metadata

tags:
  - generated_from_trainer
metrics:
  - accuracy
base_model: Rajan/NepaliBERT
model-index:
  - name: nepali_complaints_classification_nepbert3
    results: []

nepali_complaints_classification_nepbert3

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

Loss: 0.2687
Accuracy: 0.9494
F1-score: 0.9483

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 2e-05
train_batch_size: 16
eval_batch_size: 16
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
lr_scheduler_warmup_steps: 50
num_epochs: 5

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1-score
1.4921	0.22	500	0.8642	0.7235	0.7143
0.7781	0.45	1000	0.6241	0.7974	0.7923
0.5865	0.67	1500	0.5342	0.8243	0.8125
0.4625	0.89	2000	0.4250	0.8576	0.8553
0.3648	1.11	2500	0.3856	0.8759	0.8725
0.3001	1.34	3000	0.3424	0.8899	0.8891
0.2723	1.56	3500	0.3199	0.9007	0.8981
0.2538	1.78	4000	0.2898	0.9085	0.9066
0.231	2.01	4500	0.2676	0.9203	0.9189
0.1478	2.23	5000	0.3029	0.9210	0.9187
0.1666	2.45	5500	0.2580	0.9283	0.9271
0.1519	2.67	6000	0.2573	0.9308	0.9292
0.1498	2.9	6500	0.2746	0.9328	0.9306
0.1112	3.12	7000	0.2564	0.9398	0.9389
0.0903	3.34	7500	0.2726	0.9403	0.9393
0.1036	3.57	8000	0.2664	0.9398	0.9385
0.1043	3.79	8500	0.2614	0.9459	0.9447
0.0972	4.01	9000	0.2499	0.9453	0.9443
0.0663	4.23	9500	0.2643	0.9469	0.9458
0.0683	4.46	10000	0.2688	0.9474	0.9462
0.0671	4.68	10500	0.2657	0.9491	0.9481
0.0605	4.9	11000	0.2687	0.9494	0.9483

Framework versions

Transformers 4.38.2
Pytorch 2.1.0+cu121
Datasets 2.18.0
Tokenizers 0.15.2