hinglish-finetuned

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

• Loss: 2.0786

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: 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: 25
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
3.3784	1.0	80	3.0527
3.0398	2.0	160	2.8067
2.9133	3.0	240	2.7252
2.7872	4.0	320	2.5783
2.6205	5.0	400	2.5050
2.5979	6.0	480	2.4654
2.5655	7.0	560	2.4091
2.5412	8.0	640	2.3630
2.4479	9.0	720	2.3754
2.3724	10.0	800	2.2860
2.3842	11.0	880	2.2812
2.3411	12.0	960	2.2038
2.2617	13.0	1040	2.1887
2.3141	14.0	1120	2.1966
2.2115	15.0	1200	2.1248
2.2363	16.0	1280	2.1006
2.2191	17.0	1360	2.1248
2.1856	18.0	1440	2.0872
2.2009	19.0	1520	2.0299
2.2364	20.0	1600	2.0193
2.1785	21.0	1680	2.0227
2.1934	22.0	1760	2.0540
2.1479	23.0	1840	2.0381
2.0973	24.0	1920	1.9885
2.1376	25.0	2000	2.0142

Framework versions

• Transformers 4.18.0
• Pytorch 1.10.0+cu111
• Datasets 2.1.0
• Tokenizers 0.12.1