banglat5-finetuned-new-method-new

This model is a fine-tuned version of csebuetnlp/banglat5 on the None dataset. It achieves the following results on the evaluation set:

• Rouge1 Precision: 26.1936
• Rouge1 Recall: 27.4839
• Rouge1 Fmeasure: 26.0831
• Rouge2 Precision: 10.5258
• Rouge2 Recall: 11.6544
• Rouge2 Fmeasure: 10.651
• Rougel Precision: 24.2917
• Rougel Recall: 25.5269
• Rougel Fmeasure: 24.1921
• Gen Len: 8.6402
• Loss: 2.8106

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: 0.0001
• 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: 5

Training results

Training Loss	Epoch	Step	Rouge1 Precision	Rouge1 Recall	Rouge1 Fmeasure	Rouge2 Precision	Rouge2 Recall	Rouge2 Fmeasure	Rougel Precision	Rougel Recall	Rougel Fmeasure	Gen Len	Validation Loss
4.1475	1.0	202	28.4204	31.0145	28.5051	12.7947	13.7716	12.4938	26.467	28.917	26.5293	11.3871	3.2546
3.741	2.0	404	25.0778	26.2399	24.7974	9.4574	10.3436	9.4444	23.2601	24.4534	23.0588	8.6824	2.9268
3.2953	3.0	606	25.9864	26.6166	25.5881	10.0107	10.6915	9.9945	23.7579	24.458	23.4435	8.4144	2.8387
2.9438	4.0	808	26.1936	27.4839	26.0831	10.5258	11.6544	10.651	24.2917	25.5269	24.1921	8.6402	2.8106

Framework versions

• Transformers 4.40.2
• Pytorch 2.2.1+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1