mt5-small-nepali-final

This model is a fine-tuned version of aku47/mt5-small-nepali on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 1.7983
• Rouge1: 0.1664
• Rouge2: 0.0885
• Rougel: 0.1542
• Rougelsum: 0.1541
• Bertscore Precision: 0.7713
• Bertscore Recall: 0.6397
• Bertscore F1: 0.6988
• Gen Len: 21.0

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: 1e-05
• train_batch_size: 2
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 16
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 100
• num_epochs: 5

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Bertscore Precision	Bertscore Recall	Bertscore F1	Gen Len
3.0413	1.0	144	2.0856	0.1652	0.0878	0.15	0.1502	0.7758	0.6413	0.7017	20.9917
2.8167	2.0	288	1.8682	0.1726	0.0867	0.1556	0.1555	0.7745	0.6437	0.7026	21.0
2.6382	3.0	432	1.8175	0.1654	0.0859	0.1522	0.1521	0.7715	0.6408	0.6996	21.0
2.6705	4.0	576	1.7983	0.1664	0.0885	0.1542	0.1541	0.7713	0.6397	0.6988	21.0

Framework versions

• Transformers 4.53.0
• Pytorch 2.7.1+cu118
• Datasets 3.6.0
• Tokenizers 0.21.2