hing-roberta-NCM-run-2

This model is a fine-tuned version of l3cube-pune/hing-roberta on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 3.3647
• Accuracy: 0.6483
• Precision: 0.6369
• Recall: 0.6325
• F1: 0.6341

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: 3e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1
0.8973	1.0	927	0.8166	0.6483	0.6545	0.6576	0.6460
0.6827	2.0	1854	0.9071	0.6526	0.6444	0.6261	0.6299
0.4672	3.0	2781	1.1600	0.6764	0.6657	0.6634	0.6643
0.3388	4.0	3708	1.7426	0.6548	0.6406	0.6442	0.6418
0.2786	5.0	4635	1.9385	0.6505	0.6484	0.6437	0.6434
0.1794	6.0	5562	2.3158	0.6472	0.6564	0.6365	0.6388
0.12	7.0	6489	2.6961	0.6591	0.6458	0.6531	0.6466
0.1298	8.0	7416	2.7196	0.6505	0.6523	0.6307	0.6342
0.0941	9.0	8343	2.5853	0.6548	0.6406	0.6426	0.6415
0.0696	10.0	9270	2.8386	0.6613	0.6616	0.6314	0.6348
0.0722	11.0	10197	2.9658	0.6537	0.6356	0.6356	0.6355
0.0509	12.0	11124	3.3286	0.6429	0.6262	0.6192	0.6214
0.0444	13.0	12051	3.1654	0.6483	0.6347	0.6302	0.6319
0.0341	14.0	12978	2.9509	0.6537	0.6430	0.6394	0.6401
0.0345	15.0	13905	3.3416	0.6656	0.6514	0.6488	0.6499
0.0303	16.0	14832	3.3874	0.6419	0.6267	0.6339	0.6272
0.0245	17.0	15759	3.2854	0.6570	0.6428	0.6420	0.6421
0.0174	18.0	16686	3.2863	0.6602	0.6569	0.6427	0.6465
0.0136	19.0	17613	3.3674	0.6494	0.6361	0.6341	0.6349
0.0111	20.0	18540	3.3647	0.6483	0.6369	0.6325	0.6341

Framework versions

• Transformers 4.20.1
• Pytorch 1.10.1+cu111
• Datasets 2.3.2
• Tokenizers 0.12.1