracy: 0.9536 - F1 Score: 0.9534 Validation Accuracy: 0.9408 Validation F1 Score: 0.9407 Epoch 4/1000 - Loss: 0.0975 - Accuracy: 0.9698 - F1 Score: 0.9698 Validation Accuracy: 0.9457 Validation F1 Score: 0.9461 Epoch 5/1000 - Loss: 0.0722 - Accuracy: 0.9777 - F1 Score: 0.9777 Validation Accuracy: 0.9518 Validation F1 Score: 0.9520 Epoch 6/1000 - Loss: 0.0570 - Accuracy: 0.9801 - F1 Score: 0.9801 Validation Accuracy: 0.9574 Validation F1 Score: 0.9573 Epoch 7/1000 - Loss: 0.0426 - Accuracy: 0.9838 - F1 Score: 0.9838 Validation Accuracy: 0.9601 Validation F1 Score: 0.9601 Epoch 8/1000 - Loss: 0.0403 - Accuracy: 0.9850 - F1 Score: 0.9850 Validation Accuracy: 0.9646 Validation F1 Score: 0.9646 Epoch 9/1000 - Loss: 0.0340 - Accuracy: 0.9853 - F1 Score: 0.9853 Validation Accuracy: 0.9623 Validation F1 Score: 0.9624 Epoch 10/1000 - Loss: 0.0307 - Accuracy: 0.9857 - F1 Score: 0.9857 Validation Accuracy: 0.9640 Validation F1 Score: 0.9640 Epoch 11/1000 - Loss: 0.0297 - Accuracy: 0.9873 - F1 Score: 0.9873 Validation Accuracy: 0.9618 Validation F1 Score: 0.9618 Epoch 12/1000 - Loss: 0.0279 - Accuracy: 0.9867 - F1 Score: 0.9867 Validation Accuracy: 0.9607 Validation F1 Score: 0.9607