nepali_t5

This model is a fine-tuned version of rujengelal/nepali_t5 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 2.6633
• Bleu: 6.3134
• Gen Len: 15.9835

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: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu	Gen Len
3.0928	1.0	17734	2.8330	5.4935	15.9053
3.101	2.0	35468	2.8127	5.5409	15.8787
3.0165	3.0	53202	2.7814	5.6622	15.9238
2.9973	4.0	70936	2.7532	5.8108	15.8996
2.8885	5.0	88670	2.7294	5.9077	15.8805
2.8114	6.0	106404	2.7074	6.1401	15.9749
2.7791	7.0	124138	2.6905	6.1567	15.9531
2.7729	8.0	141872	2.6782	6.1865	15.9688
2.7128	9.0	159606	2.6699	6.2233	16.063
2.7398	10.0	177340	2.6633	6.3134	15.9835

Framework versions

• Transformers 4.39.3
• Pytorch 2.1.2
• Datasets 2.18.0
• Tokenizers 0.15.2